Anda di halaman 1dari 198

API/EI Research Report 545-A

Verification of lightning protection


requirements for above ground
hydrocarbon storage tanks

First edition, October 2009


API/EI Research Report 545-A

Verification of lightning protection requirements for above ground


hydrocarbon storage tanks

First edition

October 2009

Published jointly by

API
and
ENERGY INSTITUTE LONDON
The Energy Institute is a professional membership body incorporated by Royal Charter 2003
Registered charity number 1097899
Copyright © 2009 by API, and
The Energy Institute, London:
The Energy Institute is a professional membership body incorporated by Royal Charter 2003.
Registered charity number 1097899, England
All rights reserved.

No part of this book may be reproduced by any means, or transmitted or translated into a machine language without
the written permission of the publisher.

ISBN 978 0 85293 376 3

Published by the Energy Institute

The information contained in this publication is provided as guidance only and while every reasonable care has been
taken to ensure the accuracy of its contents, the Energy Institute cannot accept any responsibility for any action taken,
or not taken, on the basis of this information. The Energy Institute shall not be liable to any person for any loss or
damage which may arise from the use of any of the information contained in any of its publications.

Further copies can be obtained from:


Portland Customer Services, Commerce Way, Whitehall Industrial Estate, Colchester, CO2 8HP, UK.
t: +44 (0) 1206 796 351 e: sales@portland-services.com

or

Global Engineering Documents at (800) 854-7179 or (303) 397-7956


w: www.global.ihs.com

Electronic access to EI publications is available via our website, www.energyinstpubs.org.uk.


Documents can be purchased online as downloadable pdfs or on an annual subscription for single users and
companies.
For more information, contact the EI Publications Team e: pubs@energyinst.org
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

CONTENTS OVERVIEW

Page

Legal notices and disclaimers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv

Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v

PHASE 1

Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . i

Executive summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ii

Cul/LT-0234
Review of lightning phenomena and the interaction with above ground storage tanks . . 1

Cul/LT-0235
Review of tank base earthing and test current recommendations . . . . . . . . . . . . . . . . . . . 43

EI-EN2-04
Lightning tests to tank shell/shunt samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

EI-Vis1-01
Visit to oil refinery A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

EI-Vis2-02
Visit to oil refinery B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

EI-TN1-03
Review of burn-through and hot-spot effects on metallic tank skins from lightning strikes . . 93

PHASE 2

Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . i

Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ii

Cul/LT-0373 Lightning simulation testing to determine the required characteristics for


roof bonding cables on external floating roof above ground storage tanks . . . . . . . . . 1

Cul/LT-0401 Investigative tests on the lightning protection of submerged shunts with


parallel roof bonding cables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

Phase 1:iii
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

LEGAL NOTICES AND DISCLAIMERS

The information contained in this publication is provided as guidance only, and although every effort
has been made by API and EI to assure the accuracy and reliability of its contents, API and EI MAKE
NO GUARANTEE THAT THE INFORMATION HEREIN IS COMPLETE OR ERROR-FREE. ANY PERSON OR
ENTITY MAKING ANY USE OF THE INFORMATION HEREIN DOES SO AT HIS/HER/ITS OWN RISK. TO
THE MAXIMUM EXTENT PERMITTED BY APPLICABLE LAW, THE INFORMATION HEREIN IS PROVIDED
WITHOUT, AND API AND EI HEREBY EXPRESSLY DISCLAIM, ANY REPRESENTATION OR WARRANTY
OF ANY KIND, WHETHER EXPRESS, IMPLIED OR STATUTORY, INCLUDING, WITHOUT LIMITATION,
WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, TITLE AND NON-
INFRINGEMENT. IN NO EVENT SHALL API OR EI BE LIABLE TO ANY PERSON, OR ENTITY USING OR
RECEIVING THE INFORMATION HEREIN FOR ANY CONSEQUENTIAL, INCIDENTAL, PUNITIVE, INDIRECT
OR SPECIAL DAMAGES (INCLUDING, WITHOUT LIMITATION, LOST PROFITS), REGARDLESS OF THE
BASIS OF SUCH LIABILITY, AND REGARDLESS OF WHETHER OR NOT API OR EI HAS BEEN ADVISED OF
THE POSSIBILITY OF SUCH DAMAGES OR IF SUCH DAMAGES COULD HAVE BEEN FORESEEN.

The contents of this publication are not intended or designed to define or create legal rights or
obligations, or set a legal standard of care.

API or EI is not undertaking to meet the duties of manufacturers, purchasers, users and/or employers
to warn and equip their employees and others concerning safety risks and precautions, nor is API or EI
undertaking any of the duties of manufacturers, purchasers, users and/or employers under local and
regional laws and regulations. This information should not be used without first securing competent
advice with respect to its suitability for any general or specific application, and all entities have an
independent obligation to ascertain that their actions and practices are appropriate and suitable for
each particular situation and to consult all applicable federal, state and local laws.

API AND EI HEREBY EXPRESSLY DISCLAIM ANY LIABILITY OR RESPONSIBILITY FOR LOSS OR DAMAGE
RESULTING FROM THE VIOLATION OF ANY LOCAL OR REGIONAL LAWS OR REGULATIONS WITH
WHICH THIS PUBLICATION MAY CONFLICT.

Nothing contained in any API or EI publication is to be construed as granting any right, by implication
or otherwise, for the manufacture, sale, or use of any method, apparatus, or product covered by
letters patent. Neither should anything contained in the publication be construed as insuring anyone
against liability for infringement of letters patent.

No reference made in this publication to any specific product or service constitutes or implies an
endorsement, recommendation, or warranty thereof by API or EI.

API, EI, AND THEIR AFFILIATES, REPRESENTATIVES, CONSULTANTS, AND CONTRACTORS AND
THEIR RESPECTIVE PARENTS, SUBSIDIARIES, AFFILIATES, CONSULTANTS, OFFICERS, DIRECTORS,
EMPLOYEES, REPRESENTATIVES, AND MEMBERS SHALL HAVE NO LIABILITY WHATSOEVER FOR, AND
SHALL BE HELD HARMLESS AGAINST, ANY LIABILITY FOR ANY INJURIES, LOSSES OR DAMAGES OF
ANY KIND, INCLUDING DIRECT, INDIRECT, INCIDENTAL, CONSEQUENTIAL, OR PUNITIVE DAMAGES,
TO PERSONS, INCLUDING PERSONAL INJURY OR DEATH, OR PROPERTY RESULTING IN WHOLE
OR IN PART, DIRECTLY OR INDIRECTLY, FROM ACCEPTANCE, USE OR COMPLIANCE WITH THIS
PUBLICATION.

Phase 1:iv
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

FOREWORD

This publication has been produced at the request of the API RP 545 Task Force and the EI Electrical
Committee.

It collates a number of research reports produced by Culham Electromagnetics and Lightning Limited
(Culham) who were commissioned to investigate the lightning phenomena and the adequacy of
lightning protection measures on above ground hydrocarbon storage tanks.

Currently international, British and United States standards contain requirements relating to lighting
protection; however, these have not been verified through practical, scientific testing. As a result of
the work commissioned by the API and EI, a new Recommended Practice (RP) is being developed
which will incorporate the results of this investigation.

Suggested revisions are invited and should be submitted to the director of standards, API, 1220
L Street, N.W., Washington, D.C. 20005 or The Technical Department, Energy Institute, 61 New
Cavendish Street, London, W1G 7AR.

Phase 1:v
API/EI RESEARCH REPORT

VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS


FOR ABOVE GROUND HYDROCARBON STORAGE TANKS

PHASE 1
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

EXECUTIVE SUMMARY

CUL/LT-0234
REVIEW OF LIGHTNING PHENOMENA AND THE INTERACTION WITH ABOVE GROUND
STORAGE TANKS

This document describes the phenomena of lightning, and how it is expected to interact with various
types of tank designs, in particular with respect to the problems of hydrocarbon fires. It draws on
the experience of refinery and tank farm visits by the author, including those reported in EI-Vis1-01
Visit to oil refinery A and EI-Vis2-02 Visit to oil refinery B, and other sources. The likely strike points
on all types of tanks are described, and the current routes over tanks are shown for the fast and
slow lightning components. The shunt/rim seal region for an open top FRT is shown to be the most
susceptible to ignitions. Aluminium roof geodesic tanks appear to be the most likely type to suffer
hot spots and burn through. The electrical properties of steel as a material for tanks is described
along with descriptions of thermal and voltage sparking. Protection strategies for open FRTs, for
roofed over tanks, and LPG tanks are described. Comments are made on the problems associated
with petroleum product within the pontoons or on the surface of a floating roof. An analysis of the
operation of the shunt/shell bonding cable suggests that it would play a vital part in suppressing
sparking from the continuing current component of lightning, although it would play only a minor
role in suppressing sparking from the fast component. The principal USA/UK lightning protection and
oil industry documents are reviewed for their content on lightning protection of tanks.

CUL/LT-0235
REVIEW OF TANK BASE EARTHING AND TEST CURRENT RECOMMENDATIONS

The earthing of a storage tank may have important considerations for safety, and protection of
instrumentation on the tank, but in practice the tank is likely to be intrinsically well earthed simply
by its construction. Even so earthing rods should be (and are) generally used as recommended by
international standards. Quality of earthing has little or no significance in storage tank fire protection.
The document also discusses the likely currents which shunts may have to carry (up to 11 kA).

EI-EN2-04
LIGHTNING TESTS TO TANK SHELL/SHUNT SAMPLES

Tests at Culham replicated shunt/shell interfaces and subjected them to conducted lightning-type
currents. Even clean steel shunt/shell interfaces sparked. Fast current components produce relatively
small sparks, whereas long duration currents produce copious spark showers that are believed to be
more hazardous. (In practice good protection against the latter currents can be achieved using a roof
bonding cable.) Different shunt materials could also present less of a hazard. Currents in immersed
shunts tended to cause an eruption of fluid, due to the arc pressure.

Phase 1:ii
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

EI-VIS1-01 EI-VIS2-02
VISIT TO OIL REFINERY A VISIT TO OIL REFINERY B

Describe the features seen during two visits to refineries. Some of the practical difficulties of providing
and maintaining lightning protection in the field and over many years are observed and discussed,
and some photographs are included.

EI-TN1-03
REVIEW OF BURN-THROUGH AND HOT-SPOT EFFECTS ON METALLIC TANK SKINS FROM
LIGHTNING STRIKES

Discusses the threat of lightning strikes puncturing steel or aluminium tanks, or of causing internal
hot-spots. Aluminium is easily punctured, and so geodesic roofs which use aluminium skins < 2 mm
thick would be a hazard if they contained vapours within the flammable range. Steel skins 5 mm
thick would not be expected to be punctured by a lightning attachment. Hot-spot hazards for such
thick skins have not been investigated, but could be a hazard for severe strikes.

MAIN CONCLUSIONS AND RECOMMENDATIONS

1. Potential ignition hazards exist particularly at shunt/shell interfaces (for open top FRT) and
geodesic roofs. The nature of the hydrocarbon, as well as temperature and ventilation,
determines whether the vapour could be within a flammable range.
2. Sparking at shunts is inevitable, and is more severe for the long duration currents. Such
sparks tend to fall downwards into the seal region, and any gaps between the seal and the
shell would increase the likelihood of flammable vapour ignition. Therefore maintenance of
the seal is important.
3. The severity of sparking can be significantly reduced, by using earth cables from the floating
roof to the shell, or bonding via the ladder.
4. Immersed shunts should present a good solution to hazardous shunt sparking.

Phase 1:iii
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

CUL/LT-0234

REVIEW OF LIGHTNING PHENOMENA AND THE INTERACTION


WITH ABOVE GROUND STORAGE TANKS

CONTENTS

Page

1 Introduction: The lightning phenomenon as relevant to above ground storage tanks . 3


1.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Lightning strike rates to ground and isokeraunic levels . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Electrical parameters of lightning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.4 Lightning risk and protection principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.5 The lightning cloud approach, induction charging, and space charges . . . . . . . . . . . . . . 6
1.6 Roof surface charging of tanks and bulk product charging . . . . . . . . . . . . . . . . . . . . . . . 8
1.7 Attachment process to tanks and ground objects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

2 The interaction of lightning with storage tanks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10


2.1 Floating roof tanks (FRT) and lightning induced fires . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.2 Discussion of lightning interaction with open top FRTs . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.3 Lightning interaction with other types of above ground storage tanks . . . . . . . . . . . . . 13
2.4 Lightning interaction with geodesic lightweight roofs . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.5 Cone roof, or 'fixed roof' tanks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.6 Pressurised tanks for LPG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

3 Lightning effects on materials: electrical heating and mechanical damage;


burn through and hot-spots; thermal and voltage sparking . . . . . . . . . . . . . . . . . . . . . 17
3.1 Effects due to lightning current flow through metallic conductors . . . . . . . . . . . . . . . . 17
3.2 Effects at the lightning arc attachment point . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.3 Danger arising from sparking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

4 Floating roof rim seals and the lightning induced fire problem . . . . . . . . . . . . . . . . . . . 20
4.1 Introduction to sparking at seals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
4.2 Floating roof seal types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
4.3 Discussion of sparking sources on seals and shunts . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
4.4 Lightning risks to open FRTs from surface product and leaks into pontoons . . . . . . . . . 24

5 Lightning ignitions on roofed over floating roof tanks . . . . . . . . . . . . . . . . . . . . . . . . . 26


5.1 Introduction to roofing of open top tanks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
5.2 Lightning hazards with roofed over tanks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

6 Lightning hazards to cone roof (fixed roof) tanks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

7 Lightning hazards to LPG tanks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

8 Summary of lightning effects on above ground storage tanks . . . . . . . . . . . . . . . . . . . 31


8.1 The lightning process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
8.2 Effects on materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

Phase 1:1
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

8.3 Protection methods to prevent lightning induced rim seal fires . . . . . . . . . . . . . . . . . . . 32


8.4 Protection of roofed over tanks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
8.5 Lightning effects on roof surface product, and product in pontoons . . . . . . . . . . . . . . 33
8.6 Lightning effects on cone roof tanks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
8.7 Lightning effects on pressurised LPG tanks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

9 Conclusions and recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

10 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

Annexes:

Annex A Skin effect in steel sheets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

Annex B Review of documents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38


B.1 BS6651:1999 Code of practice on protection of buildings against lightning . . . . . . . . . 38
B.2 API 545 Recommended practice for lightning protection of above ground
hydrocarbon storage tanks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
B.3 NFPA 780 Standard for the installation of lightning protection systems , 2000 Edition . . . 38
B.4 API 650 Welded steel tanks for oil storage 10th edition. Nov 98 with add 3. . . . . . . . . 39
B.5 EEMUA 159 User's guide to the inspection, maintenance and repair of above
ground vertical cylindrical steel storage tanks, 3rd edition 2003 . . . . . . . . . . . . . . . . . . 39
B.6 NFPA 30 Flammable and combustible liquids code 2003 edition . . . . . . . . . . . . . . . . . . 40
B.7 Review of information from chicago bridge and iron . . . . . . . . . . . . . . . . . . . . . . . . . . 40

Annex C World isokeraunic levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

Figures:
Figure 1 Ref 1013 Huffines and Orville American meteorological society lightning ground
flash density and thunderstorm duration in the continental United States . . . . . . . . . 3
Figure 2 Schematic diagrams of the high current and the 'continuing' current
components of a lightning strike. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Figure 3 'Cone of protection' principle for lightning protection. . . . . . . . . . . . . . . . . . . . . . . . 6
Figure 4 The process of attachment from a lightning cloud to a tank . . . . . . . . . . . . . . . . . . . 9
Figure 5a Current routes for flash to top of shell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Figure 5b Current routes for flash to floating roof . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Figure 5c Current routes for flash to ground near a floating roof tank . . . . . . . . . . . . . . . . . . 12
Figure 6 Open top floating roof tanks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Figure 7 Typical fixed (cone) roof tank with vent at the top . . . . . . . . . . . . . . . . . . . . . . . . . . 15
Figure 8 LPG tank, all steel, pressurised . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Figure 9 Effects at the arc attachment point . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Figure 10 Typical floating roof pantograph type primary seal, with secondary seal,
shunt and foam dam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Figure A.1 Cross section of the top of the steel shell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

Tables:
Table 1 Lightning flash parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

Phase 1:2
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

1 INTRODUCTION: THE LIGHTNING PHENOMENON AS


RELEVANT TO ABOVE GROUND STORAGE TANKS

1.1 GENERAL

This section provides a summary of the present knowledge on the lightning flash process
and attachment mechanism so that the hazard to tanks can be described in terms of natural
lightning behaviour.

1.2 LIGHTNING STRIKE RATES TO GROUND AND ISOKERAUNIC LEVELS

Taking lightning generally, there is a very large variation in strike rates to ground over the
whole globe. This is illustrated in the world isokeraunic map shown in Annex C. Such a map
is not ideal for determining the risk to ground structures since it is a measure of the number
of thunderstorm days per year, which can only be turned approximately into the desired
parameter, that of strike rate/km2/year. A rough formula for this is given in Equation 1:

(1) Ng = 0.04 Td1.25 +300%/-60%

where Ng is the average strike rate in strikes/km2/year to the ground, and Td is the ceraunic
number, that is, the number of thunderstorm days a year. There are, however, maps of Ng
itself available for many parts of the world, mainly those countries that have invested in
lightning measuring systems like UK, USA, and some European countries. An example of
such a map is shown in Figure 1. Equivalent information for the UK is given in BS 6651:1999
Code of practice on protection of buildings against lightning.

Figure 1 Ref 1013 Huffines and Orville American meteorological society lightning
ground flash density and thunderstorm duration in the continental United States:
July 1999

0 0.5 1 3 5 7 9 11> Flashes km² yr

Mean annual flash density (1989-1996)

Phase 1:3
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

However, in the absence of such data, Equation 1 can be used to give an approximate value.
Knowing Ng, the local strike rate, the risk of a strike to any oil processing or storage tanks
can be established, essentially as a simple calculation of the ground surface area of the plant
multiplied by Ng. (There are a few refinements to put into this calculation to allow for ground
topology, height of structures etc, but this is the basis of the calculation of strike risk.)
Thus the likelihood of lightning damage varies very widely all over the world being
highest in humid tropical areas, (≥10 strikes/km²/year) and very low in very dry and very cold
areas (<1 strike/km²/year). (The risk of damage also depends on the type of plant: see below.)
Lightning also varies according to the season and the time of year. In the UK for example,
the highest lightning activity is in the summer. A diurnal variation is also evident. (A statistical
correlation between actual tank fires and thunderstorm days has been included in previous
LASTFIRE documentation[5].)

1.3 ELECTRICAL PARAMETERS OF LIGHTNING

The currents in any lightning flash (a flash usually comprises several components or strokes)
can range from just a few kiloamperes (kA) up to over 100 kA. This applies to short duration
strokes, each of which lasts no longer than 100 microseconds (µs).
There are also long duration components called 'continuing current' which can be of
the order of two or three hundred amperes, lasting up to a few hundred milliseconds (ms).
Most of the lightning flashes that strike oil plant will be negative ground flashes, i.e. they
come from a predominantly negatively charged region of the thunder cloud. These include
high current and sometimes continuing current components and are depicted schematically
in Figure 2. This is the main type of lightning to be considered for protection of plant.
Sometimes positive ground strokes occur, though much less frequently, but their
characteristics are different. They are usually one long current pulse; often with as high an
amplitude as negative flashes, but the duration is so long that the heating and force effects
on conductors will be much larger. The positive flashes have a lower rate of change of
current than the negative flashes. Thus some aspects of the positive flashes are more severe,
and some aspects of negative flashes are more severe, but both types have a wide range of
parameters.
Since negative ground flashes are more common, lightning protection is usually
based around the characteristics of negative flashes, but with due allowance for where the
positive flashes are more severe.
Figure 2 gives an indication of the two most important components, the fast high
current pulse and the long slow, low amplitude continuing current. Table 1 gives a summary
of the parameter ranges for positive and negative flashes.

Phase 1:4
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

Figure 2 Schematic diagrams of the high current and the 'continuing current'
components of a lightning strike. Note that the amplitude and time scales of the
two waveforms are very different.

Peak current 5 kA to
200 kA (for a very
severe strike)
Fast high current pulse

Long slow ‘continuing’


current component

200 to 400 A
typical

Less than 100 µsec 200 to 800 millisec

Negative flashes can comprise one stroke only, or a quick succession of strokes. Positive
flashes normally occur as one continuous flash, with a high peak current and continuing
current.

Table 1 Lightning flash parameters

Percentage exceeding the tabulated parameters


95 % 50 % 5%
1st return stroke kA 7 33 90
Action integral 10 A s
6 2 0,006 0,055 0,55
Negative flash

di/dt kA µs 9 24 65
charge coulomb 1,3 7,5 40
Subsequent stroke kA 4,6 12 30
di/dt kA µs 10 40 160
No of strokes 1 3 12
Action integral 106 A2s 0,6 35 250
Positive
flash

di/dt kA µs 0,025 0,65 15


charge coulomb 20 80 350

Note: some of the parameters quoted here are more severe than are generally used for
lightning protection calculations. This is because they are for the much less common positive
flashes, so a weighted average is used to obtain the working parameters for protection
design. The top limit parameters usually employed are:
1. I = 200 kA
2. Action integral = 2.25 106A²
3. Total charge = 200 coulomb
4. di/dt = 140 kA/µs

Phase 1:5
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

1.4 LIGHTNING RISK AND PROTECTION PRINCIPLES

One important fact of the lightning strike to ground phenomenon is that unless ground
objects are very high (>150 m) those objects have no influence on the occurrence of lightning
in the area. The lightning leader on its way down from the cloud is not affected by ground
objects until it is very close to them, say 100 m or less. That is, lightning occurs because of
natural processes in the cloud which are virtually unaffected by the ground nature (unless it
is mountainous or has very high structures as above). Therefore, in calculating the risk, the
presence of the refinery, tank farm etc. makes no difference to the strike rate to that region,
so simplifying the risk calculations.
Furthermore, lightning conductors only have a very limited range of 'attractiveness'
(normally represented as a 45º right circular cone around the conductor, see Figure 3) and
although many attempts have been made to improve the attractiveness of lightning conductors,
none has been shown unequivocally to be more effective than the plain lightning conductor,
often called a franklin rod. For example, radioactive rods were once promoted as having extra
attractive properties, but the claimed 'super-attractiveness' has been discredited (it has been
shown by extensive comparative tests and by reference to the physics of radioactive emitters
that the devices have no influence on the path of lightning leaders). Radioactive rods have
been banned in many countries because the attractiveness is no better than a franklin rod,
and because of the potential for causing environmental pollution and health hazards.

Figure 3 'Cone of protection' principle for lightning protection. Objects within the
cone shaped volume beneath the dashed lines, all around the vertical pole are within
the 'cone of protection', and are at much lower risk from direct strike attachments.

45° 45°

Ground level

1.5 THE LIGHTNING CLOUD APPROACH, INDUCTION CHARGING, AND SPACE CHARGES

Normally under fine weather conditions there is a steady but weak vertical electric field at
the earth's surface. This field is quite weak and very steady. No actual potential differences
can be measured between grounded objects, except those which are electrically isolated
('floating' in electrical engineering terminology), because the current flow is zero and all

Phase 1:6
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

objects are grounded i.e. joined together. The charge is distributed all over the ground itself
and over grounded objects in a way depending on their shape. Horizontal flat surfaces have
a very uniform charge distribution (i.e. a similar surface charge density everywhere). The
highest surface charge occurs on thin pointed objects like church spires, tops of aerials, tops
of lightning rods etc. Where the surface charge is highest the local electric field is the highest
(and is the point from which an upward ground streamer could rise to meet the descending
leader, described in 1.7).
Thus in fine weather, a static condition prevails with a steady vertical electric field,
virtually no ground currents and a small distributed charge on the ground and on grounded
objects and structures. This gives way to a different situation on the approach of the storm
cloud (assumed negatively charged for this discussion, as is normally the case). The approach
of the cloud gives a slow movement of the ground charge and a change in sign. The ground
now acquires a much larger distributed positive charge and there is a larger negative
vertical electric field. The build up of this positive charge is called induction charging, and
is accompanied by a slow movement of ground charge i.e. a small ground current. This
small ground current produces the slow replacement of the original small negative charge
by the much larger positive charge. By this process a large area underneath a storm cloud
has a positive charge, which drifts along the ground surface underneath the cloud as the
cloud moves. As before, all earthed objects stay at the same electric potential, but as before,
elevated sharp objects will have a very high electric field on the top, but negative not positive.
The field may be high enough on some objects to give corona or brush discharges or the
well-known maritime phenomenon of St Elmo's fire. Corona, brush discharge and St Elmo's
fire are low current effects, not like lightning, and are associated with average currents of
microamperes to milliamperes. All sharp pointed upward facing items will tend to give a small
current into the air, like leaves on the tops of trees, aerials, and lightning rods. This will often
be a silent invisible discharge of only a microamp or two.
All such items are passing current into the air, which is normally a very good insulator,
and form a space charge in the air of positive polarity; the size and distribution of the space
charge depends on the number of current emitters, their height and especially the wind
conditions, because the space charge is attached to air molecules and moves with the air.
The presence of the space charge gives a screening effect over the emission points because
it reduces the electric field, and often limits the maximum field obtained at the emitters. This
process is called space charge screening (or shielding).
Sometimes the surface charge obtained by the induction process on tanks etc. is
called a 'bound' charge, but this is a misleading term and is a misnomer because the charge
is not 'bound' when it is on the surface of a conductor. It is free to move either slowly, as
the clouds drift over the earth, or quickly if a potential difference between the roof and the
shell occurs, for example through charge relaxation of the oil, resulting in sparking. The
closest any charge becomes to being 'bound' is the charge within the oil itself from frictional
charging as it passes through pipes, since its decay is determined by a time constant based
on the volumetric resistivity of the product and normally cannot discharge quickly, since the
resistivity is very high. Because the resistivity is so high, it plays no part in the conduction of
lightning current. The movement of the induction charge on the metallic roof is not 'bound'
and depends on the inductance of the roof and the arc at the sparking point, which will be
very low. (The term 'bound' charge will not be used subsequently in this document.)

Phase 1:7
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

1.6 ROOF SURFACE CHARGING OF TANKS AND BULK PRODUCT CHARGING

This process of induction charging and the surface charge caused by it on the top of tank
roofs should not be confused with bulk charging of the product by frictional effects as occur
when it is pumped through filters and pipes (see 1.5). This is a quite different effect but it will
be responsible for a small current into the shell, floor and roof of a tank as the charge relaxes.
However, both the induction charge process and the product charge decay need adequate
conductivity between the roof and the shell, which normally will be provided without the
need for an earthing cable between the tank shell, by the shunts, the ladder, and the roof
drain and so will maintain the roof at the same d.c. potential as the shell. As noted in
section 2.1(a), the bonding cable is required for lightning purposes, to short circuit the slow
current component. With certain reservations the shunts are usually adequate to carry the
fast lightning current pulses, and are the principal current route for the fast component on
and off the roof.
It is important to note that in a tank that has an adequately earthed roof, the lightning
cloud induction charging process does not cause static charging of the product. The tank
roof and shell form a 'Faraday cage' with no electric fields inside, and therefore there is no
charging mechanism. Thus lightning has no connection with bulk charging or discharging of
the product, as long as the floating roof is effectively earthed to the shell.

1.7 ATTACHMENT PROCESS TO TANKS AND GROUND OBJECTS

Prior to a lightning flash the electric conditions are usually fairly static, with the induction
charge on the ground, and on the shell and roof of tanks, with the highest vertical electric
field regions being the top of the shell, including the top of any hand rails, lights or other metal
objects on the top of the shell, and for a big tank, the tops of items on the roof such as the
support leg tubes. This will be particularly so for large tanks when they are above two thirds
full or so. This is an important consideration because in the subsequent attachment process
any of the items mentioned above could be locations for a lightning strike attachment.
In the usual scenario, the lightning process starts in the clouds, with a stepped leader
descending to earth. The stepped leader often exhibits branching on its path to ground as
the leader attempts to find the best route to ground. The path of the stepped leader is very
irregular because of random variations in the local air conditions and other factors. When
the stepped leader is within about 100 m or less from the tank (or ground), the electric
field at ground level rises sharply, and the electric field on the highest items becomes high
enough to launch an upward streamer towards the down-coming leader. In fact, two or
more streamers may rise almost simultaneously from ground objects (tanks, vents, trees etc.)
but only one usually is successful in making the connection to the downward leader. When
this connection is made the return stroke starts, the high current pulse which gives the bright
flash and the thunder, see Figure 4.
The first such stroke is called the 'first return stroke' and is usually the stroke with the
highest peak current (average of about 30 kA) and is the longest of the high current pulses,
although it is only up to a few tens of microseconds long. Sometimes this is followed by a
subsequent stroke (or strokes) of lower amplitude. These are called the fast components.
Between subsequent strokes, or immediately following the last stroke, 'continuing current'
may flow, referred to as the slow component. Most flashes, including all subsequent strokes
and continuing current, take no longer than one second in total, and often much less.
It is standard nomenclature to name the point at which the lightning arc connects
with the ground or structure as the 'attachment point'. Names such as the 'entry' point or
'impact' point are not adequate names because it is not useful to distinguish between an

Phase 1:8
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

'entry' point, or 'exit' point. The term 'impact' implies that something is thrown down from
the sky; all that does happen is that the air is ionised on the route of the current, but no
actual downward movement occurs except electrons which pass along the ionised path.
The return stroke current has a high amplitude, (5 kA to 200 kA), and is fast (a
few microseconds), and flows as an inductively dominated sheet of current over all objects
and the ground. Because of the high rates of change of current in the return stroke and
subsequent strokes, the high self-inductance (L) of the roof to shell earth cable (where fitted)
would produce a high voltage along it from the relation:

(2) V = L di/dt

For L equal to say 15 µH, and di/dt equal to say 100 kA/µs, V = 1.5 megavolts (momentarily),
and this voltage will ensure that the current will flow through the shunts with the potential
for sparking. A further point to note is that once a strike has occurred, to the shell, the
roof, or nearby, no type of lightning air termination (Franklin rod, radio-active, ESE, or anti-
lightning dissipation array) will affect the subsequent route of current.
From wherever the attachment point is, the current will flow as a sheet of current
over all conducting objects, up and down the shells over tanks, across the roofs, along the
pipes etc. as the current spreads out over a large area, discharging the surface charge.
Thus the role of the air termination system is over once the first return stroke path has
been established; the current path is dominated by the shape of objects, their conductivity,
and especially the inductance. This applies equally to nearby strikes (i.e. strikes within a few
tens of metres) as it does to strikes to the tank itself.

Figure 4 The process of attachment from a lightning cloud to a tank: a stepped


leader descends to close to the ground object (tank, building, etc). When it is close
an answering streamer rises from the top of the tank shell, or other salient items,
to connect to the down coming leader. When they meet the high current lightning
flash commences

The direction of movement of the leader is


from cloud to ground.

The point where the lightning strikes, the ‘attachment point’ is


not ‘chosen’ until the leader is within 100 to 150 m of the
object it finally attaches to. This is termed the striking distance.

When the leader approaches the tank or building, a


streamer rises to meet the leader. When they join, that is
the start of the first return stroke, the high current pulse.

Phase 1:9
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

2 THE INTERACTION OF LIGHTNING WITH STORAGE TANKS

2.1 FLOATING ROOF TANKS (FRT) AND LIGHTNING INDUCED FIRES

Well maintained FRTs are generally a very safe form of storage, but a survey by LASTFIRE has
shown that out of the 55 rim seal fires investigated for the survey, 52 of these were caused by
lightning[5]. It is prudent therefore to see if this predominant cause of fires can be eliminated
or at least substantially reduced.
The lightning hazards are as follows:

(a) Lightning current attachment to any part of the tank, its shell, instrumentation, on
the top (i.e. the dip tube) or to any part of the tank floating roof is a hazard. Strikes to the
ground nearby and nearby objects can also be a hazard. Wherever lightning attaches it will
cause a large current to flow over the tank shell, across the rim seal shunts and so cause
sparking (See Figures 5a, 5b and 5c). The parts of the tank most likely to be struck on a tank
are shown in Figure 6.
It is not true that lightning will take only the direct route to earth, (i.e. it will not only
follow a lightning conductor straight down to ground). It will instead share between paths
according to inductance which means, for a strike to the top of the shell, some will flow over
the outer skin of the shell to ground, some on the inner skin down to the level of the floating
roof, then across the seal on to the floating roof itself, across the roof, and on to the shell the
other side, up and over it down to the ground on the outside.
Note that current does not flow through the metal sheet of the shell, only on the skin
surface (See Annex A).This will lead to incendiary sparking as it crosses the seals between
the floating roof and the shell. Calculations performed for another similar safety review of a
floating roof tank in the UK showed that a voltage, of the order of 60 kV is initially available
to drive current through the shunts with the consequent risk of sparking. This mechanism
of incendiary sparking has not been described previously in the literature, except in an AEA
report[6] and the LASTFIRE lightning report[17], but is a very real possibility in open topped
floating roof tanks.
A strike to the floating roof itself (or to any projecting item on it) will be worst of
all, because all the current has to cross the seals. (The static bonding cable, if fitted, will not
prevent this effect as explained above, although it might limit the sparking to only that from
the fast return strike, but the slow continuing component might well be shunted by the
bonding cable.)

(b) At the moment of a nearby ground stroke the very large electric field existing as
the leader approaches will very rapidly reduce to almost zero causing the induced charge on
the shell and floating roof to flow away. As with the lightning current effect discussed at a),
because the discharge is very fast (approximately 10 Ωs or so), the bonding cable, which will
normally be several metres long, will not prevent the sparking between the floating roof and
the shell owing to the very large inductance of this cable. However there will be very much
less energy in such discharge currents than there would be if the tank was directly struck.
The oil product itself is not charged by this electric field mechanism, charge resides
only on the top surface of the metallic roof. Because the product is so resistive, it plays no
part in the current transfer process for lightning currents on the tank so for this purpose it
can be regarded as an insulator. Resistivity values are given in API 2003 Protection against
ignitions arising out of static, lightning, and stray currents, where it is shown that even the
most conductive black oils are 109 Ωcm compared with ≈10-5 Ωcm for steel, a difference of
14 orders of magnitude. Even if crude oil had water in it, bringing the resistivity down by

Phase 1:10
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

three orders of magnitude, it would make no difference to the principle that the oil is still so
resistive it functions as an insulator for lightning pulses.

(c) Indirect effects – i.e. currents induced or injected into cables and any part of the
electrical system by the flow of the main current in the tanks or pipes can cause sparking
in level and temperature gauges which might occur in regions of a flammable vapour air
mixture in the dip tube for example.

(d) St Elmo's Fire (corona, or 'brush' discharge) might be a problem if the tank is, say,
two thirds full or more, due to corona on vapour vents such as the roof support pipes, but is
unlikely to cause a problem with the roof below two thirds full height.

Figure 5a Current routes for flash to top of shell. Note that the fast high current
pulse flows down the inside of the shell and via the rim seals, and across the top of
the floating roof. (Only two routes are shown; in practice, current flows all over the
top of the roof and crosses the rim seal all around the perimeter of the roof).

Lightning flash

Current flows down


outside of shell.

Ground level

Figure 5a: Current routes for flash to top of shell. Note that the fast high current pulse flows down the
inside of the shell and via the rim seals, and across the top of the floating roof. (Only two routes are shown;
in practice, current flows all over the top of the roof and crosses the rim seal all around the perimeter of the
roof). Figure 5b Current routes for flash to floating roof. Note that the fast high current
pulse flows across the floating roof in all directions to the rim seals and shunts, and
then up and over the shell to ground. (Only if the roof is high is this a likely strike
point.)

Lightning flash to roof

Ground level

Phase 1:11
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

Figure 5c Current routes for flash to ground near a floating roof tank. The current
spreads all around from the strike attachment point, including to the tank, up and
over the tank and down the far side as shown by the typical current flow lines and
arrows. This current flow plan would only apply to the fast high current pulse. The
continuing current would flow along the ground and the tank floor only.

Lightning flash to
ground.

Ground level

(e) All of the effects described in a) to d) can occur, to a lesser or greater extent,
with nearby strikes i.e. those within 150 m radius. The effective range over which item c)
can be a problem is even larger, since electrical transients can travel a long distance along
instrumentation cables and pipes.

Figure 6 Open top floating roof tanks. Possible strike location points for a tank:
(a) anywhere around the top edge of the shell and (b) the floating roof (very low
probability).

a)

b)

45

Effect of 'cone of protection':


−− Only the top edge of the shell, shown in heavy red, and the region in red on the roof
can be attachment points for lightning strikes.
−− When the floating roof is high, the red area is larger, and conversely, when the roof is
low, the red area is reduced to a small zone in the centre of the roof. No other parts
of the tank are liable to take a direct attachment, including the rim seal region.

Phase 1:12
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

2.2 DISCUSSION OF LIGHTNING INTERACTION WITH OPEN TOP FRTs

Generally speaking, all of the hazards from lightning tend to be worse when the tank roof
is high, although in mitigation there is more chance that wind action will disperse more
efficiently any vapour that may be present. On the other hand, even with the roof low, roof
to shell sparking is still a potential problem, and a fire started low down in say a crude oil tank
may be made worse by exposed wax on the shell inner wall, and the accumulation of vapour,
and the enclosing effect of the shell.
A further point that affects the situation very strongly is the type of seal design and
its condition of maintenance. Numerous different seal designs are still used, dating back from
tanks built about 50 years ago to recent tanks. Some tanks have very good shunts across the
top seals, with a strong pressure on the tank shell, while others have far fewer shunts and
less pressure. It is not obvious a priori which design is the best. A further variable is the type
of petroleum substance in the tank, i.e. whether it is crude oil, which can deposit a waxy
layer on the tank shell; or gasoline, where the tank shell is clean but rusty. A further variable
is the height of the secondary seal, compared to the height of the foam dam (see Figure 10).
The foam dam presumably has a most unwelcome effect, which is to form a 'pool' of vapour
between the dam and the tank shell. If the secondary seal/shunt contact point is higher by
tens of centimetres than the dam, it is unlikely to be in a region of combustible vapour/air
mixture. Seal designs are discussed further in section 4.

2.3 LIGHTNING INTERACTION WITH OTHER TYPES OF ABOVE GROUND STORAGE


TANKS

The other types of storage tanks used for petroleum product storage comprise three main
types:
(i) Tanks similar to floating roof tanks but with a light roof of either aluminium or fibre
glass, e.g. the geodesic type roof, used for excluding water from the tank and for
minimising vapour loss. Also used if the floating roof has to be replaced by an open
tank since a much lighter construction floating roof may be employed. These are
one of the more vulnerable types of tank to lightning unless certain precautions are
taken.
(ii) Cone roof fixed roof tanks, with trusses and sometimes vertical supports for a
shallow cone shape roof. These are usually all steel, and, as with the tanks in 2.3(i)
are atmospheric tanks, i.e. they operate at the same pressure as the atmosphere,
either through permanently open vents, or sometimes with a pressure – vacuum
vent. These tanks are not designed to withstand a pressure difference between inside
and out. Sometimes these tanks have some type of floating roof (a 'floater'); the
internal floating roof is however usually a much lighter construction than the open
topped floating roof. As for (i) above, these tanks can be vulnerable to lightning
because of the possibility of a flammable vapour/air mixture inside.
(iii) Cylindrical or spherical pressurised tanks for LPG. These tanks are designed to store
liquefied gasses such as propane and butane under pressure, and are normally fully
welded (and gas-tight) steel tanks with no possibility of ingress of air into the tank.
The tanks in (i) and (ii) will often enclose a vapour/air mixture above the liquid, which,
under certain circumstances, might be flammable, but the LPG tanks being fully
sealed and pressurised will, under all normal circumstances, contain only the liquefied
gas. They are inherently the safest storage tank with respect to lightning risks, and
skin effect keeps the lightning current on the outside of the steel sphere, so the tank
contents are totally unaffected by lightning.

Phase 1:13
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

2.4 LIGHTNING INTERACTION WITH GEODESIC LIGHTWEIGHT ROOFS

There are two types to consider: (a) aluminium conducting roofs, and (b) fibre-glass or similar
non-conducting roofs.

2.4.1 Conducting e.g. aluminium roof tanks

This type of tank has several characteristic aspects that make it potentially vulnerable to
lightning. It does not have the problem of open topped FRTs arising from the current flow
over the floating roof and across the rim seal gap where the shunts are located. This is
because the induction charging and the discharge current paths will be over the top of the
conducting roof, and there will be negligible fields and currents inside. However the roof will
be liable to a direct attachment to it, with a possibility of burning through thin aluminium skin
panels, and so dropping molten aluminium and sparks into the internal volume, which could
contain a flammable vapour/air mixture. Furthermore an attachment to the roof or to the
top edge of the shell, or even to a pole fixed to the side of the tank as a lightning conductor,
could still produce a fire or explosion risk from sparking owing to current flow over the roof
structure where there are corroded or poor contact joints between the aluminium skin and
support structure.
The hole burning problem will be caused mainly by the long slow component of
current, because holes are not burnt in aluminium by the fast high amplitude pulse because
its duration is too short. Conversely the fast high amplitude current of the return strike is
responsible for the joint sparking because the pulse is so fast that inductive current sharing
and skin effect dominate the current flow streamlines (and not resistance) so that current
from the strike attachment point will flow all over the entire outside skin of the roof and
tank, including through corroded or resistive joints giving sparking. All welded construction
is to be preferred, so ensuring good joints, but that is not consistent with having a 'lift-
off', or bursting disc characteristic for the roof for pressure relief. Adding simple lightning
conductors (i.e. the usual narrow tapes) would make almost no difference to the problem
because of the inductive effect mentioned above which will drive current all over the entire
outside surface of the tank. The current would not confine itself to relatively thin conductors
when there is an adjacent large area of metal.

2.4.2 Non-conducting e.g. fibre glass roof tanks

Such a design (although infrequently met with) gives the potential problems of the open
top FRT, but with the added danger of the large volume above the internal floating roof.
This volume could contain a vapour/air mixture, which could not disperse as easily as for an
open top FRT. Strikes directly to the floating roof itself would be very unlikely, since it would
require the lightning arc to puncture the fibre glass. If this were to happen though, there
would be a severe risk of an explosion if the tank contained an vapour/air mixture. Ventilation
or inert blanketing of the volume above the floating roof would be desirable, unless a set of
lightning protection conductors were laid over the top of the fibre glass roof. Conductors
laid over the roof and bonded into a square mesh of something like 2 m² would go a long
way to reducing the current crossing the internal floating roof and hence crossing the seals.
The fast return stroke and restrikes would be the main lightning wave-form of concern. The
slow components are unlikely to be a problem because of the low resistance to ground from
the top of the shell.

Phase 1:14
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

2.5 CONE ROOF, OR 'FIXED ROOF' TANKS

The contents of fixed roof tanks are a major factor in safety because of the risk of forming
a flammable vapour/air mixture above the liquid. For products where this can happen there
is a risk of explosion, unless inert gas blanketing, better seals on internal floating roofs, or
ventilation prevent this. Most of the discussion above for aluminium roof tanks applies to fixed
roof tanks, although on account of their more substantial construction, the hazards for fixed
roof tanks are less. The ideal construction for lightning protection would be an all-welded or
very well riveted design so that there are no bad joints between the adjacent steel plates, and
so no sparking when the fast pulse of the return stroke passes over the tank. Many fixed roof
tanks conform well to this requirement, and so are generally very safe. The main details to
be taken care of are the vents, which should be situated where they cannot receive a direct
strike, and if the vents are not of the pressure–vacuum type (which are normally closed),
flame traps would be needed to avoid ignition of the tank via an attachment to the outlet
of the vent. However, flame traps are not without their problems, because they have been
known to become blocked, resulting in collapse of the tank. With a steel cone roof, there is
no tendency for current to flow inside the tank so the problems with open FRTs do not occur.
However, good house-keeping is essential e.g. not having loose bolts on inspection or access
covers or on the bolts securing the vents to the tank.
There is little or no advantage to be gained from installing the usual thin lightning
conductor tapes on fixed roof tanks, because as explained above, they would carry an
insignificant part of the current.

Figure 7 Typical fixed (cone) roof tank with vent at the top. The construction is
usually all steel at least 3/16' or 5 mm thick, and welded or riveted. This is generally
a lightning proof design. These tanks are usually much smaller in diameter than
floating roof tanks.

Pressure - vacuum
vent

Phase 1:15
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

2.6 PRESSURISED TANKS FOR LPG

This is the safest type of tank because the storage of the liquefied gas under pressure prevents
a vapour/air mix from forming in the tank. Also because the tanks are hermetically sealed all
welded steel spheres or cylinders, lightning current will stay on the outside as a very thin skin
effect current, and will flow to ground without interacting in any way with the tank contents.
Thus there is no possibility of incendiary sparking and no mechanism by which sparking could
affect the contents. Thus it is to be expected that these tanks would have a very good safety
record, which is borne out in practice.

Figure 8 LPG tank, all steel, pressurised, basically a very safe tank for lightning
because air cannot get into the tank and there are no sites equivalent to the rim
seal of a floating roof tank, where sparking could occur.

Filling, gauging etc area, should be


protected from a direct strike by an
Spherical pressure vessel holding
air termination
liquefied gas under pressure

Steel or reinforced concrete


supports

Figure 8: LPG tank, all steel, pressurised, basically a very safe tank for lightning because air cannot
get into the tank and there are no sites equivalent to the rim seal of a floating roof tank, where
sparking could occur.

Phase 1:16
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

3 LIGHTNING EFFECTS ON MATERIALS: ELECTRICAL HEATING


AND MECHANICAL DAMAGE; BURN THROUGH AND
HOT-SPOTS; THERMAL AND VOLTAGE SPARKING

3.1 EFFECTS DUE TO LIGHTNING CURRENT FLOW THROUGH METALLIC CONDUCTORS

Almost without exception, all of the metal structure and ancillaries of oil storage and
processing plant that are in locations where they might carry lightning current are so massive
that there is no risk either mechanically or thermally. That is, the mechanical forces caused
by the flow of current (electrodynamic forces) and the heating effect ('joule heating'), will
normally be negligible compared to the strength and thermal capacity of the steel. This
applies even to the usual type of tank rim seal shunts. For even a severe lightning strike, cross
sectional areas of only about 10 mm² of copper, 15 mm² of aluminium, or 20 mm² of steel
are required to carry the full current of a severe strike. Since the current disperses over many
different routes on its way to earth, only within a few centimetres of the attachment point is
there a high current density and action integral. Therefore there are exceedingly few places
where conducting material will be thin enough to be damaged by lightning current.

3.2 EFFECTS AT THE LIGHTNING ARC ATTACHMENT POINT

Where a lightning arc attaches to any part of the tank, a certain amount of erosion of the
metal will occur. For the high current pulse there is usually an insignificant amount of erosion
because of the short duration of the pulse. However the long duration, or continuing current,
lasts sufficiently long to burn through thin metal sheets or cause hot-spots on the back of
thicker metal sheets.
With the usual thickness of steel, (5 mm minimum) there is a negligible risk of lightning
being able to melt through steel sheet at the attachment point. The continuing current
component is able to melt through just over 1mm if the continuing current charge transfer
exceeds about 70 coulombs[7]. Only a very severe strike exceeds this. (For comparison, aircraft
are certificated to the 2 % severity level i.e. only 2 % of strikes exceed the level, which for
continuing current is 200 coulombs.) However it is not possible that 5 mm could be melted
through although the inside surface will become quite warm. As described in section 2.4.1,
burn through has important implications for geodesic dome tank roofs with thin aluminium
skins, and might be important for cone roof tanks, but generally not for open floating roof
tanks. Also, it is an experimental observation that galvanised steel sheet is very resistant to
burn through, even when as thin as 1 mm[7]. With arc attachment to steel sheets of 5 mm or
so thick there is no risk of sparks from the inside surface to other nearby items because the
resistive voltages produced by the continuing current are very small.

Phase 1:17
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

Figure 9 Effects at the arc attachment point. As well as the molten metal pool (which
may go right through the metal, leaving a hole) a hot-spot forms on the underside
of the metal plate. It is only the long slow continuing current component that causes
this type of damage.

Lightning arc attaching to Molten metal pool


metal plate

Zone of hot-spot on metal under surface

Figure 9: Effects at the arc attachment point. As well as the molten metal pool
(which may
A hazard for go
veryright through pipes
small diameter the metal, leaving
exists if they are atathe
hole)
top ofa process
hot spot forms vessels
or storage on the
underside of the metal plate. It is only the long slow continuing current component
or other exposed locations. If the wall of the tube is very thin it could suffer burn through
that causes
from this type
a lightning of damage.
attachment; such items should always have a lightning air termination above
them, which may be a 'natural' termination like a crane jib, hand rail, light support etc. Only
in very few cases will an actual lightning conductor be required. Thus, in summary, very
few locations will be at risk from burn through or any other significant attachment point
damage.

3.3 DANGER ARISING FROM SPARKING

Apart from aluminium geodesic roofs, sparking is probably the most likely cause of fires
from lightning in FRTs owing to the tendency for current from any strike on the tank or
closely nearby to drive current across the roof, via the shunts or via any other metal making
intentional or unintentional contact between the roof and the shell wall.
There are two types of spark to consider; thermal and voltage sparks:
(a) A thermal spark is defined as a minute piece of incandescent material which has
been ejected from some sparking site, usually a place where tens or hundreds of
amps or more are passing through a very poor joint, such as the contact point from
a shunt on to the inner shell wall of an open floating roof tank, or from a poorly
bolted flange joint etc. The white hot metal sparks falling from welding operation are
examples of thermal sparks, which are actually very small particles of metal, burning
as they fly through the air. Usually they are less effective as incendiary sources than
voltage sparks.
(b) A voltage spark is quite different. It occurs in a location with a small gap between
conducting items where the lightning creates a voltage large enough to cause
electrical break-down of the air or vapour/air mixture in the gap. A spark in a car
engine spark plug is an example of a voltage spark.

Phase 1:18
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

Voltage sparks with an energy above 0,2 mJ could be sufficient to ignite product vapour/air
mixtures if they are slightly richer than the stoichometric mixture. For mixtures away from
the optimum, the energy requirement is higher, but even so the energy in lightning induced
sparks is likely to be many orders of magnitude higher than this; that is, lightning induced
arcs are very powerful ignition sources.
Both types of spark are very incendiary as shown in work at Culham as part of their
research into aircraft fuel system lightning hazards; in particular to examine and understand
thermal and voltage sparking[8, 9, 10, 11].
A poorly contacting rim seal shunt on an open FRT is an example of where sparks
might occur under lightning conditions. The small contact area and, in practice, the presence
of surface treatments or contaminants are conducive to sparking. If there is an oxide layer on
the shell or the shunt, the spark will initially be a voltage spark to break down the insulation,
followed by current flow in a poorly contacting area giving a flurry of thermal sparks.
Normally, non-sparking joints for high current purposes need large areas of contact with clean
surfaces and high contact pressure. The problem is potentially worsened by the fact that the
electrodynamic force arising from current flow between the shell and the shunt will be in a
direction so as to push the shunt away from the tank wall, so making the sparking worse,
although in practice the force will be very small unless the roof is very high and the strike
occurs to the shell very close to the position of the shunt. It should be noted that because
the shunts are normally within the cone of protection of the shell, there is a negligible risk
that a strike will ever occur directly to the seal region or to the shunts themselves. Thus
the principal sparking problem should be considered as a secondary effect of lightning, i.e.
due to current flow from an attachment point which is at least some little distance away.
Although previous documents have said that lightning might damage the shunts, the more
recent work based on lightning testing shows that while no physical damage is likely to
occur, sparking is nonetheless a real hazard[17].

Phase 1:19
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

4 FLOATING ROOF RIM SEALS AND THE LIGHTNING INDUCED


FIRE PROBLEM

4.1 INTRODUCTION TO SPARKING AT SEALS

The detailed design of seals including (a) the primary, and (b) secondary seals (where fitted), and
(c) the location and design of the shunts (where fitted) and maintenance, play an overriding
part in whether incendiary sparking and fires might occur at the seals of open topped FRTs.
As described in section 2, a lightning strike anywhere in the vicinity will drive a current across
the surface of the roof and will flow on and off the roof via the shunts or other contact
points. Where the shell is either corroded or contaminated with non-conducting material
e.g. wax etc. or where there are no shunts, there will be sufficient inductive voltage to cause
break-down of whatever gap there is in the form of a voltage spark, and the consequent
current flow arcing may give thermal sparks also. Thus incendiary sparking is inevitable and
the possibility of a fire will then depend on the location of the sparks, and whether vapour
leaks create a flammable vapour/air mixture in the region of the sparks. Thermal sparks for
this situation are more of a problem than voltage sparks, because the voltage spark (or arc) is
localised at the shunt. Thermal sparks on the other hand spray out and fall from the sparking
site and so they can spread their effect over a larger volume, especially below the source of
the sparks. A discussion on the current required to produce such sparks is given in 4.3.3.

4.2 FLOATING ROOF SEAL TYPES

4.2.1 Primary seals

These can be divided up into two main classifications, liquid (or product) mounted (i.e. the
seal has the stored product as a liquid in contact with it), or vapour mounted seals (i.e. where
the seal is above the liquid surface, having a vapour or vapour/air space between the liquid
and the seal). The three main types of primary seal are metallic shoe type mechanical seals,
liquid filled seals and foam seals. The metallic shoe type is the oldest, and uses a metal shoe
system to slide against the shell, whereas the other types use (normally) a rubberised material
to contain the liquid or foam, so that the rubberised material rubs against the shell. Whereas
the metallic shoe system provides some sort of electrical contact to the shell from the roof,
the other two are in principle insulating or at least very resistive, and might allow static
charge to leak away, but would have no effect on lightning. The electrical contact between
the shell and the roof given by the metallic shoe seal is probably a big disadvantage, since it
could give sparking in a place where there is a possibility of a vapour/air mixture, very close to
the liquid, or in the volume between the primary and secondary seals. The use of insulators
for the mechanical shoe system will prevent sparking at these locations as long as there are
shunts or other conducting paths from roof to shell as alternative safe current routes. Thus
such insulators must be used in conjunction with shunts to provide the safe conduction
path.

Phase 1:20
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

4.2.2 Secondary seals

These seals, as the name implies, are a secondary protection against vapour leaks, and
are always mounted above the primary seals, usually in place of the weather shields that
have hitherto been used on some floating roof tanks. In some countries there are statutory
requirements to fit secondary seals for storage of refined product with a relatively high vapour
pressure, to minimise emissions. Secondary seals normally take the form of a flexible wiper
mounted on the rim and pressing against the shell, or the 'compression plate' type secondary
seals. The compression plates are usually stainless steel at an angle of approximately 35º
to the vertical and holding a wiper against the shell, and sealed by a rubber diaphragm all
around the entire circumference of the tank. These usually have shunts fixed to the secondary
seal so that they contact above the top of the secondary seal where it presses against the
shell.
There should never normally be liquid in the space between the primary and secondary
seals, but of course, the better the secondary seal works, the more likely it will be that vapour
accumulates in the space between them.

Figure 10 Typical floating roof pantograph type primary seal, with secondary seal,
shunt and foam dam. The shunt is intended to provide electrical continuity from the
roof to the shell.

Shunt, stainless steel, usually fitted to the top of


the secondary seal

Secondary seal
Foam dam

Floating roof
Primary seal

Weight to hold shoe


against the shell

Shell

Figure 10: Typical floating roof pantograph type primary seal, with secondary seal, shunt and
foam dam. The shunt is intended to provide electrical continuity from the roof to the shell.
One of the main problems in seal design is the need to accommodate mechanical imperfections
of the tank shell, such as ovality, distortion, effects of settlement of the foundations, and
raised items on the shell wall such as rivets or imperfect welds. Thus there are many factors
which can make it difficult to ensure 100 % contact of the seals, both primary and secondary,
and if mechanical failure is included too, it shows that to minimise losses, and hence to reduce
the fire risk, frequent inspection and good maintenance would seem to be essential. One of
the main reasons for poor contact between the shell and shunts is out-of-roundness of the
shell, often caused by subsidence of the tank shell. Where this opens up gaps between the

Phase 1:21
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

secondary seal and the shell, it will often prevent contact of the shunts locally. Gaps between
the secondary seals and the shell will allow sparks to fall into the vapour area between the
seals from the nearest shunts which are still in contact.

4.2.3 Shunts

These have been fitted for many years to minimise the electrical problems with tanks and
often take the form of a spring contact strip, made from stainless steel, fastened to the top
of the weather-shield or secondary seal that presses against the shell. They are normally
spaced ≈3 m (10') apart. As described earlier, these are particularly important for lightning,
since although static charging effects should be taken care of by the flexible bonding cable,
this lead has a very high impedance for fast lightning pulses, and does not carry much of the
current in a strike, apart from the slow component. The shunts, which will have a far lower
impedance, will provide the fast pulse lightning current route on and off the roof. However,
depending on the absolute current amplitude through each shunt, the physical condition
of the contact area and the force with which the shunt presses on the shell wall, it may or
may not produce incendiary voltage and/or thermal sparks (see 4.3.3). Also depending on its
position relative to the seals and the condition of seal maintenance, sparks produced by the
shunt might (or might not) be within, or fall into, a region of flammable mixture.

4.3 DISCUSSION OF SPARKING SOURCES ON SEALS AND SHUNTS

4.3.1 Sparking on metallic shoe type seals

These seals are normally classified as vapour mounted, although in the usual design, the
metal shoe is at least partly immersed in the liquid. However, the rim vapour space (i.e. the
space between the roof and the shoe) is often vented to air, so allowing direct evaporation
from the liquid surface into this volume in which air can freely circulate. Thus with certain
stored products or crude, it is possible the rim space contains a flammable mixture which
could be ignited by a spark. As referred to in 4.2.1, some mechanical shoe type seals are
equipped with insulators to prevent current flow through the mechanism and so prevent
sparking. If they are not so equipped this is a severe risk area, and even with insulators there
is still a severe risk unless there is a good low impedance alternative path for lightning current
to prevent flash-over of the insulators, such as well installed and maintained shunts. Where
insulators have not been fitted, there are innumerable possible sparking positions: (a) where
the shoe makes poor contact with the shell wall, (b) at joints between the shoe sections, and
(c) at any of the linkage joints of the pantograph and weight hangers which are within the
rim vapour space. There are a large number of such joints in a big storage tank, see Figure
10. (Note sparking within the liquid is not incendiary because of the absence of air.) The
problem of sparking, as explained above, is probably worst when the roof is high and when
a strike occurs to the shell top or the roof itself, so giving the highest currents in those shoes
or shunts close to the lightning attachment point. As stated previously, it is extremely unlikely
that a strike could occur directly to the seals, shunts or the rim vent outlet, even with the roof
very high because of the cone of protection principle. Also, since the rim vent will normally
be very close to the shell, it is very unlikely that the electric field could be high enough at the
exit of the vent to give corona or streamering which otherwise might have caused ignition.
Thus in summary, the principal fire problem is sparking at:
(a) Joints in the pantograph and the shoe if there are no insulators.
(b) Joints in the pantograph and the shoes if there are insulators but no shunts or any
other distributed low impedance connections from the roof to the shell wall.
There is also a problem with sparking at the shunts themselves.

Phase 1:22
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

4.3.2 Sparking with liquid and foam filled primary seals

These seals are in effect insulating as far as lightning currents are concerned (although some
manufacturers offer static discharging material for the seals which has no effect on lightning).
Only if there is a bump device or similar incorporating conducting material or any other
aspect of the primary seal which is highly conducting would it be possible to have sparking.
Thus, putting other considerations aside, there should not be sparking at the primary seals
themselves induced by lightning, but they still require conducting weather shields or secondary
seals with shunts to give the necessary electrical contact to the shell. The whole roof cannot
be left connected by only the earthing lead to the shell because of the risk of a direct stroke
to the roof which would flash-over seals even if there were gaps of several inches. Such a
flash-over would produce a very hot and dangerous incendiary arc. Generally, lightning is not
a problem with liquid and foam filled seals, as long as the seal region is free from inadvertent
metallic conducting paths, but they will need shunts to provide low impedance connections
from roof to shell, and for these shunts to be placed above the secondary seals. Experimental
data reported in other documents in this programme show that good electrical contact at
shunts is in practice difficult to achieve, and that sparking is likely to occur.

4.3.3 Sparking at shunts and weather shields

Shunts are designed to provide electrical contact between the shell and the roof for lightning
and static discharge purposes. (The roof bonding cable is not intended for lightning, but it
provides a reliable low resistance connection between shell and roof and could well prevent
sparking from the slow component, as well as conducting static.) Shunts are spaced about 3 m
apart around the seal. As discussed in section 3.3, they do not have the required qualities
to be non-sparking contacts, since the contact area is not clean enough or big enough or
under sufficient pressure. Determining the threshold level at which sparking occurs can only
be effected experimentally, especially when the range of surface conditions is taken into
account, including the use of surface treatments to the shell walls and the use of electro-
conductive paints[14]. In work on aircraft fuel systems, sparking occurs in fuel pipe joints,
which have poor contact, at currents as low as 1 kA or so, and unless the roof is very low the
current will easily exceed this value. With the roof high, a strike to the shell itself of average
intensity might be in the order of 30 kA, and the two nearest shunts might be required to
carry up to 40 % of the current, about 6 kA each. This is probably well above the sparking
threshold for a typical shunt/shell contact. Thus the use of shunts is not a complete cure for
the sparking problem, unless methods can be devised to improve the contact conditions so
that the threshold for sparking is raised considerably, or the shunt is positioned where sparks
do not matter.
(The experimental work on shunts, and discussion on shunts' susceptibility to
lightning, which was mostly carried out subsequent to this review, is given in EI-EN2-04
Lightning tests to tank shell/shunt samples.)
There are two alternative approaches to the problem, which are concerned with
the location of the shunt (and the sparking) with respect to the zone containing flammable
vapours. If the shunt contact is located above the top of the secondary seal, given reasonable
maintenance of the seals, the sparking will be above the vapour/air zone and as long as there
are no other accidental contact points within the zone, any sparking will occur in a region of
a very weak mixture, and any thermal sparks occurring should be prevented from falling into
the critical zone by the wiper at the top of the secondary seal (see Figure 10 for illustration of
this principle). A complication with this scheme is the height of the foam dam. Under normal
conditions the dam will tend to accumulate within it any vapour leaking from the seal region.
Thus it could be imagined that under fairly still air conditions, and with the roof low, and

Phase 1:23
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

with a leaking seal system, that a flammable vapour/air mix could accumulate in the region
between the foam dam and the seal. Thus to avoid ignition of this it is desirable to have the
shunt contact point as high as, or higher, than the top of the foam dam. This may not be
practicable on most tanks because of constraints on the foam dam height with respect to the
top of the secondary seal for fire fighting reasons. Also the natural dispersion of the vapour
might make this precaution unnecessary, (for example, by the drainage holes at the bottom
of the foam dam), so that in practice there may be no need to have the shunt contact point
so high. This is a point for discussion and further review.
The alternative method referred to above is to arrange for the roof-to-shell electrical
contact to be always immersed in the stored liquid. With this location, any sparks that occur
are within the liquid and do not have the necessary air available. This has some disadvantages
such as maintenance inspections and repairs and the possible interference with the mechanical
shoe system, and wax scraper etc, but there are important advantages in having current
flow in a position where there is guaranteed to be no air. This contact system should be
compatible with mechanical shoe seals, just as well as with liquid or foam seals, but the
voltage available to cause sparking on the insulators of the mechanical shoe system would
need to be investigated theoretically and experimentally. (It might be possible to incorporate
the contact system into a wax scraper as long as it is always immersed in liquid.)
Owing to the diversity of types of weather shields it is difficult to make general
comments, but any type which corresponds electrically to a secondary seal with a contact to
the shell above its rubbing edge at the top will have a similar general characteristic. That is,
sparks which occur at the contact points will only be prevented from causing ignitions if they
are outside the interspace, and are prevented from falling into the interspace.

4.4 LIGHTNING RISKS TO OPEN FRTs FROM SURFACE PRODUCT AND LEAKS INTO
PONTOONS

4.4.1 Lightning risk from product on the top surface of open floating roof and roofed
over tanks

Where liquid product is on the top surface of the floating roof the lightning hazard depends
almost entirely on the exposure of the product and its vapour to direct ignition by the flash
itself, or from corona, upward streamers, etc. Thus, the height of the roof will play a large part
in determining the actual risk of ignition, since the area for a possible lightning attachment
is much smaller when the roof is low than when it is high, and the electric field and corona
effects are less when the roof is low. (See Figure 6, where it can be seen that a large annular
area of the roof will not normally be struck with the roof low. Note that any type of lightning
protection system which aims at inhibiting strikes to the top of the shell runs the risk of
diverting the strike to the floating roof itself which makes the problem worse.)
Any type of 'roofing over' will achieve the objective of eliminating direct attachments
to any part of the floating roof and so eliminating ignition directly from the lightning arc.
Roofing which is non-conducting as defined previously will inhibit direct arc ignition only,
but not ignition from corona etc, whereas a conducting roof will protect the tank completely
except as discussed in section 5 concerning the build-up of flammable vapours arising from
the exposed surface product and the consequent explosion risks from sparking in the roof
structure.
Although much less likely, there are mechanisms whereby the surface current flowing
over the floating roof surface (from a strike to the shell for example) could cause ignition
from sparks at loose metallic items on the roof, debris, chains, the rollers of the steps onto
the roof etc, and this is an example of where good maintenance and inspection procedures

Phase 1:24
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

will minimise the amount of such metal on the roof. Effective operation of the shunts will,
of course, minimise the current flow between the roof and the steps. A possible problem
arises from loose contact between the bonding cable and the roof skin which should be
minimised by sheathing the cable over its entire length so that apart from the bolted ends,
no inadvertent contact can occur.
Generally the current from a lightning strike will not cause an ignition when the
current flows along a metal surface covered in product because the current density will
be, for the most part, much too low, and the conductivity of the steel too high to cause
significant temperature rise, but loose conducting objects in contact with that surface will
conduct and might spark, but such sparks are only likely to be incendiary if they are just
above the liquid surface. This mechanism could ignite a heavy vapour/air mixture lying on the
tank roof even with no actual liquid present.

4.4.2 Product within floating roof pontoons and the lightning risk

As discussed previously, any vapour/air mixture is not at risk when enclosed in a welded gas
tight structure. This forms a 'Faraday cage', or more accurately a 'Maxwell shield' around the
mixture, so eliminating all sparking from electric or magnetic fields or current flow. However,
the pontoons will normally have access hatches, and sometimes vents on them, or in some
designs the entire top of the pontoon is detachable, which could completely nullify the
inherent protection of a sealed volume. The worst risk is probably a direct strike to an unbolted
or poorly bolted pontoon or hatch cover, or to a vent, giving sparking which could ignite an
explosion in the pontoon. To minimise this risk, the hatches and vents should be close to
the shell wall, (so they are always within the 45 º cone of protection even with the roof at
the top), or be close to alternative air terminations (like roof support legs) which prevent
attachments directly to the hatches or vents. Where possible, all joints in hatches, vents
etc. should be welded preferably, and not bolted, since welding gives a more satisfactory
electrical connection that prevents sparking. As with exposed oil on the top surface of
the floating roof, maintenance procedures should be in force which ensure that such an
eventuality occurs for only a very small percentage of the life-time of the tank. All metal,
removable pontoon covers are very unsafe for conducted or arc attachment effects because
sparking is likely. Non-conducting covers would be safer for conducted current because there
will be no sparking joints.

Phase 1:25
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

5 LIGHTNING IGNITIONS ON ROOFED OVER FLOATING ROOF


TANKS

5.1 INTRODUCTION TO ROOFING OF OPEN TOP TANKS

Where open topped tanks are equipped with roofs e.g. geodesic domes, the lightning risk
can become very different, and depends on whether the roof is mainly conducting or non-
conducting. A conducting roof provides a 'Maxwell shield' over the tank, which minimises
current flow within the tank, including the prevention of current flow over the internal
floating roof, often referred to as a 'floater'. Thus the seal problem is very much reduced but
sparking in the roof structure or a burn through of thin skin material gives a risk of explosion
should it be possible for a flammable mixture to occur in the space above the floater.
An insulating roof structure is very different, and would still be subject to the problem
of sparking at the rim seal and have an explosion problem depending in detail on how the
fixed roof was constructed, and as before, whether a flammable mixture could occur in the
space above the floater. These points are discussed in more detail in 5.2.

5.2 LIGHTNING HAZARDS WITH ROOFED OVER TANKS

5.2.1 Conducting roofs

For tanks with a conducting roof (which must have multiple bonds to the shell to qualify as a
conducting roof), very little lightning current will flow inside the shell and across the floater.
Thus a major source of ignition, that of sparking at the rim seal area, is removed. This is
because the induction charge built up by the storm cloud is confined to the top outer surface
of the fixed conducting roof and to the outside of the shell, so the current from a strike
discharges this charge and in so doing flows over the outside i.e. the top surface of the roof.
This current flow process on the external fixed roof is also assisted by skin effect in the metal
and inductive current flow considerations. Thus there is little or no current within the tank.
The safety of the tank mainly depends on the detailed design of the fixed roof and whether
the leak rate past the seal is low enough, or the ventilation rate is high enough to prevent a
flammable mixture building up above the floater. If a flammable mixture existed in the tank,
the fixed roof could be vulnerable to lightning in several ways:
(a) A direct attachment to a thin roof skin (e.g. 1 mm aluminium) giving burn through
and/or molten metal or sparks dropping into the lower part of the space between the
fixed roof and the floater.
(b) A direct attachment to a vent from the roof which either has no flame trap, or if it
does, has a poor joint beneath the flame trap which sparks.
(c) Corona or streamers from a vent with no flame trap.
(d) Current flow across poor joints between metal components of the roof, resulting
from an attachment or a nearby strike, so producing voltage or thermal sparks.

Thus a direct strike to the tank has the potential for producing a fire or explosion by all the
mechanisms (a) to (d), but a nearby strike could only produce an explosion by mechanisms
(c) and (d).
Thus to produce the worst event requires a strike directly to the fixed roof itself, and
is thus a lower probability event than the corresponding 'worst case' event with an open
top floating roof tank, where a strike to nearby the tank or to the tank itself could produce

Phase 1:26
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

incendiary sparking. As discussed in section 1.2, the risk depends on the effective area for
collection of strikes which could affect the tank, which for an open roof tank will be larger
than a roofed over tank. With suitable design and construction techniques, a roofed over
floating roof tank (with a metal roof) could be much safer than an open top floating roof,
as long as sparking and burn-through of the roof were prevented, or precautions made
to ensure that there is always a non-flammable mixture in the vapour/air space above the
floating roof.
Prevention of sparking in vent pipes requires that the vent is within the 45 º cone of
protection of a lightning rod, which must not be fixed to the vent in any way except with a
non-conducting stay. This precaution prevents current flow in the vent pipe below the flame
trap.

5.2.2 Non-conducting roofs

Roofs incorporating large amounts of non-conducting material e.g. fibre glass are difficult
to make generalisations about because of the possible use of some structural metal and
whether it is interconnected or not. Where small quantities of metal are used and they are
not interconnected, the problem with seal sparking will be practically unaltered from that
described in section 4, because the metal will not form a 'Faraday cage' or 'Maxwell shield',
and will not alter the tendency for current to flow over the floating roof from a strike to
the top edge of the shell or nearby, nor will it prevent the induction charge process from
occurring on the floating roof. To prevent this, and to prevent the current flow over the
floating roof would require an interconnected metallic system in the roof with a mesh size no
bigger than approximately 1/20 of the tank diameter, and preferably finer, depending on its
height above the floating roof when the tank is full. A geodesic roof with a metal framework
and glass fibre panels might meet this criterion. Where the metal is not interconnected, not
only does the tank have the potential for rim seal sparking, but also the problems described
in 5.2.1 associated with sparking in the fixed roof, depending on where and how metal is
used in its construction.
It is important as referred to previously that adequate precautions are taken to
ensure that a flammable mixture cannot occur in the tank. A further risk for structures with
small quantities of metal might be structural failure of fibre glass (resulting in the possibility
of roof collapse) if lightning currents cause explosive arcing within it, e.g. between non-
interconnected metal components or in fine wires used for X-raying the lay of the fibres.
Where the roof is entirely insulating or has insufficient metal to form a conductor for
lightning over the top, it might be necessary to provide metallic lightning conductors over the
top of the tank, and bonded to the top of the shell, to prevent lightning damage to the roof.
The design of lightning conductor systems is described in BS6651:1999[4], NFPA 780[12], DIN
and IEC standards, where design data for fixed and catenary systems are given.

Phase 1:27
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

6 LIGHTNING HAZARDS TO CONE ROOF (FIXED ROOF) TANKS

Along with open floating roof tanks, these are the other most common type of tank
encountered in bulk storage petroleum tank farms. They are normally all steel, having a steel
shell generally similar to a floating roof tank, and an almost flat conical shaped roof of steel,
supported by internal trusses, and occasionally with upright support stanchions too (see
Figure 7). These tank roofs are usually a much heavier and more substantial construction,
usually welded or riveted, than the type of tank discussed in section 5.2. The top is usually
very smooth and free from hardware, the most usual items being vents of various types and
access hatches and instrumentation items. Thus there are very few items to act as lightning
air terminations for a lightning attachment to the tank; the vents, hatches, and instrument
heads being the only ones.
Thus it would be more possible for this type of tank for a strike to the actual roof
skin, as distinct from the very low possibility of a strike to the skin of an open floating roof
tank. So for this type of tank in particular, knowledge of the burn through and hot-spot
characteristics and skin thicknesses usually encountered are necessary. These are usually steel,
which has good burn through resistance and its thickness is usually sufficient (5 mm or 3/16'),
but the hazard arises from the fact that the large volume above the liquid and below the
conical roof may contain a flammable vapour air mixture. This potentially is a lightning risk,
although with the usual thicknesses of steel, neither complete burn through nor a hot-
spot hot enough and lasting long enough to cause an ignition is likely, but these factors will
be tested as part of the experimental programme. Tank roof thinning due to corrosion is a
possible cause of problems, and it is possible that the corrosion could bring down the skin
thickness from a minimum of 5 mm to much less, so making hot-spots or burn through more
possible. However, against that possibility, it is unlikely that a strike will occur exactly to a very
thinned area, and in any case the attachment is more likely to occur to slight projections like
raised rivet heads or the edges of overlapped steel sheets. For a smooth top tank of welded
construction this might not be the case, and attention should be paid to roof corrosion (no
part of the tank skin other than the roof is ever likely to take a direct attachment).
A further possible cause of ignition is from sparking associated with the fast high
current return stroke of lightning. This is a similar problem to that described for geodesic
roofs in section 5.2.1, and is associated with sparking at poor joints between adjacent metal
parts forming the conducting path. With the fast high current lightning impulse, current tries
to flow as a sheet of current all over the outside skin of the tank and inductive voltages are
large and drive current through any bad joints. However in practice, this is likely to be a much
lower probability of an ignition source than with geodesic roofs, because the typical cone
roof tank is much more substantially built, with solid welded or riveted joints in the cone skin
(the shell, of course must be welded or riveted to be liquid tight, and so is unlikely to be a
sparking hazard), although the detail of the cone roof fastening to the shell might not be so
spark free depending on the method of construction.
That leaves the vents, breathers, hatches and instrumentation as the only other
sources of ignition. Continuously open vents which are potential attachment points for
lightning and where the lightning could attach very close to the open vent (from which a
vapour/air mixture is venting) should be equipped with a flame trap. The flame trap should
preferably have welded joints between it and the tank to avoid the risk of sparking in the pipe
between the tank and the flame trap. If bolted joints are used, they must be solidly bolted
to make a very sound current path to the cone roof without sparking. Standard pressure –
vacuum breather vents are more satisfactory in that they are likely to be closed for most of
the time and so less likely to allow a vapour/air ignition to propagate into the tank. This type
is generally approved by the industry for use without flame-traps (see review of NFPA 30

Phase 1:28
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

Flammable and combustible liquids code in Annex B). Loose hatches could also be a hazard
with sparking at the loose bolts. Generally, non-conducting hatches pose less of a problem
for this, because there is no tendency for current to pass through a poor joint.
Sparking at instrumentation is a general problem with all types of floating roof and
atmospheric tanks. The major factor for safety of both cone roof and geodesic roof tanks is
the type of petroleum product stored in the tank, whether it has been 'spiked', and whether
any system is used in the tank to reduce the possibility of a flammable vapour/air mixture.
Thus tanks where the liquid volatility is in the range of producing a flammable vapour/air
mixture, and which don't have an internal floating roof, or any inerting method, are most
at risk.

Phase 1:29
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

7 LIGHTNING HAZARDS TO LPG TANKS

Owing to the all welded gas tight construction of these tanks, and the absence of air or
oxygen inside the tank, these tanks are the most safe against lightning effects. Lightning
current flow in the tank will be as a thin skin effect dominated layer on the outside skin only,
thus no current will interact with internal parts of the tank or its contents and there is no
significant risk. However, it is possible for damage to occur to instrumentation if it is exposed
to a direct strike at the top of the tank. If lamp support poles on the tank do not fulfil the
function of an air termination over all the exposed pipes and instrumentation at the top of
the tank, a pole should be installed to give 45 º cone of protection over the instrumentation.
Hand rails and all other metal support structure should be assessed for their ability to provide
cone of protection coverage.
Overall, LPG tanks, both the spherical or cylindrical types, are the safest type of tank
against the hazards of lightning.
To the authors' knowledge, this type of tank has never been involved in a lightning
induced fire or explosion incident, and the reasons for this are very clear owing to its
construction which keeps lightning current out of the petroleum liquid/gas volume, and
equally important, maintains an oxygen free atmosphere in the tank.

Phase 1:30
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

8 SUMMARY OF LIGHTNING EFFECTS ON ABOVE GROUND


STORAGE TANKS

8.1 THE LIGHTNING PROCESS

8.1.1 The lightning strike rate to ground has a strong geographical dependence, being highest in
the tropics and lowest in cold or dry regions.

8.1.2 Lightning intensity, i.e. the magnitude of strikes is very variable (5 kA to over 100 kA) and
does not have a correlation with the strike rate to ground, but there is a correlation with the
type of strike, whether positive or negative. Positive strikes are generally more severe, but are
uncommon except in particular areas of the world such as Sweden and Japan.

8.1.3 Unless very tall, such as sky-scrapers, masts and towers etc. or mountains, the presence of
ground objects has no effect on the strike rate to ground. Strikes are initiated in the cloud
and progress to ground by a leader. The path is essentially random.

8.1.4 Metallic structures, which usually include all above ground storage tanks and refinery/
processing equipment, usually function as their own air terminations, but sometimes special
lightning conductors are installed. These are often called Franklin rods, or finials. The metallic
structures also act as their own down conductors, so that lightning protection requires only
the addition of adequate earthing by means of earth rods or connections to the site earth
system.

8.1.5 As a part of the whole lightning phenomenon, induction charging of the ground underneath
the lightning cloud occurs; this is a slow process. At the moment of a lightning flash this
charge is fully or partially discharged in a high amplitude fast current pulse, giving a surface
current over all grounded objects, including over open topped floating roofs.

8.1.6 The lightning current from its attachment point on the tank or structure does not travel
straight down to ground. The route it follows is usually the one of least inductance, resulting
in the current flowing over the whole surface of structures as it spreads out radially from
the attachment point and finds a route to earth discharging the induction charge. During
this fast discharge process, the high impedance of the earthing cable causes the current to
discharge via the rim seals, through the shunts or the mechanical shoe seal or through the
secondary seal. The product is too resistive by many orders of magnitude to play a part in the
lightning current conduction process.

8.1.7 The zone of protection concept suggests that with the tank roof low there is a low probability
of a strike to the roof. The most likely attachment points would be around the top of the
shell. With the roof high, a strike could occur to the roof but would normally attach to the
tops of the support legs or other stress raisers. Thus it is extremely unlikely that a strike would
occur to the rim seal area or the shunts directly, or to the inside or outside wall of the shell
below the top, or to the flat decking of the roof.

Phase 1:31
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

8.2 EFFECTS ON MATERIALS

8.2.1 Because storage tanks are steel structures, and good conductors, the thermal effects of
current flow through the steel are normally insignificant. Most metals are able to carry
lightning currents very safely, since only very small cross section areas of steel, aluminium and
copper are required to carry the entire lightning flash currents. These are 20 mm², 15 mm² and
10 mm² respectively. Thus shunts etc. will not be damaged by lightning to the tank, either
directly by the current or from force effects. The sparking at the contact point will result in
negligible erosion of the shunt material.

8.2.2 The lightning attachment point is the only place where any obvious damage to steel might be
seen, but it is believed that steel as thin as 2 mm may be sufficient to prevent burn through
with a severe strike. 5 mm is certainly satisfactory, although experimental confirmation is
being obtained of this and of the temperature reached on the back surface (the hot-spot) of
the steel sheet.

8.2.3 It is possible that thin small bore pipes (for example on the tops of storage and process
vessels) might be damaged by a direct attachment to them. Such pipes and instrumentation
on the tops of tanks should not be exposed to direct strike attachment. Some simple form
of air termination above them is needed to prevent a direct strike. A similar principle should
apply to equipment on the tank gauger's platform; it should be within the 45 º cone of
protection of a lightning conductor, which could be a metal lamp post, hand rails or a simple
steel pole.

8.2.4 Lightning is a high voltage and a high current phenomenon, so high voltages are available to
cause flash-over or high voltage breakdown of insulators. Thus any scheme to protect against
lightning using insulators must allow for the very high voltages, and provide sufficiently
large insulators, or an alternative safe path for the current which limits the voltage stressing
the insulators. The high currents can cause incendiary sparking at poor joints in metallic
components.

8.2.5 Thermal and voltage sparking is probably the principal cause of rim seal fires. This arises from
the current crossing the rim seal; and because of, in practice, the poor joint which occurs
between the shunt and the shell, which is not a satisfactory design for passing current with
no sparks. The poor joint is due to the presence of rust, wax, surface coatings etc. on the shell
and insufficient contact area and pressure.

8.2.6 By contrast, fully welded gas tight steel tanks, as used for LPG, which have no equivalent of
the rim seal are, in fact, self protecting, and require earthing only.

8.3 PROTECTION METHODS TO PREVENT LIGHTNING INDUCED RIM SEAL FIRES

8.3.1 Low leakage seals maintained in good condition are a primary protection principle.

8.3.2 Since lightning currents are difficult to prevent (except by roofing over the tank with metal,
or having a mesh type catenary system over the entire tank, which is not very practicable) it is

Phase 1:32
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

necessary to provide a method of permitting the surface discharge currents to flow between
roof and shell without giving incendiary sparks. This could be achieved by two approaches:-
(i) Shunts which do not spark under any conditions (very difficult).
(ii) Shunts located where any sparks do not occur in, or fall into, a region of flammable
mixture (more easily attainable).

8.3.3 To achieve the objective of 8.3.2(ii) there are two main approaches; either have the shunt
contact points above the secondary seal, where any sparks produced are outside (and cannot
fall into) a region of possible flammable mixture, or have the contact points below the
liquid level, so the possible sparking sites are entirely submerged in liquid which excludes air
from the spark source. Submerged contacts will require some theoretical and experimental
assessment to determine their feasibility.

8.3.4 The mechanical shoe type seals are the more difficult to prevent sparking but fitting insulators
to the hangar system (as provided by some tank builders) and by providing good shunts,
above the secondary seal or immersed within the liquid, could cure the problem as long
as care was taken in the design of the immersed contacts to minimise inductance, and so
minimise the voltage stressing the insulators. Likewise, liquid and foam filled seals should
be fitted with shunts, either above the secondary seal or as immersed contacts, as discussed
above.

8.3.5 In general a bonding cable and the shunts spaced 3 m apart are required on all open roof tanks.

8.4 PROTECTION OF ROOFED OVER TANKS

8.4.1 With conducting roofs the rim seal problem is largely eliminated, but depending on the detail
of the roof construction, and whether a flammable mixture might be present above the
floating roof, there might be a risk of explosion from sparking etc. in the fixed roof.

8.4.2 With non-conducting roofs, the rim seal problem is almost as severe as with fully open
tanks; also, sparking in the roof structure could be a problem, again, depending on detailed
construction and the presence or otherwise of a flammable mixture in the tank. It is possible
that lightning could induce roof structural failure through arcing in confined spaces which
damages structural members. In some cases the installation of a lightning protection system
might be beneficial in the form of tapes attached across the top of the roof, or a catenary
system above it.

8.5 LIGHTNING EFFECTS ON ROOF SURFACE PRODUCT, AND PRODUCT IN PONTOONS

8.5.1 Product on the surface of the roof could, in principle, be ignited by lightning, particularly
with the roof very high, but it is fairly unlikely to be from direct strike ignition since the strikes
would normally attach to roof support legs and other elevated objects. The ignition is likely
to come from sparking at roof level caused by the presence of loose metal objects, chains,
the earth bonding lead if unsheathed etc.

8.5.2 Product in pontoons. This could be an explosion hazard if the design allows sparking at vents
or on the hatch covers to ignite any enclosed flammable mixture resulting from it.

Phase 1:33
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

8.6 LIGHTNING EFFECTS ON CONE ROOF TANKS

8.6.1 Cone roof tanks can be compared to conducting geodesic roof tanks, except that the usual
construction method is much more substantial, so giving less risk of sparking in the roof
structure.
Also, because the roof is usually 4 mm thick steel, the burn through and hot-spot
risk is very much reduced. This will be subject to experimental confirmation. A conducting
roof prevents the rim seal sparking problem of open top floating roof tanks, so that if the
cone roof tank has an internal floater, sparking at its rim seal is very unlikely. This, however,
does not remove the need for a bonding cable from the floater to the shell for electrostatic
discharge purposes, and shunts might be necessary in certain instances.

8.6.2 Vents, hatches, gauge heads etc, are possible attachment points for lightning and hence the
possible sites for incendiary sparking. Vents other than the pressure-vacuum breather type
should have a flame trap if the vent is situated somewhere on the tank where it can take
a direct strike. Good maintenance should include ensuring that all hatches etc. with metal
covers are firmly bolted, and no other loose connections exist to vents or instrument heads.

8.7 LIGHTNING EFFECTS ON PRESSURISED LPG TANKS

All steel welded and pressurised tanks containing LPG are very safe from lightning and the
risk to the tank itself is negligible. However, they do need earthing, as for all petroleum
plant, and no equipment on the top of the tank should be exposed to the risk of a direct
strike attachment. A lamp support pole or a simple steel pole is sufficient to provide cone of
protection coverage over the top of the tank.

Phase 1:34
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

9 CONCLUSIONS AND RECOMMENDATIONS

The conclusions and recommendations made in this desktop review have in some cases
now been modified by the output from the complementary testing programme. Where this
is the case the appropriate testing documents are referenced in section 10 and additional
comments made.

9.1 There are some very useful documents which have laid the ground work for a more thorough
investigation of the seal sparking problem. However, most of the standards on lightning
protection available to the oil industry, or prepared by oil industry organisations, are out of
date and/or insufficiently detailed and so do not give adequate advice for the protection
against lightning of open top floating roof tanks and other types of above ground storage
tanks.

9.2 Protection of open floating roof tanks is feasible, by due consideration of the mechanisms of
sparking, and by applying protection methods which prevent sparks in regions of flammable
mixtures. This is possible for the main types of primary and secondary seals, although there
may be a considerable cost penalty in equipping some types of tank, especially pantograph
types where there are many joints and hinges in metal components.

9.3 It is useful to know the thickness of metal skins needed to provide protection against direct
lightning attachments. EI-TN1-03 Review of burn-through and hot-spot effects on metallic
tank skins from lightning strikes discusses this further by reviewing available data. Steel skins
of 4 mm thickness or more appear to have adequate protection against puncture, but hot-
spots in 4 mm skins could be hazardous. Aluminium skins of 3 mm are easily punctured by
lightning attachments, and generally geodesic skins are thinner even than this (e.g. 1,5 mm).

9.4 The threshold level for sparking at the contact of shunts to the shell wall has been addressed
in EI-EN2-04 Lightning tests to tank shell/shunt samples. This report notes the low levels
of current required to initiate sparking, as well as the severity of sparking. It is especially
severe from the long duration currents. It also shows how protection from the latter can be
substantially improved by relatively simple bonding of the tank roof.

9.5 Where fixed roofs are installed over floating roof tanks, the lightning problem might be
alleviated considerably, but only if the roof is electrically conducting, sufficiently thick to
eliminate burn through and incendiary hot-spots, and the space above the floating roof is
maintained free of a flammable mixture. A non-conducting roof, depending on details of the
roof and the presence of a flammable mixture, could be safer.

9.6 It would appear that the liquid immersed shunt concept is worth pursuing since the advantages
of it sound very promising as a way to prevent incendiary sparks from occurring in a region
that could have a vapour/air mixture. This approach has been supported by testing in EI-EN2-
04 Lightning tests to tank shell/shunt samples.

Phase 1:35
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

10 REFERENCES

[1] IEC 1024-1 Protection of structures against lightning – Part 1: General principles,
1990
[2] National Institute of Standards and Technology on behalf of the National Fire
Protection Association Early streamer emission air terminals lightning protection
systems; Literature review and technical analysis, January 1995.
[3] CIGRE 95 SC 33 (WG 01) 17 IWD Statement of CIGRE Working Group 33.01
'Lightning', Appendix C, May 1995.
[4] BS6651:1999 Code of practice on protection of buildings against lightning1.
[5] LASTFIRE Project analysis of incident frequency survey, June 1997.
[6] AEA Technology The susceptibility of the elf oil milford haven refinery to damage
from lightning strikes, September 1994.
[7] Culham Puncture threshold tests to a range of materials, Prepared by Helen Costigan,
July 1990.
[8] S J Haigh, R E Baldwin, C R Banks Factors affecting the detection of voltage sparks
by flammable gas and optical techniques, International conference on lightning and
static electricity, Oklahoma City USA, 1988.
[9] AEA Technology (Haigh, Baldwin, and Hardwick) Fuel ignition hazards from thermal
sparks, International conference on lightning and static electricity, University of Bath,
1989.
[10] AEA Technology (Haigh, Hardwick) Parameters influencing the hot-spot ignition
of aviation fuel/air and ethylene mixtures, International aerospace and ground
conference on lightning and static electricity, Atlantic City USA, 1992.
[11] AEA Technology (S J Haigh) Lightning ignition hazards to heated JP8 aviation fuel,
16th International aerospace and ground conference on lightning and static electricity,
Mannheim, Germany, 1994.
[12] NFPA 780 Standard for the installation of lightning protection systems, 2000
edition².
[13] SIPM-MFE No 144/84 The influence of various types of roof seals on the lightning
protection of floating roof tanks, Shell memorandum 1984.
[14] SIPM-MFE No 8/86. Earthing of floating roofs of storage tanks to prevent rim fires,
Shell memorandum 1986.
[15] Hamelin, Hubert, Stark, Waters Panneau a pointes multiples pour la protection contre
la foudre. Etude experimentale et interpretation theorique, Report no DPh/EP/81/230,
June 1981.
[16] J L Bryan, R G Bierman GA Erikson (for NFPA) Report of the third party independent
evaluation panel on the early streamer emission lightning protection technology,
Sept 1999.
[17] LASTFIRE Project. Large atmospheric storage tank fire project. Lightning protection of
floating roof storage tank, May 1998.

1 See Annex B.1.


2 See Annex B.3.

Phase 1:36
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

ANNEX A
SKIN EFFECT IN STEEL SHEETS

For fast pulses of current, the skin effect principle requires current to flow as a thin skin layer
on each side of the conductor, but not through it from one side to the other (see Figure A.1).
This accounts for the current flow lines in Figures 5a, 5b and 5c.

Figure A.1 Cross section of the top of the steel shell, showing how current flows
as a thin skin layer on each side of the metal in a layer less than 0,5 mm thick. The
fast pulse current cannot flow through the steel from one side to the other; it has
to flow over the surface. Only very long duration pulses can flow through the steel
from one side to the other i.e. pulses lasting at least many tens of milliseconds.

Current flows as a skin layer in


the steel plate, and does not
flow in the middle.

Cross section of the top of the


shell. The metal plate is >5 mm
thick, but the current only
penetrates <0,5 mm.

Figure A-1: Cross section of the top of the steel shell, showing how current
flows as a thin skin layer on each side of the metal in a layer less than
0.5mm thick. The fast pulse current cannot flow through the steel from one
side to the other; it has to flow over the surface. Only very long duration
pulses can flow through the steel from one side to the other i.e. pulses
lasting at least many tens of milliseconds.
Phase 1:37
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

ANNEX B
REVIEW OF DOCUMENTS

Note: This review was undertaken at the time of the investigation, not at the time of
publication.

B.1 BS 6651:1999 CODE OF PRACTICE ON PROTECTION OF BUILDINGS AGAINST


LIGHTNING3

This document is unchanged from the previous 1992 edition in respect of items affecting
storage tanks, although there are many other detailed alterations. As in the 1992 edition, no
specific mention is made of protecting FRTs or any other type, although it is still a very useful
general purpose protection document.

B.2 API RP 545 RECOMMENDED PRACTICE FOR LIGHTNING PROTECTION OF ABOVE


GROUND HYDROCARBON STORAGE TANKS

This is under preparation, and will encapsulate findings from this investigation.

B.3 NFPA 780 STANDARD FOR THE INSTALLATION OF LIGHTNING PROTECTION


SYSTEMS 2000 EDITION4

This document contains rather more than BS 6651 on tank protection, and although it goes
some way towards useful protection advice, it needs to be updated to take into account the
recent findings on tank sparking mechanisms and protection techniques.
In 6.4.1.1 item 3 says '...all vapour or gas openings shall be closed or provided with
flame protection...'. It would have been useful to mention pressure-vacuum vent valves
specifically as suitable vents, without the need for flame traps.
In 6.4.1.2 on floating roof tanks, there is some very confusing advice especially in
section (b) on protection, where it is stated '... Tanks without a vapour space at the seal shall
not require shunts at the seal'. This is not in accordance with good practice, and the present
study confirms that it is good practice always to have rim seal shunts.
On the question of nomenclature, the first paragraph of 6.4.2.1(a) refers to the rims
of tanks, apparently meaning the top of the shell, rather than the usual usage where the rim
is the region of the floating roof nearest to the shell.
In summary, this document could be improved by editing to bring it in line with
current practice.

3 Since publication of this report, this standard has been superseded by BS EN 62305: 2006 Protection against
lightning.
4 Since publication of this report, this standard has been superseded by NFPA 780 Standard for the installation
of lightning protection systems 2008 edition.

Phase 1:38
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

B.4 API 650 WELDED STEEL TANKS FOR OIL STORAGE 10TH EDITION. NOV 98 WITH
ADD 35

This document, like EEMUA 159, is a document covering the mechanical aspects of tank
design, build, and maintenance, although both documents have small sections relevant to
lightning protection.
API 650, in 3.10.2.2 states that the minimum roof plate material thickness for internal
floating roof, and fixed roof tanks shall be 3/16', nominal 5 mm, without allowance for
corrosion, although thicker plates might be required for self supporting roofs.
Appendix C, paragraph C.3.3.2 states that floating roof deck plates shall also have a
minimum nominal thickness of 3/16' (nominal 5 mm). This is of relevance to the burn through
and hot-spot phenomena, but in practice a strike directly to the deck plates is unlikely (an
attachment to a support leg or similar is more likely), and also there is likely to be liquid
directly under the plate, except in double skinned floating roofs, so an ignition would be
unlikely.
Paragraph C.3.1.3 refers to shunts and mentions API 2003; this document has now
been edited to exclude lightning related subject matter.
Appendix G refers to aluminium dome roofs (geodesic) roofs, and paragraph G.2.3
states that roof panels should be fabricated from Series 3000 or 5000 aluminium, with a
minimum nominal thickness of 1,2 mm. This is of particular interest to this study because of
the risk of burn through and tank explosions with this small thickness of aluminium.
Paragraph G.5.4 refers to bonding of the dome to the shell, which, as a minimum,
should be done by 3/16 ' stainless steel bond straps across every third support point to the
shell. Stainless steel is a very high resistivity material, so does not make ideal bond straps. (Its
resistivity is 45 times higher than copper.)

B.5 EEMUA 159 USER'S GUIDE TO THE INSPECTION, MAINTENANCE AND REPAIR OF
ABOVE GROUND VERTICAL CYLINDRICAL STEEL STORAGE TANKS, 3RD EDITION
2003

This document is also a primarily mechanical construction document with a few items relevant
to lightning protection.
It has some useful illustrations of the various types of seals, both primary and
secondary, and also information on rolling ladders for floating roof tanks, which are known
to be troublesome.
Regarding bonding cables for floating roof tanks, it recommends that 'earthing
cables should be directed away from ladders to prevent entanglement'.
Paragraph 12.7 makes the following observations:
'Shunts should be provided, spaced no more than 2,5 m (compared to 3 m usually),
around the periphery of the roof rim, and should be in open air above the secondary seal.
However, where secondary seal design allows for easy inspection, the shunts may be placed
in the void between the primary and secondary seals.'
The present study suggests that this is very bad advice, and the shunt should be
above the secondary seal, unless an immersed shunt is used, in order to prevent incendiary
sparking.
Paragraph 12.8.4, concerned with maintenance, recommends frequent inspection of
the shunts to ascertain their condition.

5 Since publication of this report, this standard has been superseded by API 650 Welded tanks for oil storage,
11th edition, 2008 with Add 1.

Phase 1:39
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

B.6 NFPA 30 FLAMMABLE AND COMBUSTIBLE LIQUIDS CODE 2003 EDITION6

This is a general flammable and combustible liquids code, with two sub paragraphs which
have a relevance to this study. These are:
−− Paragraph 4.2.5.1.6 Tanks and pressure vessels that store Class 1A liquids (i.e. flash
point below 73 ºF, and boiling points below 100 ºF) should be equipped with pressure/
vacuum vents, which are normally closed except when venting.
−− Paragraph 4.2.5.1.7 states that tanks or vessels storing Class 1B and 1C liquids (i.e.
liquids with flash points below 73 ºF, and boiling points at or above 100 ºF, (1B), and
liquids with flash points at or above 73 ºF but below 100 ºF (1 ºC)), shall be equipped
with venting devices or with listed flame arrestors. During this study, it became clear
that flame arrestors may cause blockage problems, and as a result pressure-vacuum
vents without flame arrestor are preferred.

B.7 REVIEW OF INFORMATION FROM CHICAGO BRIDGE AND IRON7

Comprising:
(i) lightning testing, an article concerning lightning test work on shunts, and
(ii) lightning testing (update).

Item (i) provides more detail of the testing and recommendations carried out by Dr Whitehead
on shunts. It is a rather confusing account, because it appears to indicate the use of two sets
of shunts, one above the primary seal, and one above the secondary seal; however later
in the article the need for shunts above the secondary seal is referred to. It seems that this
is similar to the EEMUA 159 recommendation for a possible shunt position between the
primary and secondary seals. As stated above, the present study would indicate that the
shunts should always be above the secondary seals, and should be the only current route
at the rim between the shell and the roof. The only alternative to this is if immersed shunts
were used.
Item (ii) refers to a Lightning Test update, and to the development of an improved
insulator for use with the insulated hangar system for shoe seals. This is a worthwhile route
to take for reducing fire risks in floating roof tanks.

6 Since publication of this report, this standard has been superseded by NFPA 30 Flammable and combustible
liquids code 2008 edition.
7 Not published.

Phase 1:40
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

ANNEX C
WORLD ISOKERAUNIC LEVELS

Figure C.2 World Isokeraunic Levels

Phase 1:41
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

CUL/LT-0235

REVIEW OF TANK BASE EARTHING AND TEST CURRENT


RECOMMENDATIONS

CONTENTS

Page

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

2 Tank base earthing: requirements for earth resistance to optimise
lightning protection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

3 Effect of physical characteristics of tanks on earthing and lightning safety . . . . . . 47


3.1 Tank weight and dimensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.2 Earthing and bonding cable for the shell to the floating roof of a FRT . . . . . . . . . 47

4 The effect of the environmental membrane on the earthing of the tank . . . . . . . . 49

5 The effect of tank bonding and grounding on impressed current


cathodic protection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

6 Recommendations for the test stroke parameters for sparking


and burn-through tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
6.1 Recommended current parameters for the shunt sparking tests . . . . . . . . . . . . 51
6.2 Recommended continuing current pulse for hot–spot and burn through tests,
and for literature comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

7 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

Figures:
Figure 1 Showing definitions of a1, a2, and a3 . . . . . . . . . . . . . . . . . . . . . . . . . . 51

Phase 1:43
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

1 INTRODUCTION

This report covers the grounding and bonding aspects of above ground storage tanks, and
their influence on lightning induced fires. Their impact is judged to be negligible for fires, but
important for other reasons.
The report also includes an assessment of the current amplitude and waveform
appropriate for lightning simulation tests (the 'test stroke'). This 'test stroke' is intended for
laboratory assessment of sparking, hole burn through and hot-spot formation.
Cathodic protection of steel tanks and pipes is also discussed.

Phase 1:44
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

2 TANK BASE EARTHING: REQUIREMENTS FOR EARTH


RESISTANCE TO OPTIMISE LIGHTNING PROTECTION

Lightning safety for tanks is described in CUL/LT-0234 Review of lightning phenomena and
the interaction with above ground storage tanks, and depends almost entirely on conditions
in regions where there might be a vapour/air mixture or exposed combustible liquid or wax
etc, that is, it depends on:
−− For open top FRTs: conditions around the rim seal area and other details associated
with the floating roof, like liquid on the top, loose hatches etc.
−− For roofed over FRTs: whether the roof is conducting or not, sparking or burn-through
on the roof structure for conducting roofs, or for non-conducting roofs, rim seal
problems as for open FRTs.
−− For cone roof tanks: construction of the cone roof, and vents etc. on it, and whether
a flammable vapour/air mixture is contained within.
−− LPG tanks: generally no problem areas.

There are no known occurrences of lightning induced fires around the base of tanks, or
underneath tanks, which are the only places where inadequate earthing of the tank would
result in sparking or other sources of ignition. The initial attachment process of the lightning
strike would be unaffected by the tank earth resistance, since it will always be sufficiently low
to permit the lightning leader attachment process to occur, even if the resistance to earth of
the tank were many tens of ohms.
The clear conclusion that can be drawn from these possible problem areas referred
to above is that none is critically dependent, or even dependent at all, on the tank earthing.
This is because the tank will inevitably have earth conductivity through either its massive steel
structure in contact with the ground (see section 4), or through the earth rods (four usually)
that are specified in some tank earthing standards. There will also be additional earthing
through the many pipes and cables that connect to each tank.
Thus, even with only say four earth rods, the tank is earthed well enough to conduct
the strike to ground. Since the earthing points are so remote from areas of flammable
vapour/air mixes (such as the rim seal or roof), effects at the earth have little or no bearing
on lightning induced fires.
As described in CUL/LT-0234 Review of lightning phenomena and the interaction
with above ground storage tanks section 2.1 and Figures 5a and 5b, current flow paths from
a strike to the top of the shell or the roof involve the whole tank, that is, there is a distributed
path to earth all over the outside of the tank, and down to ground, and the effects at the
areas of current flow into the ground are so remote and 'decoupled' from the tank contents
that there is no problem.
This is not to say however that the tanks can be left unearthed. For personnel
safety reasons, for mains earth faults, and for safety of both the electricity supply system
and instrumentation, the individual tanks in a tank farm should be earthed according to
recognised standards, as referred to below. The earthing systems of adjacent tanks should
also be interconnected to each other, and to the instrumentation/control centre, to help limit
lightning transients in the instrumentation circuits.
Typical recommendations in BS 6651:1999[1] are that each building or structure should
have a resistance to earth of less than 10 Ω, (or in some cases 4 Ω), and this recommendation
can be applied to tanks. In practice, because of the multiple interconnections between tanks
through earth armouring on cables, instrumentation earths and pipe-work, the combined earth
resistance of a tank farm will be very low, usually much less than 1 Ω, without any recourse
to special earthing measures. As a precaution, however, each tank should be equipped with

Phase 1:45
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

earth rods distributed equally around the circumference, to ensure that adequate earthing is
achieved even if an insulating membrane is incorporated under the base. Such earthing will
also help protect the cathodic protection system. Furthermore, as indicated above, the earth
systems of adjacent tanks must be bonded together.

Phase 1:46
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

3 EFFECT OF PHYSICAL CHARACTERISTICS OF TANKS ON


EARTHING AND LIGHTNING SAFETY

3.1 TANK WEIGHT AND DIMENSIONS

In view of the factors concerning tank earth resistance in section 2, it may be deduced that
tank weight, diameter, whether full, almost empty, or at an intermediate level has almost
no bearing on lightning induced fires or explosions (although roof height and fill level may
both have a bearing on other ignition mechanisms). Likewise the underlying material and
conditions under the tank have little or no direct influence on lightning induced fires and
explosions, but all these factors have an indirect influence. The diameter of the tank affects
the strike risk to a particular tank, since the collection area of the tank (i.e. its probability
of being struck) goes up as the square of the diameter. Also, for an open FRT, the height
and diameter of the tank, and the height of the roof inside it, influence the likelihood of a
strike directly to the floating roof. These factors though are not related to earthing. Likewise,
the condition of the ground under the tank affects ignitions indirectly, because ground
subsidence can make the tank go out of round, and cause widening of the rim seal gap. In a
severe case this might open up to such an extent that the seals cannot keep the gap sealed.
Again however, this is not an earthing concern.

3.2 EARTHING AND BONDING CABLE FOR THE SHELL TO THE FLOATING ROOF OF A FRT

Principally for lightning, and to a lesser extent for discharging of static charge in the petroleum
product, the floating roofs of all FRTs (including roofed over types and cone roof tanks with
floaters), require good electrical conductivity between the shell and the floating roof.
This is ideally provided by a cable to provide a very low resistance connection between
the roof and the shell, and should have an outer insulation jacket over its entire length from
the connection point on the shell to the connection point on the roof. This is so that when it
is lying loosely on the deck of the roof it does not make accidental contact with other deck
items and initiate sparking.
Having a low resistance between the shell and floating roof via the cable will largely
prevent arcing/sparking from the 'continuing current' component of a lightning strike to
the roof itself. By having a low resistance path it is likely that the volt drop along the cable
during the continuing current flow from the roof to the shell will be low enough to permit
any arcs at the shell to shunt contact points to extinguish. By contrast, the continuing current
from strikes to the top of the shell would be largely confined to the shell alone, because of
its massive construction from thick steel, the resistance will be so low as to prevent voltages
across the shunts being high enough to permit arcing and sparking.
This is a different situation from that described in section 2.1 of CUL/LT-0234 Review of
lightning phenomena and the interaction with above ground storage tanks, in which current
paths for the fast current component (the first return stroke and restrikes) were described.
For these components, skin effect and the inductance of the tank dominate current flow
routes, and result in current flow through the shunts, and consequent sparking.
The static charge effect for which the roof has to have electrical contact to the
shell is associated with bulk electrical charging of the liquid when being pumped into the
tank. Fast flow through pumps, filters, etc, causes a charge exchange process leaving a
net charge in the liquid. This charge relaxes over a period of time inversely proportional to
the liquid conductivity; during the charge relaxation process, charge flows to the roof skin,

Phase 1:47
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

(and to the shell and floor) so there must be a path to earth for this charge. For static, the
shunts, the ladder and other metallic items make a sufficiently low resistance path. If there
were no electrical path, the roof would charge up to a high voltage until it sparked over to
somewhere on the shell with the possibility of an ignition. Having a low resistance bonding
cable thus assists lightning and static safety. Static charging and lightning are otherwise not
related, because the stored liquid is always so resistive that it plays no part in conduction
of lightning, and lightning current flow in the tank does not cause static charging of the
liquid. For comparison, the electrical bulk resistivity of typical metals is of the order of 10-7
Ωm, whereas petroleum liquids are of the order of 107 Ωm, a difference of 14 orders of
magnitude, i.e, 100 million million times more resistive.

Phase 1:48
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

4 THE EFFECT OF THE ENVIRONMENTAL MEMBRANE ON THE


EARTHING OF THE TANK

In section 2, the discussion concluded that tank earthing is not an important contributing
factor to the prevention of incendiary sparks or of rim-seal fires, or explosions in roofed over
tanks. The environmental membrane (where used, and its use is by no means universal)
would be expected to have the effect of reducing the conductivity between the tank floor
and the ground. However, where earth rods are installed, and where there are many pipes
and other metallic connections to the tank, the tank will be adequately earthed to meet the
electrical safety requirements. As explained previously the quality of the earthing will have a
negligible effect on rim seal fires, explosions etc.
Thus in summary, the existence or otherwise of an environmental membrane is not
relevant to the prevention of lightning induced fires or explosions, but the tank earthing
should nonetheless conform to electrical safety requirements.

Phase 1:49
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

5 THE EFFECT OF TANK BONDING AND GROUNDING ON


IMPRESSED CURRENT CATHODIC PROTECTION

The use of impressed current cathodic protection (i.e. the use of an electrical power source
and a sacrificial anode to minimise the corrosion of steel components in harsh environmental
conditions), in itself has little effect on the lightning fire and explosion safety of above ground
storage tanks. In the UK it would appear that it is very rarely used, perhaps because there are
few places with severe environmental conditions necessitating the installation of a cathodic
system. All tanks must be earthed for electrical and lightning safety considerations; however
pipe connections might use insulated joints to confine the cathodic protection current to the
tank (or structure) that requires the protection. Lightning could however affect the cathodic
protection system owing to the probability that some fraction of the lightning current might
cause transient damage to the cathodic system power supply and monitoring circuits.
It is often recommended that any pipe joints which have insulating gaskets in them
for the purpose of cathodic protection isolation should have those joints equipped with gas
discharge tube protectors to allow current to bypass the insulated joint at the time of a strike.
This is so that the lightning current can spread between the various tanks and into the whole
tank farm earth system. It is better for minimising effects to the instrumentation that the
current can disperse over a wider area than just the one tank which is struck. This is what gas
discharge tubes permit by making parallel current paths available.

Phase 1:50
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

6 RECOMMENDATIONS FOR THE TEST STROKE PARAMETERS


FOR SPARKING AND BURN-THROUGH TESTS

6.1 RECOMMENDED CURRENT PARAMETERS FOR THE SHUNT SPARKING TESTS

Shunt sparking may occur through two types of current pulse, the fast current component,
which for natural lightning may be in the range from 5 kA to 200 kA peak, although the
median strike value is approximately 30 kA. In some instances (see section 3.2) the continuing
current component may also cause shunt sparking. However the fast current component is
likely to be the predominant cause.

6.1.1 Current pulse recommendations for the fast current pulse

For strikes to the top of the shell, the current pulse likely to go through any individual shunt
can have a very wide amplitude range. This is because of three main factors: the current in
the lightning flash, the vertical distance a1 between the strike point on the top of the shell
to the shunt, and the shunt-to-shunt spacing a2 around the floating roof (see Figure 1). To
minimise the number of variables to deal with, it will be assumed for the first calculation that
there will be a median strike (30 kA) with a2 = 3 m (10 feet). (Adjustments for different values
can be made after initial testing.) This leaves the primary variable as a1, to which must be
added the horizontal distance between the strike point to the nearest shunt, a3.

Figure 1 Showing definitions of a1, a2, and a3. View is looking at the inside of the
tank wall from roof level

Top of shell

a3
shunt
shunt a1

a2

Level of roof

There are two extreme cases for this, either directly above a shunt a3 = 0 (maximum), or at
half way between two shunts a3 = a2/2 (minimum). Thus for the roof as high as it can go,
and a strike directly above a shunt (a3 = 0) a value of a1 of 1 m will be assumed. The adjacent
shunt will then carry approximately 35-40 % of the current which is approximately 11 kA.
For a strike halfway between shunts i.e. a3 = a2/2 the current in the two nearest shunts will
be approximately 7 kA. These values will scale up and down with the absolute magnitude of

Phase 1:51
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

the lightning strike, so for Ilightning = 100 kA, the values would be 37 kA and 23 kA respectively.
These are the most severe current parameters ever likely to be carried by shunts.
The least severe will be when a1 is a maximum, that is when the roof is as low as it can be.
Then the current per shunt will be down to approximately 700 A.
For strikes directly to the roof, the current will spread out in all directions over it and the
average current per shunt will be simply Ilightning divided by the number of shunts. Again the
maximum current per shunt will be when the roof is high and the strike is close to the shell,
giving probably maximum values as above of 11 kA or 7 kA for a median strike. Thus for
shunt tests the current range has to be from 11 kA to 700 A.

6.1.2 Current pulse recommendations for the 'continuing current'

The only situation that the shunts may be involved in continuing current is likely to be with
strikes to the roof itself of an open FRT. In this situation current per shunt is very difficult
to predict because of the effect of the individual shunt arc resistance on the proportion of
current it will carry. However it is likely that an individual shunt could carry around 20 % of
the total current which for a fairly severe continuing current component of 400 A would be
80 A. The duration would be up to 0,5 s.

6.2 RECOMMENDED CONTINUING CURRENT PULSE FOR HOT–SPOT AND BURN


THROUGH TESTS, AND FOR LITERATURE COMPARISONS

These tests are for simulating the whole lightning current arc attaching to one position (see
Figure 9 in CUL/LT-0234 Review of lightning phenomena and the interaction with above
ground storage tanks). Thus, the entire lightning pulse parameters for continuing current
may flow into one attachment point. Thus, parameters up to 400 A for 0,5 s apply.

Phase 1:52
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

7 REFERENCES

[1] BS 6651:1999 Code of practice on protection of buildings against lightning

Phase 1:53
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

EI-EN2-04

LIGHTNING TESTS TO TANK SHELL/SHUNT SAMPLES

CONTENTS

Page

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

2 Exposed shunt tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58


2.1 Test set-up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
2.2 Exposed shunt test results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

3 Immersed shunt tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64


3.1 Test set-up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
3.2 Immersed shunt test results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

4 Conclusions from shunt tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68


4.1 Exposed shunt tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
4.2 Immersed shunt tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

5 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

Annexes:

Annex A Typical applied current waveforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

Figures:
Figure 1 Test set-up for tests to shunt/shell interfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
Figure 2 Test rig used for immersed shunt tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
Figure A.1 Typical fast component waveshape (scaled comp. D) . . . . . . . . . . . . . . . . . . . . . . . . 70
Figure A.2 Typical long duration current waveshape (scaled comp. B) . . . . . . . . . . . . . . . . . . . 71

Photographs:
Photograph 1 Typical photograph from this test series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
Photograph 2 Typical sparking from fast component at left, and the long
duration current, at right . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
Photograph 3 Long duration current sparking effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
Photograph 4 Typical sparking from steel shunt at left and copper at right . . . . . . . . . . . . . . . 63
Photograph 5 Effects of fast component only (12 kA) on a steel shell with waxy deposit applied . . 63
Photograph 6 Shunt configurations in these tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
Photograph 7 Shunt immersed just below fluid surface (50 mm) . . . . . . . . . . . . . . . . . . . . . . 66
Photograph 8 Shunt now 140 mm below the surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
Photograph 9 Shunt also 140 mm below the surface, but now a wire loop shunt . . . . . . . . . . 67
Photograph 10 Surface of the shell sample after some of the tests . . . . . . . . . . . . . . . . . . . . . . 67

Phase 1:55
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

SUMMARY

This report describes lightning simulation tests to samples representing interfaces between
shunts and tank shells. Currents at various levels were conducted into the tank shells via
improvised or representative shunts, and types and levels of sparking observed.
Tank shell samples included new (bright) steel, rusty steel, and new steel with a waxy
crude deposit.
Test were carried out to both external shunts and shunts immersed in fluid (water
for the tests).
These tests were carried out to obtain an understanding of the factors which might
influence the creation of hazardous arcing and sparking at shunt/shell interfaces.
The first test series was to heavily rusted white product tank shells, in which severe
arcing occurred at the shunt even for relatively modest currents and charge transfers.
Long duration currents crossing the interface produce the most intense sparking,
in particular showers of sparks (burning metallic particles). The hazardous nature of these
sparks is linked to the metal involved, and can probably be modified by use of different shunt
materials.
Later tests were carried out to assess new bright steel shells; however these too
produced copious sparking when used with representative shunts.
Similarly new shells coated with waxy deposits (crude oil residue) produced copious
sparking.
Effects at submerged shunts were finally studied. If an arc is formed at the shunt
contact then it was found than an eruption of liquid can take place.

Phase 1:56
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

1 INTRODUCTION

During visits to two oil refineries as part of this programme, several different types of tank
shell were made available for inspection. These are described in some detail in EI-Vis1-01 Visit
to oil refinery A and EI-Vis2-02 Visit to oil refinery B, but it was found that the white product
tanks were generally heavily rusted, whilst the crude oil tanks had heavy waxy deposits on
the shell wall.
Shunts in the form of stainless steel strips are installed at intervals (typically 3 m)
around the floating roof, to provide a path for lightning currents between the shell and the
roof. However under the typical conditions which exist at the shell surface it seemed unlikely
that many, or indeed any, of the shunts would be making good electrical contact with the
shell.
These tests were conducted to help understand the behaviour of these interfaces
between the shell and the shunt whilst carrying low level lightning currents.
The following samples were tested with different current waveforms and different
polarities:
−− Shell samples
−− rusted white product shell samples from oil refinery A;
−− new steel samples (bright and rust free), and
−− new steel samples with waxy crude oil deposit.
−− Shunts
−− Ssainless steel (improvised), stainless steel (from oil refinery B), and copper
(improvised)
−− Submerged Shunts
−− stainless steel (from oil refinery B), and wire loop, against new samples with
waxy crude deposit, tested under 50 mm and 140 mm of water.

Phase 1:57
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

2 EXPOSED SHUNT TESTS

2.1 TEST SET-UP

Figure 1 Test set-up for tests to shunt/shell interfaces.

Live
Shunt

Earth

Flat steel Screw jack


shell (variable shunt
position)

Insulating
Supports
Fibre Optics

Floor
Experimental arrangement sits in a partially
darkened room, under observation by cameras
and a fibre optics/photomultiplier system.

Note: A slightly different version was used to test the submerged shunts (see section 4).
The test set-up (Figure 1) had a shunt mounted on a screw jack which allowed the
shunt contact to be moved between tests. This could be done remotely, from outside the test
facility. Similarly the camera within the facility was operated remotely.
In the first test series the shunts were improvised out of stainless steel and copper
strip. Later tests used representative shunts from oil refinery B, but these tests gave results
similar to those from the improvised stainless steel shunts.
Photograph 1 shows a typical result photograph, showing also the screw jack
mechanism which was manufactured for these tests.
For this test the interface is a clean steel surface with a waxy crude oil deposit applied.
At the shunt/steel interface there is an intense localised arc produced, which is not usually
directly visible in the photographs but produces a bright glare. In addition, where there are
waxy deposits on the wall of the tank there is usually a local flash-flame, which is assumed to
be burning of the vapours which have been evaporated from the wax at the arc root.
Both the arc and the flash-flame would of course be an effective ignition source,
although they are fairly localised at the contact point of the shunt, and might be some
distance above the seal interface.

Phase 1:58
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

Perhaps of more concern would be the showers of spark particles which are ejected
from the interface, especially by the long duration currents, even when the injected charge
transfer is at a relatively low level. In Photograph 1, the slow current components were not
applied, and the spark particles are few, and fast moving.
These bright spark particles are not simply hot, they are burning in the air. Some
metals burn fiercely and are very hot, whilst others such as copper barely react and are
relatively cool. Some metals are therefore more hazardous to use than others in flammable
environments[1].
Generally the more hazardous a spark particle, the brighter it appears on film.

Photograph 1 Typical photograph from this test series, recorded as a four second
time exposure. This photograph records a test to a steel shell which has had a waxy
deposit applied. The apparatus in the foreground is the screw jack mechanism for
remotely repositioning the shunt against the test sample shell.

Spark
particles
Flash flame

Shunt

Wax

Phase 1:59
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

2.2 EXPOSED SHUNT TEST RESULTS

The following observations were made during the test to the exposed shunts (i.e. located
outside the fluid).

2.2.1 Current components, waveforms and levels

Lightning can be assumed to consist of at least one very fast high energy transient (typically
40 kA peak, 100 µs duration) component, plus a long duration current (typically 200 A, 0.5s).
The separate effects of these different components were investigated in these tests. The fast
components (at levels of only 9 kA) produced an arc spot throwing out small spark particles
at high speed, whereas the long duration currents (at levels of only 6 C) produced a relatively
intense shower of bright, large sparks. Typical test waveforms applied during these tests are
given in Annex A.

Fast

40 kA, 100 µs

Long Duration
200 A, 0,5 s

2.2.2 Sparking effects from fast components

The fast components will predominantly distribute inductively, which means the current seeks
to spread out and follow as many parallel paths as it can. We would expect, for a strike to
the top of the tank shell, that these fast components would spread out over the tank roof,
and that arcing would occur at many of the shunts. This is unavoidable with present shunt
designs. It is worth noting that these fast components seem to produce showers of spark
particles which are small and fast (Photograph 2 left). They do not produce as heavy showers
of spark particles as the long duration currents (Photograph 2 right).

2.2.3 Sparking effects from long duration currents

The long duration currents are found to create heavy spark showers on crossing shunt
interfaces. Such sparks are believed to be more hazardous than the sparks produced by the
fast components[1]. In addition, because these particles are larger and fall downwards, they
are more likely to present a hazard at the seal level, especially if there is a gap between the
seal and the shell.
The long duration currents are relatively slow, and inductive effects are much less
important in determining how they flow. For example, a rate of current change of 1 kA/ms
in an inductance of 20 µH would generate a voltage of only 20 V. Therefore arcing over at
the shunts during the long duration currents appears unlikely so long as a parallel path is
present. So to protect against the effects of a tank roof strike (and against the hazardous type
of sparks noted above) the provision of a DC current path will provide benefits. Such a path
could be a long cable, or via the ladder (see also 2.2.8).

Phase 1:60
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

2.2.4 Shunt material

Initially shunts were improvised from thin sheet steel 50 mm wide; later tests with a
representative shunt from oil refinery B gave similar results. We know from previous
investigations [1] that other materials produce very different spark types, an effect demonstrated
here using a copper strip as a shunt. This showed that, for the rusty surface, the copper
appeared less hazardous (Photograph 4). Whilst copper is not a practical solution it points
the way to possible improvements.

2.2.5 State of the shell surface

In the second series, tests were carried out to both clean new steel with a bright finish, and
similar samples to which a waxy crude oil deposit had been applied. This application was
done by dissolving some of the waxy deposits obtained from oil refinery A in petrol, painting
on a layer of perhaps 0,5 mm thick, and allowing the wax to harden again over several hours.
The new samples still produced sparking at the same test levels, but somewhat less intense
(Photograph 3 left). The samples with waxy deposits generated sparking, but also a localised
flash-flame, which is assumed to be ignition of vaporised wax near to the root of the arc
(Photograph 3 right, and Photograph 5).

2.2.6 Current polarity

The current polarity was reversed for some of the tests with the copper shunts, since with
copper/steel at the interface the anode and cathode materials are then different. However
this did not appear to be significant.

2.2.7 Contact load

On this test rig the contact load was not quantified, but was set by feel to be similar to
loading observed at the refineries.
Where a comparative test was done with a higher contact loading on the rusty
panel, sparking still appeared at similar levels as it had for a relatively light contact loading.
Because of the rust interface layer it is likely that the contact pressure would not normally be
significant for white product tanks. However softer wax could be penetrated by higher loads,
and steel wire shunts also penetrated wax in the submerged shunt tests, at relatively low
loading. However the tanks seen during the site visits showed no sign of the shunts having
penetrated the wax, a layer of which appeared intact along the vertical shunt line.

2.2.8 Parallel paths

In a strike to the shell rim current can either flow down the shell, or spark across to flow
across the roof. We would expect the fast component to divide between both paths, but the
long duration currents to flow largely down the shell. In an attempt to see how this could
be simulated, a second cable approximately 3 m long was wired in parallel across the bank
output. Although the fast bank current divided equally between these paths, the arc did not
continue with the long duration currents; all these components passed through the parallel
short. It indicates that good bonding of the tank roof to the shell, via either ladder bonding
or by additional cable, would reduce the ignition hazard from lightning currents, even in the
event of a tank roof strike.

Phase 1:61
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

Photograph 2 Typical sparking from fast component at left¹, and the long duration
current, at right². The significant parameters are at a few % of severe threat levels.
Both tests used steel shunts.

¹ 4kA (fast) ² 830A 6C (long duration)

Photograph 3 Long duration current sparking effects at a new steel shell sample at
left, and the same sample with a crude wax deposit (right). Both tested at 830 A, 6 C. For the
wax deposit samples sparking was usually more severe than shown and generally
similar to that in Photograph 2 (right). However this photograph also shows the
flash-flame effect due to the wax deposit.

Phase 1:62
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

Photograph 4 Typical sparking from steel shunt at left and copper at right. The
differences for these two shunt materials were consistent over several tests to this
rusty shell. Both tests were at 830 A, 6 C.

Photograph 5 Effects of fast component only (12 kA) on a steel shell with waxy
deposit applied. The spark particles here are less hazardous, but at the shunt itself
there is both an arc, and a local flash flame.

Phase 1:63
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

3 IMMERSED SHUNT TESTS

3.1 TEST SET-UP

The tests to the immersed shunts were carried out using the test rig below. This is similar to
that used for the previous tests (see Figure 1) except that the moving contact could not be
used because of space limitations.

Figure 2 Test rig used for immersed shunt tests

Live

Earthy

Flat steel Video


plate

Shunt

Water
Insulating
Supports Steel plate with
Shunt
waxy crude layer

Floor

The clean steel shell plate was coated with a layer of the waxy crude oil deposit from oil
refinery A , prior to installation in the water bath. The water is being used to investigate how
fluids (such as fuels) respond to small submerged arcs. The shunt used for most tests was one
of the stainless steel types obtained from oil refinery B. In initial tests the edges of the shunt
were taped over, effectively providing a small spacer so that the shunt could not make metal
to metal contact with the tank shell (i.e. ensuring that the waxy deposit had to break down
and a submersed arc be formed).
Later tests investigated a bent steel wire as a shunt contact. These shunt configurations
are shown in Photograph 6.

Phase 1:64
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

Photograph 6 Shunt configurations in these tests. Most tests used the strip shunt at
left. Final tests used the bent steel wire.

During the tests the surface of the fluid was under observation with a video camera, and this
provided the main method of interpreting the effect. A digital camera also monitored the
fluid surface, but it was difficult to synchronise it with the fraction of a second subsequent to
the arc, so it was a fortuitous event if it captured anything useful.

3.2 IMMERSED SHUNT TEST RESULTS

Tests were carried out using both the fast components at typically 11 kA, 3 000 A²s, and the
long duration current at 750 A, 5 C. It was found that at both levels there were eruptions
of water from the bath, which occurred some short time ~20 ms - 100 ms after the impulse
began.
Because of the time delay in visible movement of the water occurring, it is not
believed that burning particles will erupt along with the water, and there was no sign of any
during these tests.
Initial tests were carried out with the shunt immersed only a short way below the
water (50 mm). Later tests were with an immersion of 140 mm.
Results were variable, in that if the shunt was against an intact layer of wax then
significant upheaval of the water occurred, but in a subsequent test (with the waxy layer
displaced) the upheaval was less. It appears that the formation of an arc results in an
overpressure (and possibly a shock wave), which acts on the water, the waxy deposit and also
the shunt. In at least one test the shunt was bent away from the wall by the impulsive force.
It is believed that a less rigid and more 'springy' mounting of the shunt would overcome this
problem.
Photograph 10 shows the waxy deposit after a few tests to different locations, and
shows how the wax has been 'blown away' around the arc points.
When the wire shunt was used the majority of tests produced less visible effect. This
is believed to be because a light pressure on the wire could easily cause the waxy layer to be
penetrated and for metal to metal contact to occur.
The results are shown in the following photographs.

Phase 1:65
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

Photograph 7 Shunt immersed just below fluid surface (50 mm). Arc products at
the shunt are visible, as is the narrow upward plume of water, which was a typical
observation for these near-surface shunts (long duration current). Plumes were
generally larger for the fast components.

Photograph 8 Shunt now 140 mm below the surface. At 11 kA of applied fast


component current a more large scale upheaval occurs. This is a view of the water
surface in a similar view as Figure 9.

Phase 1:66
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

Photograph 9 Shunt also 140 mm below the surface, but now a wire loop shunt.
At 11 kA of applied fast component current a violent upheaval again occurs here.
However for most tests with the wire shunt there was little or no disruption of the
water surface.

Photograph 10 Surface of the shell sample after some of the tests, showing the
effect of the pressure in displacing the waxy crude deposit around the arc points,
where pitting is visible.

Phase 1:67
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

4 CONCLUSIONS FROM SHUNT TESTS

4.1 EXPOSED SHUNT TESTS

−− Showers of sparks produced by the long duration currents appear much brighter and
intense than those produced by the fast components; they are believed to present
the more serious risk of ignition.
−− Sparks from the long duration currents can be substantially reduced or eliminated by
tank roof bonding using a cable, or via the ladder.
−− Reduction in hazard may also be provided by use of alternative shunt metal.
−− Factors such as surface condition of the shell, contact load, or current polarity had
minimal influence on the level of sparking.

4.2 IMMERSED SHUNT TESTS

These tests were to shunts immersed under water, but with a crude oil waxy layer at the
shunt/shell interface.
−− If immersed shunts are used then it is quite likely that eruptions of fluid a couple of
feet or so above the surface could occur.
−− No hazardous sparks were seen to be ejected with the liquid.
−− The scenario could cause a problem if fluid spray were to penetrate the seals into an
environment where sparks might be present from an attachment to the top of the
tank shell. This would be a seal design and maintenance issue.
−− The effect can also distort a conventional strip shunt used under fluid, and cause it to
lose contact. However this is believed to be resolvable by simple design modifications.
This distortion did not occur to the exposed shunts.
−− A wire loop shunt appeared to provide better continuity, and may also be less prone
to distortion under fluid.

Phase 1:68
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

5 REFERENCES

[1] Haigh, Baldwin, Hardwick, Fuel Ignition Hazards from Thermal Sparks International
Conference on Lightning and Static Electricity (ICOLSE) 1989.

Phase 1:69
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

ANNEX A
TYPICAL APPLIED CURRENT WAVEFORMS

Figures A.1 and A.2 show the two typical types of waveform applied in the test. These are
the fast high energy component, and the long duration component. Note the very significant
difference in timescales for these two types of waveform. The fast high energy component is
applied within 100 µs, and the long duration component within 10 ms. The exact shape of
the waveforms is not believed to be of significance in the results.
For the fast, high energy component the arc is formed very quickly, and with an
overpressure at the arc location which ejects any molten metal at high speed (Photo 2, left).
The long duration components have a fairly static burning arc which ejects molten
particles more gently, and the mechanism is similar to that of an arc welder.
The 8/20 µs current pulse which is often defined is of somewhat lower energy
than the fast test waveform used here, so for the same peak current we would expect a
similar effect, but perhaps less intense arcing. (The 1,2/50 µs waveform which has also been
discussed is in fact a voltage waveform, not a current waveform. Here it is the current which
is responsible for the effects observed.)
For the long duration currents the primary parameter in the evolution of sparks is the
charge transferred in Coulombs.

Figure A.1 Typical fast component waveshape (scaled comp. D)

30µs

Phase 1:70
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

Figure A.2 Typical long duration current waveshape (scaled comp. B)

10ms

Phase 1:71
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

EI-VIS1-01

VISIT TO OIL REFINERY A

CONTENTS

Page

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
1.1 Shunts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
1.2 Steel tank shell material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
1.3 Other information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

2 Observations on visit to refinery A (first visit of review) . . . . . . . . . . . . . . . . . . . . . . . . 76

3 Detailed observations on shunts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77


3.1 Tank 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
3.2 Tank 2 for Avtur (aviation turbine fuel) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
3.3 Tank 3 for crude oil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
3.4 Tank 4 for crude oil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

4 Detailed observations on the shell inside skins where the shunts contact . . . . . . . . . . 78

5 Types of primary and secondary seals used at refinery A . . . . . . . . . . . . . . . . . . . . . . . . 79

6 Objectives for subsequent refinery visit(s) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

Photographs:
Photograph 1 Spacing of shunts on secondary seal prior to assembly . . . . . . . . . . . . . . . . . . . 80
Photograph 2 Shunt above a crude oil tank secondary seal . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
Photograph 3 Different shunt type above a crude oil tank secondary seal . . . . . . . . . . . . . . . . 81
Photograph 4 Powdery rusting of new steel shell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
Photograph 5 Inner surface of Avtur tank shell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

Phase 1:73
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

SUMMARY

This note describes a visit made to oil refinery A as part of a review of the lightning protection
of storage tanks. Observations are made on the type and condition of tank shells and shunts,
on both crude oil and petroleum product tanks.

Phase 1:74
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

1 INTRODUCTION

The objectives of the visit were to provide the following information:


Sites ideally to have several tanks, and various types of tanks, including:
−− Ones by different manufacturers.
−− Different types of rim seal: including primary only, and both primary and secondary.
−− Both the Chicago Iron and Steel type of pantograph seal.
−− Liquid filled and foam rim seals.

1.1 SHUNTS

It would be advantageous to be able to get down on to the floating roof of tanks to examine
the seal region and the shunts. It would be beneficial to see how the shunts vary from tank to
tank and their typical geometry (e.g. spacing, installation and spring loading), and if possible
perhaps obtain some shunts from the maintenance department during the visit.

1.2 STEEL TANK SHELL MATERIAL

Primarily the visit would examine the state of the contact surface of the shells of several
tanks, to see how corroded or contaminated they get. If possible collect a sample of any
deposits formed on the inside of tanks.

1.3 OTHER INFORMATION

Which might best be obtained separately from the visit, would be the type of stainless steel
used in shunts i.e. what grade is typical, and whether other metals are used.
−− Paint if any that is used on the inner surface of shells.
−− Possible source of good specimens of tank shell material, shunts, paints.
−− Information about the range of fluids stored, and of their flashpoints.

It is understood investigation of some of these latter points is already underway.

Phase 1:75
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

2 OBSERVATIONS ON VISIT TO REFINERY A (FIRST VISIT OF


REVIEW)

Access was provided to four tanks, two of which were undergoing maintenance (tanks 1
and 2) and two large crude oil tanks (3 and 4) the former of which we were able to inspect
actually on the tank floating roof. On tank 2 we were able to inspect the tank from the top
of the stairway, which was still nonetheless a good inspection vantage point. Tanks 1 and
2 were accessible to us for inspection under the floating roof, as well as on top of it (both
floating roofs were safely 'landed').
Several specimens of tank shell material taken from an old tank being refurbished
were already cut to size (about 500 mm square), and ready to bring to Culham, also some
waxy crude oil residue was already prepared for us to bring away.

Phase 1:76
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

3 DETAILED OBSERVATIONS ON SHUNTS

The four tanks provided a range of details on tank shunts.

3.1 TANK 1

New secondary seals and shunts were being installed on this tank, the shunts were made
of some grade of stainless steel, ~0,5 mm thick, and 18 cm wide. The length from the
attachment bolts to the contact point on the shell was about 16 to 18 cm. The shunts were
quite 'floppy' owing to their small thickness, but none was already fixed on the tank to
determine the contact pressure. The spacing of the shunts was approx. 3,1 m (10'), which is
in line with industry recommendations (see Photograph 1). These shunts are fitted to contact
the shell above the rubbery top edge of the secondary seals, as can be seen in Photograph 1.
The secondary seals used substantial metal plates to which the shunts were bolted so as to
provide good contact back to the floating roof.

3.2 TANK 2 FOR AVTUR (AVIATION TURBINE FUEL)

This tank was undergoing maintenance. The secondary seals to be installed on this tank were
like the seals on tank 1, with similar shunts, but the shunts were spaced 4,8 m apart, judging
by the spacing on the secondary seal units still to be installed. This is about 5,5' wider than
the usual recommendation.

3.3 TANK 3 FOR CRUDE OIL

This tank was in use, but full so the roof was at the top allowing access on to it easily. The
shunts were very 'gunged up' with the thick black wax that was very hard. The shunt spacing
appeared to be the same as for tank 1 (3,1 m) and the shunts were generally similar to those
being fitted to tank 1. Contact force varied from about 1,5 lbs force to 4,5 lbs. However in
many cases the shunts appeared not to be in contact with the shell wall because the hard
wax had built up behind the shunt (see Photograph 2).

3.4 TANK 4 FOR CRUDE OIL

This tank was also in use, with the floating roof about one eighth of the way down, so roof
access was not available. However from the top of the stairway, it could be seen that the
shunts were a different type, being longer and narrower as in Photograph 3. Also the spacing
appeared to be about twice as much as tank 1. A further shunt on this tank appeared to be
quite distorted, perhaps suggesting that this type of shunt is not as reliable mechanically as
the broader type discussed in the comments on tank 1.

Phase 1:77
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

4 DETAILED OBSERVATIONS ON THE SHELL INSIDE SKINS


WHERE THE SHUNTS CONTACT

There were two types of surface effects on the shells seen.


On the crude oil tank 4 the tank shell was covered in at least a thin layer of wax on
top of any corrosion. It is possible that the deposit protects the surface against corrosion, but
the shell could not be examined in detail on this visit. (The main surface of tank 3 shell could
not be seen because the roof was within a couple of inches of its top limit, and tank 4 was
a little too low for roof access.) The waxy deposit appeared to present an insulating layer for
the shunts.
On tanks other than the crude oil, the shell surface was invariably corroded with thick
hard, even dimpled, iron corrosion material. Even on a tank with almost new shell plates
(tank 1) the steel surface very quickly becomes corroded (see Photograph 4). Normally there
were no waxy deposits formed on Avtur or motor spirit tanks, but this of course promotes
serious corrosion. It was noted that on motor spirit tanks the oxide layer can be lifted off in
thin sheets. The Avtur tank was not that badly corroded due perhaps to the extra protection
provided by the heavier liquid (Avtur being an aviation kerosene, compared with the much
lighter and more volatile motor spirit). See Photograph 5 for the Avtur tank 2. Thus the
conclusions that can be drawn are that (with the possible exception of the crude tanks) the
shell is always corroded with some form of hard or not so hard rust.
At refinery A, painting is usually restricted to the floor of the tank and the bottom
2 m or so where corrosion tends to be worse. Tanks 4 and 3 indicate that the top part of
the inside shell wall is painted too, white for tank 4 and maroon for tank 3. This is especially
significant for a strike to a tank when full, because of the high local current densities near
the strike. The insulation layer will need to be broken down for current to get through to the
metal tank, and this could promote arcing.

Phase 1:78
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

5 TYPES OF PRIMARY AND SECONDARY SEALS USED AT


REFINERY A

Unusually, all the floating roof tanks appeared to have liquid filled primary seals, in which a
liquid, usually kerosene, is used to inflate a flexible trough or bag, which presses on the shell
and on to the tank rim. This trough is of course non-conducting (although it might be static
electricity conducting, i.e. very high resistance), and so closes the gap from the shell to the
roof.
Regarding secondary seals, all the ones we saw were of the same general type,
angled steel plates with a rubber like top edge to press against the shell with shunts as
described in paragraph 3. These seals are shown in Photograph 1 prior to installation and
Photograph 2 when installed (tank 4). As well as helping with lightning protection, they are
beneficial for rainwater exclusion from the tank, for reducing emissions of vapour from the
tank and thus generally reducing fire risks. The shunts should always be placed above the
tops of the secondary seals. This is so that the contact point of the shunt is such that sparks
occurring there cannot drop into the vapour space above the primary seal.
Where the tank shell could be seen, the effect of the shunts riding on the surface
was not very noticeable. That is, the thick rust deposit was present, and barely marked by the
passage of the shunt. On the crude tanks the waxy deposit also seemed to be intact on the
shunt tracking line, although it was not possible to make a detailed inspection of these.
Samples of the waxy deposits were obtained at refinery A, but no samples of the
epoxy paint said to be used for tank protection, and no samples of shunts. From aircraft
experience, where epoxy paints are extensively used, it can be concluded that where paint
layers are used, they will be extremely tough electrically, unless the contact of the shunt has
scraped through the paint layer. Sparking is therefore almost inevitable with lightning type
currents flowing in the shunt to shell contact where there is paint.

Phase 1:79
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

6 OBJECTIVES FOR SUBSEQUENT REFINERY VISIT(S)

One point which was not clear on the crude oil tanks was the presence of corrosion (or
not) under the waxy deposits. Also whether these deposits presented an insulating layer, or
whether the shunts break through. It looked from a distance as if the waxy film was more or
less undisturbed by the shunts.
A further objective would be to see chicago bridge and iron pantograph seals, and
their secondary seals; also the condition of the shell walls, contamination, corrosion and
condition of the shunts for such pantograph tanks.

Photograph 1 Spacing of shunts on secondary seal prior to assembly

Phase 1:80
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

Photograph 2 Shunt above a crude oil tank secondary seal

Photograph 3 Different shunt type above a crude oil tank secondary seal

Phase 1:81
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

Photograph 4 Powdery rusting of new steel shell

Photograph 5 Inner surface of Avtur tank shell

Phase 1:82
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

EI-VIS2-02

VISIT TO OIL REFINERY B

CONTENTS

Page

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85

2 General design and maintenance items discussed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

3 Detailed observations on the tanks that were inspected . . . . . . . . . . . . . . . . . . . . . . . . 89


3.1 Tank A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
3.2 Tank B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
3.3 Tank C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

Figures:
Figure 1 Current flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

Photographs:
Photograph 1 Showing the corrosion inside a motor spirit tank . . . . . . . . . . . . . . . . . . . . . . . 90
Photograph 2 Geodesic tank roof at refinery B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
Photograph 3 Geodesic tank roof at an overseas refinery . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

Phase 1:83
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

SUMMARY

This note describes a visit made to refinery B as part of a review of the lightning protection
of storage tanks. Observations are made on the type and condition of tank shells and shunts,
and on atmospheric cone roof tanks. Some useful feedback was also obtained from refinery
staff.

Phase 1:84
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

1 INTRODUCTION

The aim of the visit to refinery B was to identify possible sources of tank fires from lightning
strikes, and to consider and test practical methods of improved protection.
The most significant source of ignition appears at this time to be the shunt/shell
interface, where lightning currents are capable of generating severe sparking. It was hoped
that the visit to refinery B would provide support information in particular:
−− To study the condition of typical tank shells and shunts in a typical refinery and for
both crude oil and petroleum product tanks.
−− To discuss the practical difficulties of maintenance of tank shells and shunts, to help
identify the type of design modifications that might be practicable.

Specifically it was hoped that crude shells could be inspected closely, both with waxy deposits,
and with deposits removed. Also pantograph type seals, cone roof tanks and geodesic roofs
were an object of interest for this visit.

Phase 1:85
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

2 GENERAL DESIGN AND MAINTENANCE ITEMS DISCUSSED

The design and maintenance of the tanks was discussed with refinery staff, focusing on
questions raised by Culham. The following points were noted from these discussions, and
from general observations of the tanks inspected:

−− o environmental membranes beneath the floors of tanks were fitted to tanks at


N
refinery B.

−− No cathodic protection of pipes or tanks was in use at refinery B.

−− Many of the newer tanks were built on piles because of the marshy ground. This
gives a very stable foundation for the tanks, largely preventing a potentially serious
problem with tanks, that of subsidence and distortion of the shell, cracking, leaks
etc. The distortion of the shell was a concern in relation to lightning protection and
safety, because it can be a cause of problems with the rim seals of tanks such as
crude oil floating roof tanks (FRT). Distortion can be a cause of the rim gap opening
so wide that the rim seals do not touch the shell, so exposing a gap down to the
stored liquid. No tanks at refinery B were distorted in this manner; however, on the
contrary, the tanks viewed seemed to have very good uniform contact pressure on
to the shell around the periphery, as far as could be ascertained by viewing one of
the tanks at roof level. Some of the smaller old tanks were not on piles, having only
a compacted chalk layer beneath the floor.

−− ll tanks were fitted with secondary seals above pantograph type primary seals.
A
The secondary seals were the inclined plate with diaphragm type with an insulating
rubber scuff strip along the top in contact with the shell, and with shunts generally
similar to the ones at refinery A along the top above the scuffing rubber strip, being
about 18 cm wide, and 15 cm from the bolt line to the shell contact line.

−− T he crude oil tanks at refinery B did not appear to be so heavily coated with hard waxy
deposits as at refinery A, this could be dependent on the grade of crude oil usually
stored there, but other reasons such as a somewhat higher ambient temperature, or
a more recent clean–up may be responsible. Even so it was clear that a continuous
deposit occurs over most of the shell inner surface. At the upper rim of the nearly
full tank this deposit could easily be flaked off in parts, and was approximately 1 mm
thick. The shunts had often ploughed this layer off locally, usually at one or other
shunt edge.

−− T he steel surface of the crude oil FRT which was examined appeared to have only
very light corrosion on it, again perhaps due to the type of crude which quickly
coats the shell surface with a protective thin layer of wax. By contrast the motor
spirit tank shells were very corroded, with the corrosion coming off in thin skins (see
Photograph 1). This is in accordance with observations made at refinery A, where
it was commented that Avtur tanks form a hard adherent layer of rust on the steel
shell, whereas motor spirit (petrol, gasoline) forms this very poorly adherent corrosion
so promoting relatively rapid thinning of the steel sheets which comprise the shell.
Presumably Avtur, with its lower vapour pressure and different chemical composition,
allows the steel to form the hard adherent layer that was seen at refinery A. This is
of interest for lightning work because the type of rust may well affect the degree of
sparking at shunt to shell contact points.

Phase 1:86
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

−− All the tanks looked at in detail were of a welded construction, but some of the older
cone roof tanks at refinery B were riveted. Generally, all welded tanks are better
for minimising sparking at joints etc, when carrying large lightning currents, but a
gas-tight or oil-tight riveted construction is considered satisfactory. No cone roof or
LPG tanks were looked at, but it was communicated that many of the cone roof
tanks had 'floaters' in them (i.e. a floating roof directly on top of the liquid). These
are not considered to be subject to the same lightning hazards as open top FRTs,
because there is no equivalent of the rim seal sparking problem in tanks with welded
construction conducting roofs. However there may be other issues, especially with
geodesic roofs.
One geodesic roof tank was viewed at some distance (see Photograph 2),
the roof of which was said to be of bolted construction, but the skin thickness of the
aluminium panels is not known at the present. The steel cone roof tanks have steel
at least 5 mm thick so they are able to carry the high current safely. There remains
the possibility that an incendiary hot-spot could form on the back surface of an arc
attachment point. This is discussed in EI-TN1-03 Review of burn-through and hot-spot
effects on metallic tank skins from lightning strikes.

−− All the tanks examined were equipped with flexible bonding cables from the shell
to the roof. One tank had the usual type of yellow/green insulation covered cable,
lying loosely on the floating roof deck, (which might have been a temporary cable),
another tank had two bonding cables, both fitted to automatic tensioning spools on
the top of the floating roof. These are expected to play a secondary role in lightning
protection because of their relatively large inductance, although they might be of
sufficiently low resistance to prevent sparking associated with the continuing current,
which is only a few hundred amps at most, and has a very low di/dt. They are much
too inductive to affect current flow appreciably for the fast high di/dt return stroke
and restrikes.

−− It appears to be industry practice that shunts are made from steel, but stainless
steel not carbon steel. It was noted that the shunts used on the tanks at refinery B
were grade 312 stainless steel, which is a non-magnetic austenitic type of steel. This
means that one of the metals involved in the shell to shunt contacts is uncorroded,
and in fact polished smooth by the rubbing on the shell wall as the roof moves up
and down. However it appears that the shunts do not rub the oxide layer off the shell
to any extent so that one electrode of the contact (the shell) is always corroded and/
or covered in a layer of wax etc.
It would appear that no other materials are in general use as shunts. Scrap
samples of the stainless steel shunts were provided, to be used in later tests.

−− On cone roof atmospheric tanks at refinery B, the pressure-vacuum breather vents


do not have flame traps owing to problems with them getting blocked up and so
allowing the tank to have an over or under pressure in it which can lead to collapse.
Although the vents are located high up on the roof of the tank, they do not protrude
upwards very far. Therefore the probability of them being hit directly is very low. Also,
the probability of one of the vents being open at the time of the strike is viewed as
small. Overall there is likely to be a low probability of a lightning induced ignition in
such cone roof tanks, and experience supports this.

−− Painting of tanks. Electrically, from the standpoint of lightning safety, the only
concerns with painting are where paint is on a surface upon which the floating roof
tank shunts slide, that is, the inside surface of shells of either open FRT tanks, or non-

Phase 1:87
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

conducting geodesic roofed tanks. Mostly tanks are unpainted, being in effect bare
metal, which for crude oil is protected to some extent by the wax. Motor spirit, Avtur
and Avgas tanks are usually unpainted on their shell inner surface.
Tanks are sometimes painted with an epoxy paint. Where epoxy paints are
used, it should be noted that they are very tough electrically (i.e. very good insulators)
so they will inhibit electrical contact between shunt and shell. Voltage breakdown
accompanied by arcing and sparking is inevitable. (Iron rust is essentially a non-
conductor too, so whether the surface is rusty or painted it makes for high resistance
contacts for the shunts.)

−− Halon fire extinguishing systems were not fitted to tanks at refinery B, but as is
common, several outlets of foam pipes were located around the tops of the tanks.

−− From observations on the tank floors (see 3.2), it is interesting to make a comment
about overall current flow on tank shells. These comments are expanded in
Cul/LT-0234 Review of lightning phenomena and the interaction with above
ground storage tanks and Cul/LT-0235 Review of tank base earthing and test current
recommendations.

−− The tank floor is generally 8 mm thick and welded to the tank shell walls. The
significance of this is that the skin effect will cause the fast pulses of lightning current
to flow exclusively as a thin outside layer. So for a strike to the top of the shell of an
open FRT, much of the current will flow as a surface layer down the outside surface
of the shell. Some will also flow down the inside surface of the shell to the level of
the floating roof, cross via the shunts to the floating roof, flow across the top surface
of the floating roof and then to the shunts on the opposite side of the tank, up the
inside surface of the shell to the top, and then down the outside surface to ground.
Current will enter the ground by flowing on the underside surface of the tank floor
steel plates. This current flow is shown in Figure 1.

Figure 1 Current flow

Thus no fast current will flow in structures or surfaces adjacent to the tank liquid.

Phase 1:88
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

3 DETAILED OBSERVATIONS ON THE TANKS THAT WERE


INSPECTED

3.1 TANK A

This was an open top FRT for crude oil, and was equipped with primary and secondary
rim seals. The primary seals could not be observed, because the secondary seals were in
good condition, and were pressing firmly against the shell over the length of seal that was
inspected. The secondary seals were fitted with shunts similar to the ones found on tanks at
refinery A, about 18 cm wide, and pressed against the shell at about 15 cm above the bolt
line where the shunt was fixed to the top of the secondary seal. The spacing was about 3 m
(but this was not actually measured), and the build-up of wax under the shunts appeared to
be less severe than found on tanks at refinery A.
As is common on FRTs, tank A had a foam dam immediately next to the secondary
seals, the top of the foam dam being higher than the top of the secondary seals. The shunt
contact point may have been at about the foam dam height or just below. Thus it would
be possible for sparks from a shunt to drop into the region between the foam dam and the
secondary seal, even though the shunt contact point was much higher than the dam. This is
because the thermal sparks (small pieces of incandescent material) can travel large distances
and still be very bright. However, on this particular tank, the secondary seal made good
contact against the shell so that it looked to be very unlikely that thermal sparks could drop
into the volume between the primary and secondary seals, a region where there would be a
relatively high probability of a flammable vapour/air mixture being present.

3.2 TANK B

This open top FRT was undergoing maintenance and the roof was supported on its legs. The
primary and secondary seals had been removed for the maintenance programme, so they
could not be inspected. The region below the floating roof had been completely cleaned
out as part of the maintenance and refurbishment programme, so the condition of the shell
could not be assessed. The floor of this tank was suffering corrosion and was being replaced with
8 mm thick steel sheets that will be welded into place to be oil tight. The significance of this
observation is discussed in Section 2.

3.3 TANK C

This was an FRT for motor spirit, (petrol) also in its maintenance programme, and so allowed
detailed inspection both below and above the floating roof. As with tank B, the seals had
been removed so the construction details could not be seen. During discussion on this tank,
EEMUA 159 User's guide to the inspection, maintenance and repair of above ground vertical
cylindrical steel storage tanks was referred to.

Phase 1:89
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

Photograph 1 Showing the corrosion inside a motor spirit tank. Although the
picture is slightly blurred, it shows the flaking away of the rust on part of the
surface

Photograph 2 Geodesic tank roof at refinery B

Phase 1:90
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

Photograph 3 Geodesic tank roof at an overseas refinery

Phase 1:91
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

EI-TN1-03

REVIEW OF BURN-THROUGH AND HOT-SPOT EFFECTS ON


METALLIC TANK SKINS FROM LIGHTNING STRIKES

CONTENTS

Page

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95

2 Installations vulnerable to ignition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96


2.1 Flammable range . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
2.2 Tank types and fuel/air headspaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

3 Background on puncture effects from lightning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98

4 Parameters which influence puncture and hot-spot formation . . . . . . . . . . . . . . . . . . . 99

5 Puncture effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

6 Hot-spot effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

7 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102

Figures:
Figure 1 Components of a typical lightning strike . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

Tables:
Table 1 Some examples of the volume percentage of various vapours
which define the flammable range. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

Phase 1:93
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

SUMMARY

This report describes the threat from puncture and hot-spots in igniting vapour/air mixtures
in storage tanks. A review of data from different sources suggests that puncture of steel skins
>4 mm thick is not a threat. Hot-spot effects may be a hazard at this thickness and possibly
greater, but that is not so clear.

Phase 1:94
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

1 INTRODUCTION

For some types of petroleum product storage tanks there is a possible threat from lightning
puncturing the roof of the tank. If the vapours in the head-space within the tank are within
the flammable range then such a puncture would be expected to result in an ignition. An
ignition of enclosed vapours will result in large overpressure and severe structural damage.
Similarly the effect of the arc would be to heat the roof skin locally, and it is possible
that, even if puncture does not occur, the local hot-spot could itself present an ignition
hazard.
This report reviews available data on lightning effects on metal skins. It assesses
whether there is such an ignition hazard from these sources. There is a scarcity of relevant
data, but it is clear that steel skins in excess of 4 mm thick do not have risk of puncture.
There is similarly a lack of relevant data on hot-spot effects, and although there does
not appear to be a significant risk, it is proposed that some supplementary testing be carried
out to confirm this.

Phase 1:95
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

2 INSTALLATIONS VULNERABLE TO IGNITION

Types of tank for which this could be a threat are discussed briefly in this section. Types of
tank are also covered in Cul/LT-0234 Review of lightning phenomena and the interaction with
above ground storage tanks.

2.1 FLAMMABLE RANGE

Both ignition by hot-spots and by a puncture are not particularly sensitive to the vapour
concentration. That is, the parameters of the hot-spot required to cause ignition, in terms of
temperature/time/size, do not vary much across the flammable range[11]. On the other hand
ignition by spark sources is strongly dependent on vapour concentration, so that low energy
sparks of perhaps 0,5 mJ can cause an ignition, but only if the mixture is in the flammable
range. Higher energy sparks will be required to ignite mixtures away from the optimum
proportions.

Table 1 Some examples of the volume percentage of various vapours which define
the flammable range.

Product Min % vol Max % vol


Propane 2,2 9,5
Ethylene 2,75 34
Gasoline 1,5 7,5
Avtur 1,3 8

The conditions which lead to these concentrations occurring will vary widely.
For the aviation industry this limited range of flammability has helped prevent fuel
ignitions and explosions. Older fuel types were very volatile and under most flight/lightning
strike conditions the tank vapour space was too rich to burn. Modern jet turbine fuels are low
volatility and usually too lean to burn.
Where catastrophic incidents from lightning have occurred it has been when these
fuel types were unwittingly mixed, leading to a fuel within the tank of dangerous volatility.
In the case of the TWA800 explosion in 1996, Avtur was brought into a hazardous state
primarily because the centre fuel tank was adjacent to air conditioning equipment and
became unintentionally heated[12].

2.2 TANK TYPES AND FUEL/AIR HEADSPACES

It follows then that the installations potentially at risk are only those where flammable
vapour/air mixtures could be present. These could include:
−− fixed roof tanks;
−− geodesic tank covers, or
−− pontoons on floating roof tanks (if fuel has leaked into one of the pontoon cells).

Phase 1:96
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

For fixed roof tanks, complete equilibrium is likely to have formed between the fuel and
the vapour space. This is because of the large surface area interface between the liquid and
vapour, and also because there would be little or no air flow other than through the vent.
Tanks potentially at risk would be those holding the less volatile fluids, such as some alcohols,
which form a flammable equilibrium mixture at normal temperatures. Motor spirits would
become dangerous typically at temperatures below 0 ºC.
For geodesic tanks, or tanks where there may be limited visibility of the vapour
surface, as well as some ventilation, then the more volatile liquids would be those likely to be
more dangerous. This is because the equilibrium vapour air mixture (which would be too rich
to ignite) is less likely to have formed, and there will be a tendency to have more air present
than would be the case at equilibrium.
It is believed that many fresh crude oils have flashpoints below 10 ºC, which would
make crude oils relatively hazardous in an enclosed environment. That is, it may well form an
equilibrium vapour ratio which is in the flammable range.
It would be interesting to source any data on monitoring of vapour space flammabilities
in fixed roof and geodesic tanks, and near to floating tank seals.

Phase 1:97
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

3 BACKGROUND ON PUNCTURE EFFECTS FROM LIGHTNING

Puncture of metal skins by lightning strikes has been a subject of concern for many decades in
the aviation industry. The threat for aircraft is that the thin aluminium wings directly enclose
the fuel.
Experimental investigations were being performed on puncture of metal skins as far
back as 1949[1].
In most of these studies it was aluminium skins (<3 mm) which were the primary
concern. However the reported work on aircraft skins is applicable to geodesic roofed tanks
which sometimes use thin aluminium skins (API 650 Welded steel tanks for oil storage notes
that 1,2 mm is a minimum.)
For tank roofs the plates are generally steel of 5 mm thickness, occasionally 6 mm
where there is a known corrosion problem¹.
There is relatively little published work which describes puncture or heating effects
in such thick steel sections. That which has been reviewed is discussed in section 4. It gives a
clear answer for puncture effects, but less so for hot-spots.

1 Data sourced from Culham email correspondence, not included in this publication

Phase 1:98
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

4 PARAMETERS WHICH INFLUENCE PUNCTURE AND HOT-SPOT


FORMATION

Lightning can be assumed to consist of at least one very fast high energy component (typically
40 kA peak, 100 µs duration), plus a long duration current (typically 200 A, 0,5 s). In reality
it may consist of several high energy components, but the combination of these two very
different current waveforms is a good representation, see FIgure 1.
These different components produce very different effects. The very high magnitude
fast currents are not actually significant for heating or melting of metals at the arc attachment
point. The most important effect is from the long duration current, which has an action akin
to that of an arc welder. That is it tends to melt metal at the point of arc attachment. The
heat input to the metal at the arc root is proportional to the charge transferred in Coulombs
(C) (i.e. I x t or ∫I dt). However, there are also time dependencies which determine whether a
given charge transfer will cause puncture or not[3]. Even so, test data tend to relate puncture
and hot-spot effects to a known charge transfer, and in this note it is with reference to an
assumed severe strike value of 200 C.

Figure 1 Components of a typical lightning strike

Long duration charge


transfer is the most
significant parameter for
Fast
puncture effects
40 kA, 100 µs

Long Duration
200 A, 0,5 s

Phase 1:99
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

5 PUNCTURE EFFECTS

Most of the test data relate to thin panels, and it might be assumed that the area of damage,
or volume melted, can be read across from these thin panels to thicker ones. In fact both
testing and modelling show that this is not the case.
For example[8] carrying out modelling for 100 C applied charge transfer shows a melt
radius of:
−− 4,8 mm in 0,8 mm ferrous steel (volume 58 mm³),
−− 3mm radius ~ <1 mm depth in 10 mm ferrous steel (volume ~ 15 mm³).

The difference occurs because there is much more effective conduction of heat away from
the arc root with thicker panels, so that less material is melted.
In another example[2], tested 5 mm thick samples including both stainless and ferrous
steels, with both 100 C and 400 C (that is, either 200 A or 800 A applied in 0,5 s). Molten
pool volumes were significantly higher with the higher charge transfer i.e.:
−− 100 C: pool volume 64 mm³
−− 400 C: pool volume 400 mm³

Significantly neither stainless nor ferrous steels were punctured at this level. The depth of
penetration of the molten pool was 2,1 mm to 2,5 mm for 100 C, and 3,5 mm to 3,7 mm
for 400 C.
Following the work in Cul/LT-0234 Review of lightning phenomena and the interaction
with above ground storage tanks and Cul/LT-0235 Review of tank base earthing and test
current recommendations a threat of 200 C is being considered an adequately severe level
against which to protect.
Testing[7] concludes that for 200 C steels can be punctured at 2,5 mm thick, but that
at 3,5 mm there is no possibility of puncture.
From these tests[2, 7] we can be clear that there is no possibility of a 5 mm steel shell
being punctured, and with a safety margin of at least 1 mm to allow for corrosion reduced
thickness.
Aluminium can be punctured by 200 C at up to 3 mm[7], and extensive aircraft testing
experience shows that a 2 mm skin can easily be punctured by only 40 C or so. Since typical
geodesic roof thickness is only 1 mm² then puncture of such roofs will occur easily at a
relatively low strike level.

Phase 1:100
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

6 HOT-SPOT EFFECTS

A less clear conclusion can be drawn about the possibility of hot-spot effects leading to
ignition of a vapour/air mixture.
Measurement of rear surface hot-spot temperatures has been reported in the
literature, as well as modelling effects.
Temperatures for 4 mm steel up to 940 ºC, and 750 ºC for 5 mm steel[7] have been
reported.
It is not easy to relate a maximum temperature to an ignition threat. The ignition
threshold depends on hot-spot size and duration as well as temperature. Work carried out
suggests that for hot-spots of small sizes (< 100 mm²) and durations of perhaps a second or
two then temperatures of around 900 ºC would be required for ignitions to occur[11].
Therefore it appears an unlikely threat for 5 mm steel tank shells, but a possible one
for 4 mm.
It is proposed that some hot-spot ignition tests be carried out using the new steel
sheet samples.

Phase 1:101
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

7 REFERENCES

Some of the original notes made by the author in carrying out the review have been retained
to give some background to each document. These are incorporated as text below the
reference.

[1] Hagenruth Lightning stroke damage to aircraft AIEE Transactions 68, 1949

[2] Zischank, Fisher et al Simulation of lightning continuing current effects on metal


surfaces.
Very relevant, much data.

[3] Costigan, Burrows Puncture threshold tests to a range of materials, UKAEA Culham
(Internal Report) July 1990
Tests to a variety materials, results for steel mainly presented on the thinner ones,
where puncture occurred. For aluminium only 3 mm was punctured by 200 C, 4 mm
was not at 215 C, but was molten at rear, i.e. zone penetrated. Highlights danger
of simply specifying charge transfer on these time scales. Require above a minimum
time, and above a minimum current. Tests to steel were only on thin panels of 1 mm
or so.

[4] Kern, Zischank Melting effects on metal sheets and air termination wires caused by
direct lightning strokes. ICLP 1988 (Graz).
Noted that loss of material is dependent on polarity, although assume this is for
smaller gaps. For Q=200 C, aluminium punctures up to 3 mm, stainless, brass, copper
up to 2 mm.
Steel 4 mm, Cu 5 mm, Al 7 mm. Stainless steel sheets 4 mm thick show rear surface
discoloration T>400 ºC.

[5] Mariani, Rodriguez Controlling the risk of fire caused by lightning in hydrocarbon
storage tanks ICLP 2000 (Rhodes)
NFPA 780, IEC 1024-1 state apparently that self protected needs 4,8 mm skins. This
report presents a mathematical model, and although this suggests that 4,8 mm OK
it is a confused argument, based on resistance. Highlights the difficulties of corrosion
reducing the thickness of the skins.

[6] Kern The heating of metal sheets caused by direct lightning strikes. ICLP 1990
(Interlaken)
Similar to Kern's Cocoa report ref [7]. This one included effects of insulating internal
layers. However 4mm Steel @200C gives only 400 ºC, but this is at 10 mm radius.
Gives steel types V2A as 960 and St37 as 1100 in the maximum, achieved in perhaps
10 seconds.

[7] Kern Simulation and measurement of melting effects on metal sheets caused by
direct lightning strikes. ICOLSE 1991 Cocoa Beach.
Looks at some experimental simulation aspects, such as electrode spacing. Then,
aluminium up to 3 mm can be punctured, steels up to 2,5 mm, for 200 C. Looks also
at temperature rise, but difficulties experienced with speed of response. Notes how
insignificant Joule heating is compared to Long Duration effects. 4mm steel reaches

Phase 1:102
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 1

940 over 3 mm and 550 over 10 mm radius.

[8] Brocke, et al The effects of long duration lightning currents and their simulation ICLP
2000
Also 2001 ICOLSE. Modelled diagrams showing difference between thin sheets and
thick sheets. Thin sheets undergo much wider damage because of less cooling effect
by conduction. Notes that Joule heating is not so significant.

[9] Nagai Current injection experiment into a metal sheet ICLP 1992 (Berlin)
Notes the effects of surface water, explosive overpressure of an arc of 70 kA created
through fluid can be sufficient to puncture 0,4 mm of steel. Could be of interest in
the tests to submerged shunts.

[10] Uhlig, Gondot et al Experimental simulations of lightning impacts on aeronautical


structural materials
Concerned mainly with effects on aircraft aluminium, but has interesting
measurements on forces and on arc voltages

[11] Haigh Hot-spot ignition characteristics of Aviation/Fuel/Air and Ethylene/Air mixtures


ICOLSE 1992 (Atlantic City)

[12] Aviation Rulemaking Advisory Committee, Task Group 2, Explosion supression,


www.fire.tc.faa.gov/pdf/TG2.pdf.

Phase 1:103
API/EI RESEARCH REPORT

VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS


FOR ABOVE GROUND HYDROCARBON STORAGE TANKS

PHASE 2
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

ACKNOWLEDGEMENTS

This suite of reports has been produced by Culham Electromagnetics and Lightning (Culham) at
the request of the American Petroleum Institute (API) and Energy Institute (EI). API and EI gratefully
acknowledge the input into these reports during their development.

Tests contained in these reports were conducted by:


Chris Chessum
Stephen J Haigh
Philip G Leichauer

Inductances and resistances were put together with assistance from Brian J C Burrows as consultant
to Culham.

The first series of tests were carried out over two days in February 2007. On the second day tests
were witnessed by Mark Scanlon (EI) and Sonia Quintanilla (EI). The second series of tests were
carried out over two days in August 2007.

The API RP 545 Task Force and EI’s Electrical Committee provided technical direction to the project,
and reviewed and commented upon drafts of the reports.

At the time of publication, the API RP 545 Task force comprised:


Louis Barrios Shell Global Solutions (US) Inc.
Chip Breitweiser HMT Inc.
MarkGoodman Jacobs Engineering Group Inc.
Paul S. Hamer Chevron Products Company
Bruce Kaiser Lightning Master Corporation
Joe Lanzoni Lightning Eliminators & Consultants, Inc.
Victor Minak ExxonMobil Research & Engineering
Rick Mondler ConocoPhillips
George L. Morovich TEMCOR
Philip E. Myers Chevron Corporation
Art Neubauer Arseal Technologies LLC
Gordon Robertson API
Marilyn Shores Explorer Pipeline Company
Richard J. Virgilio Foster Wheeler USA Corporation

At the time of publication, the EI's Electrical Committee comprised:


Phil Carpenter Chevron Limited
Mel Cockerill Total
Duncan Crichton BP
Robert Denham Health and Safety Executive
Jeff McQueen Shell U.K. Oil Products Limited
Geoff Fulcher F.E.S. (Ex) Limited
Kevin Hailes BP
Terry Hedgeland Consultant
Darren Hughes Petroplus Refining and Marketing Limited
Tom Ramsey (Chair) ExxonMobil
Paul Taylor British Pipeline Agency Limited
Stephen Wilkinson ConocoPhillips

Project coordination was carried out by, in order of involvement:


Mark Scanlon (EI)
Sonia Quintanilla (EI)
Andrew Sykes (EI)

Technical editing was carried out by Andrew Sykes (EI).

Phase 2:i
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

EXECUTIVE SUMMARY

1 INTRODUCTION

In API/EI Research report: Verification of lightning protection requirements for above ground
hydrocarbon storage tanks: Phase 1 a review was made of present installation practices, and
their effectiveness in prevention of lightning induced ignitions. It was based on the review of
two UK refineries, discussions with operators, as well as general experience of the industry.
The lightning susceptibilities of installations in general were also discussed in this phase of
the investigation.
The significant findings from the investigation can be summarised as follows:

1.1 SOLID VENTED TANKS

Were considered relatively well protected from ignition hazards due to puncture or hot-spot.
This was discussed in more detail in EI-TN1-03 Review of burn-through and hot-spot effects
on metallic tank skins from lightning strikes. It was concluded that for steel skins >4 mm
puncture was less likely, and although hot-spots might occur, for steel skins >5 mm they would
probably not be hazardous. Any vents or other outlets can present a risk of lightning ignition
propagating back into the tank, and this may be the case even if suppressors or pressure-
vacuum vents are used (because the overpressure from lightning may defeat these).

1.2 GEODESIC TANKS

Were not studied in detail, but ignition hazards were considered more likely if lightning
attaches to a geodesic roof which is of aluminium panels <2 mm thick, as aluminium is readily
punctured. There may also be the possibility of sparking at bolted connections, although
puncture would be the greater risk.

1.3 FLOATING ROOF TANKS

Are recognised by industry as being susceptible to rim fires following lightning strikes. It is
clear that shunts around floating roof tanks rarely make good metal to metal contact with
the tank wall. This is especially true for crude oil tanks where a deposit tends to build up
between the shunt and the wall. Preliminary testing showed that sparking at such shunts is
inevitable, if lightning strikes either the floating roof, or the shell rim. A further programme
of testing was carried out to study effects at shunts, and at submerged shunts. These are
summarised in section 4, with the full test results in Cul/LT-0401 Investigative tests on the
lightning protection of submerged shunts with parallel roof bonding cables.

Phase 2:ii
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

2 SUMMARY OF CURRENT FLOW AND SITES OF ARCING IN


FLOATING ROOF TANKS

In order to understand how lightning might affect a facility, and therefore be able to protect
against it, it is essential to understand how the lightning current will flow. The fast high
energy lightning current will not follow a single bond path, but will generally spread out
between as many paths as possible.
If we look at a generic floating roof design we can see what this means (Figure 1).

Figure 1 Typical floating roof tank installation of primary seal, secondary seal, and
shunt. The shunt is intended to help provide electrical continuity from the roof to
the shell, but in practice any other fortuitously touching or nearly-touching contacts
will also carry current.

Shell
Shunt
Secondary seal
Roof bonding cable

Primary seal fabric

Metallic shoe

Floating roof

Weight (or spring) to hold


shoe against the shell

In the event of a strike to the roof, or to the rim, the current will spread out and divide
between the available paths, even if it has to flow across poor contacts. Interfaces such as
shunts and metallic shoes are inevitably going to generate arcing, because it is not possible
to provide a good bond there.
Arcs formed around the metallic shoe between the primary and secondary seal will
occur in a vapour space that may be flammable. A way to lower this threat is to provide
isolation in the pusher plate or pantograph arrangement so that this cannot be a current
path. This isolation should be designed to hold off tens of kV, and so a flashover distance of
at least 75 mm/3" should be maintained.

2.1 ARCING AT SHUNTS

Even if the arcing within the secondary seal can be prevented, there still remains the threat of
arcing at the external shunts. Arcs at this location will be a threat if there is flammable vapour
present this far above the secondary seal. This seems less likely in normal circumstances;
a bigger threat is that showers of sparks generated at the shunt can then fall behind the

Phase 2:iii
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

secondary seal, where flammable vapour is more likely to be present.


Two possible approaches have been studied for mitigating shunt arcing effects:
−− use of roof bonding cables, and
−− use of shunts submerged beneath the fluid surface.

A selection of the findings from the experimental tests are included in section 3, to provide a
rationale for the recommendations proposed.

Phase 2:iv
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

3 SUMMARY OF TESTING STUDYING SHUNT ARCING EFFECTS


AND EFFECTS OF ROOF BONDING CABLES

The installation of roof bonding cable(s) is beneficial for dissipating static charges, but also
has significant benefit for lightning protection. However roof bonding cables will not provide
absolute protection, and the degree of protection they provide will depend on the tank
design. To explain this we have to also understand some properties of lightning currents.

3.1 THE LIGHTNING THREAT

The lightning waveform comprises different current components which behave differently
both in the way they flow, and in the degree of arcing and sparking which they cause.
Figure 2 shows typical current peaks and durations for different components of a
severe lightning strike. Note that the waveform shown in Figure 2 is distorted for clarity.
The fast component (200 kA, 100 µs) is actually of x 1 000 higher amplitude than the long
duration currents, but the long duration components last x 10 000 longer. So this schematic
plot hides the vast differences in scale of the two currents. Because of this difference in scale
the different components behave differently, both in the way they flow and in the risk they
pose.

Figure 2 Typical components of a lightning attachment. The fast high energy


component contains the very high peak current and most of the energy, but the
long duration component current presents the risk of hot-spots and melt-through
on tank skins, and also the risk of sparking.

If we tried to drive the fast current through a bonding cable, the cable’s high frequency
impedance (inductance) would give a voltage drop from roof end to shell end of perhaps
a few million V. So of course this doesn't happen; breakdowns will occur at much lower
voltages at several locations around the roof rim. Current then flows across glancing contacts
such as shunts, as well as pusher plates etc. if they are available paths.
The long duration currents are essentially DC currents, and are not deterred by the
inductance of the bonding cable. Therefore most of the long duration current could flow
through the roof bonding cable(s) if the resistance were low enough.

Phase 2:v
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

Figure 3 Typical parts of the current waveform flowing through shunt, and low
resistance cable

Spreads out to flow mainly


through shunts, pushers etc

Flows mainly through roof bonding


cable if it is <50 mΩ or so.

As there are many shunts in an actual tank installation, the expected effect in practice is that
the fast component will divide between many shunts, whilst the long duration component will
tend to take the lowest resistance path. In the worst case all of the long duration component
could flow through one shunt.
Because of this we might expect a shunt to have to carry up to:
1. A fraction of the expected severe strike fast components i.e. 10 kA-20 kA out of the
full 200 kA.
2. A relatively large proportion of the long duration components if not all (200 C, i.e. 400
A average for 0,5 s).

The purpose of the tests in this investigation was to demonstrate the potential hazards from
these different parts of the waveform, and to see how the hazards could be reduced. In
particular it was to see how shunts could be made more effective and less spark prone (in
conjunction with roof bonding cables).

3.2 BRIEF RESUME OF TEST RESULTS

3.2.1 Ignition hazards from arcs

Initial tests looked at the sparking effects of the different current components in a lightning
strike.

Photograph 1 Shunts carrying Photograph 2 Shunts carrying


fast component (16 kA) only. slow component (1,5 kA).

Phase 2:vi
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

In Photographs 1 and 2 there is an arc at the contact point between shunt and shell, which
may ignite flammable vapours, if they are present, at this location. However, the shunt is
normally in a relatively ventilated region outside the secondary seal. A bigger threat than the
arc can be the burning spark particles, and the difference in such sparking in the photographs
is striking. The faint, straight lines in Photograph 1 are small, fast moving sparks. Due to their
lower heat capacity and rapid speed they pose less of an ignition hazard than the slower,
larger sparks shown in Photograph 2.
Photograph 3 shows flammable vapours behind a simulated secondary seal being
ignited by such sparks.

Photograph 3 Sparks from the slow components resulting in an ignition of flammable


vapours. A gas was fed into a trough below the shunt to simulate a secondary seal gap.

It is worth noting that a lightning attachment to the rim of a tank shell will also result
in copious sparking falling down from the rim towards the secondary seal. Whilst a tight
secondary seal is good protection against this sparking being an ignition hazard, there are
other possibilities to reduce the hazard from this (such as attracting the lightning to locations
other than the rim, e.g. air terminals, walkway, etc).

3.2.2 Effects of roof bonding cables

Several tests were carried out in which the arcing at the shunt/shell interface was monitored
and recorded as a light signal. Examples from these results are given in Figure 6 for two
different test configurations. In each case a bonding cable of 100 µH inductance was
connected to bypass the shunt (this is roughly the inductance of a 130 m long cable), and
the resistance of the cable was varied from 40 mΩ to 210 mΩ.

Phase 2:vii
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

Figure 6 Current against light at shunt/shell interface for two different bonding
cable configurations (both signals –ive going). The arc is rapidly extinguished in the
right hand case where a bonding cable with lower resistance is used.

C ulham Lightning
C ulham Lightning
LIG HTNING panel diagnos tics -C ulham
LIG HTNING panel diagnos tics -C ulham 30-Aug-6 100
30-Aug-6 100

0A 0
0

-100
-100

-200 -200
P lot 1 P lot 1
Current A

-300 -300
E TE S T~3\OIL\D254\1S LOB A~1.WF T E TE S T~3\OIL\D247\1S LOB A~1.WF T
-400 sca ling 1 = x 1.000 -400

V
sca ling 1 = x 1.000
V

-500 -500

-600 -600

-700 -700

-800 -800

900 A -900
0 20 40 60 80 100 120 140
-900
160 0 180 20 40 60 80 100 120 140 160
Millis econds Millis econds

0. 05 0

0 0
-0. 05

P lot 2 -0. 05 P lot 2


-0. 1
E TE S T~3\OIL\D254\2LAR C _~1.WF T -0. 1 E TE S T~3\OIL\D247\2LAR C _~1.WF T
Light signal

sca ling 2 = x 1.000 -0. 15


sca ling 2 = x 1.000
-0. 15
V

-0. 2

V
-0. 2

-0. 25 -0. 25

-0. 3 -0. 3

-0. 35 -0. 35

-0. 4
0 20 40 60 80 100 120 140 -0. 4 160 180
Millis econds 0 20 40 60 80 100 120 140 160
Millis econds

210 mΩ 40 mΩ

In Figure 6, in the left hand case a stable arc continues at the shunt/shell interface for most
of the test waveform. In the right hand case the arc rapidly extinguishes. The crucial factor
is that the resistance of the cable has to be low enough to carry 400 A whilst developing a
voltage less than 14 V. If the voltage falls below this value at the shunts then any arcs will
extinguish.
To satisfy this, the bonding cable and its terminations should therefore have a total
resistance of <35 mΩ.

3.2.3 Tests to submerged shunts

In the most definitive tests, arcs were formed from 20 mm to 450 mm below the surface of
the fluid.
With shallow submersion (20 mm) oil, an eruption of fluid and bubbling occurred,
and sometimes sparks and a jet of sooty flame.
With submersion of 300 mm or greater under oil no ignition hazards were observed,
even without bonding cables (up to a high level of 44 kA [fast], 2 kA [slow]).
If connections straddle the fluid line then arcing is more likely to occur at or above
the fluid line, as the fluid is a good insulator.

Phase 2:viii
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

Photograph 4 Burning vapour emitted from oil surface, after test to shallowly
submerged shunt.

So if all current paths (other than a roof bonding cable) can be constrained to >300 mm
below the fluid line then the hazard is reduced.

Phase 2:ix
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

4 SUMMARY OF RECOMMENDATIONS ON THE USE OF


SHUNTS AND ROOF BONDING CABLES

Sparking cannot be eliminated at shunts, but its intensity can be reduced by low resistance
roof bonding cables. Submerging shunts beneath the fluid line can also provide improved
protection.
It may be that the shunts are not the only current path, so that some other fortuitous
contacts may exist near or above the fluid line (shoe, pusher plates, etc). In these cases any
arcing is much more likely to occur outside the fluid, and would present a hazard. In such
cases shunts or roof bonding cables will not provide any significant protection, as some
current will always flow through the primary seal.

4.1 EXTERNAL SHUNTS (NOT SUBMERGED) AND ROOF BONDING CABLES

For external (not submerged) shunts arcing caused by the fast component is unavoidable and
is not significantly attenuated by a roof bonding cable. However, the fast component arcing
is primarily a hazard near its location at the shunt/shell interface.
Arcing caused by the long duration component is more hazardous as showers of
burning particles are ejected, and these can fall down onto or behind the secondary seal. This
particular arcing threat may be reduced if low resistance roof bonding cable(s) can be fitted.
The arcing is extinguished if the resistive voltage drop down the cable(s) is <14 V. Since the
average value of long duration current is considered to be 400 A (see CUL/LT-0235 Review of
tank base earthing and test current recommendations) then this would require the bonding
cable(s) to have a combined resistance of <35 mΩ, and it would be prudent to aim for 20
mΩ as a maximum (see CUL/LT-0373 Lightning simulation testing to determine the required
characteristics for roof bonding cables on external floating roof above ground storage tanks).
The common industry practice of using more than one bonding cable (for redundancy) will
also help in this regard. It will also help current distribution if bonds are made from the roof
to two or more well separated locations on the shell rim.
For the bonding cable it would be possible to use, for example, AWG0 gauge copper
wire (8,3 mm diameter), which has a rated resistance of 0,29 Ω per kilometre. A 50 m long
roof bonding cable would then have a resistance of about 15 mΩ. However, the material
from which the bonding cable is constructed is not relevant, as long as the appropriate
resistance can be achieved and maintained without corrosion issues.
Clearly the resistance of any bonding connection must be kept low. It is practical to
achieve values of well below 1 mΩ, but care must be taken to ensure that the joint is well
sealed and does not corrode.

4.2 SUBMERGED SHUNTS

If shunts can be submerged, then this is a good solution, and if the contact point is submerged
below 300 mm, an ignition source at this interface is less likely, even without roof bonding
cables. High level tests were carried out in this configuration (using 44 kA fast component, 2

Phase 2:x
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

kA slow long duration component) without ignition. Shallower submersion depths can result
in burning particles and flame being ejected by slow current arcs, but the use of roof bonding
cables reduces this threat. So in summary:
Fast currents submersion at >100 mm gives improved protection
Slow currents submersion at >300 mm gives improved protection
submersion at 100 mm to 300 mm would require bonding cables

4.3 PROPOSED RECOMMENDATION

Eliminate any possibility of arcing at


primary seal/shell interf ace
No Higher risk of sparking between
Ensure isolation to prevent current flowing primary and secondary seals.
along in pantograph, spring, pusher plate Options A or B will make only a
paths, etc. small reduction in risk.
Isolation should be designed to hold off
tens of kV, and so a flashover distance of at
least 3”/75 mm should be maintained.
Adequate isolation provided?
Yes

O ption B - A dd external shunts and


roof bonding cable
O ption A - Use submerged shunts
Traditional shunt designs are acceptable,
Shunts should be submerged by at least although they do tend to ride above the
300 mm, otherwise they can have a similar waxy deposits on crude oil tanks, and have
design to external shunts. An external roof no direct contact with the shell.
bonding cable should still be used to For the bonding cable a resistance of 25 mΩ
guarantee bonding for static electricity. or less should be achieved from end to end,
including end terminations. It would be
possible to use for example AWG0 gauge
copper wire (8,3 mm diameter), which has a
rated resistance of 0,29 Ω per kilometre.
Using two or more parallel cables will help
reduce the resistance, and if used they
should go to well separated locations on the
shell rim. Ladders may be bonded in to help
achieve the same requirement.
Ensure that the terminations are well
sealed and will not corrode.

In Option A, all sparking is in a less hazardous location. Option B still allows a possibility that
some sparks could fall behind the secondary seal, although the use of the roof bonding cable
significantly mitigates the risk. The larger the distance between the shunt contact point and
the secondary seal, the better.

Phase 2:xi
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

CUL/LT- 0373

LIGHTNING SIMULATION TESTING TO DETERMINE THE REQUIRED


CHARACTERISTICS FOR ROOF BONDING CABLES ON EXTERNAL
FLOATING ROOF ABOVE GROUND STORAGE TANKS

CONTENTS

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.1 Rim protection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.2 Roof bonding cable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2 Test arrangement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.1 The gas trough and flammable gas mixture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.2 Simulated roof bonding cable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.3 Optical fibres . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.4 Threat levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

3 Test results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10


3.1 Low threat level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
3.2 Medium threat level 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.3 Medium threat level 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.4 Results with spacer between shunt and shell (slow currents only) . . . . . . . . . . . . . . . . . 18
3.5 High threat level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.6 Relevance of the arc extinction voltage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.7 Comparison with real tank designs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.8 Probabilities of ignition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

Annexes:

Annex A Notes on ignition probabilities arising from lightning strikes to tanks . . . . . . . . 26
A.1 Strike rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
A.2 Risk of a particular threat level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
A.3 Chance of damage at particular threat level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
A.4 Chance of exceedance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
A.5 Reality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
A.6 Suggested chance of exceedance from testing so far . . . . . . . . . . . . . . . . . . . . . . . . . . 28

Figures:
Figure 1 Test arrangement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Figure 2 Butane/air mixing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Figure 3 Typical waveform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Figure 4 Comparison of current through shunt, and light emitted from arc at
shunt/shell interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

Phase 2:1
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

Figure 5 Further comparison of current through shunt, and light emitted from arc . . . . . . . . 12
Figure 6 Example of results from medium threat level 2 tests . . . . . . . . . . . . . . . . . . . . . . . . 13
Figure 7 Applied waveform and photomultiplier record for combination A . . . . . . . . . . . . . 15
Figure 8 Applied waveform and photomultiplier record for combination B . . . . . . . . . . . . . . 16
Figure 9 Recordings for no bonding cable, and 100 mΩ, 100 μH bonding cable . . . . . . . . . 17
Figure 10 Applied waveform and photomultiplier record for combination C . . . . . . . . . . . . . 18
Figure 11 Applied waveform and photomultiplier record for combination D . . . . . . . . . . . . . 19
Figure 12 Applied waveform and photomultiplier record for combination E . . . . . . . . . . . . . . 20
Figure 13 Applied waveform and photomultiplier record for combination F . . . . . . . . . . . . . . 21

Tables:
Table 1 Test levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Table 2 Bonding cable parameter combinations for different tests . . . . . . . . . . . . . . . . . . . 14
Table 3 Summary of results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

Photographs:
Photograph 1 Shunts carrying fast component (16 kA) only . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Photograph 2 Shunts carrying slow component (1,5 kA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Photograph 3 (Combination A 2 mΩ, 14 μH) High threat level, no spacer. . . . . . . . . . . . . . . . 15
Photograph 4 (Combination B, 6 mΩ, 40 μH) Medium threat level 2, no spacer. . . . . . . . . . . . 16
Photograph 5 (Combination C, 10 mΩ, 100 μH) Medium threat level 2, 1 mm spacer. . . . . . . 18
Photograph 6 (Combination D, 40 mΩ, 100 μH) Medium threat level 2, 1 mm spacer. . . . . . . 19
Photograph 7 (Combination E, 110 mΩ, 100 μH) Medium threat level 2, 1 mm spacer. . . . . . 20
Photograph 8 (Combination F, 210 mΩ, 100 μH) Medium threat level 2, 1 mm spacer. . . . . . . 21
Photograph 9 An ignition caused by the slower sparks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

Phase 2:2
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

SUMMARY

This report describes testing carried out as part of Phase 2 of API/EI Verification of lightning
protection requirements for above ground hydrocarbon storage tanks. The report covers the
following aspects:
1. Level of protection provided by generic roof bonding cable designs.
2. Requirements for actual roof bonding cable designs.

The aims of the tests were to determine how effective different roof bonding configurations
would be in reducing sparking at the shunt/shell interface. The tests inject current across a
shunt/shell interface, with a parallel path for the current, representing different roof bonding
cable configurations. This approach is shown schematically below:

Roof bonding cable of defined


Inductance L, Resistance R

Shell
Shunt

Current
in-out

Such a roof bonding cable will carry only intermediate and long duration component current.
The fast components will not flow through the cable due to its high inductance, and instead
take the low inductance path through many of the shunts. Arcing at the shunts is then
inevitable, and the arc may ignite any flammable vapour/air mixture present at the level of
the shunts.
The arc will also generate hot burning particles, which can fall down behind the
secondary seal, a region where a flammable vapour/air mixture is more likely to be present.
The aim of this investigation is to minimise this threat from hot burning particles.
The effectiveness of the roof bonding is gauged by:
1. The ignition or not of a trough of butane-air mixture located under a secondary seal
which was away from the tank shell by 30 mm (1,2”).
2. Photographs of the shunt sparking which occurred in the tests.
3. Photomultiplier signals of the light from the arc at the shunt/shell interface.

3. gives a good indication of the presence of a current flowing through the shunt, which
quantifies the effectiveness of the bypassing roof bonding cable.

Phase 2:3
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

In all, 120 test shots were fired on the main bank and 55 on the test bench. The various tests
and their findings are summarised below:
1. Using the flammable gas below a shunt shell interface in initial tests, it was found
that the fast component of lightning at a current of 16 kA did not cause ignition of
the gas (10 tests), but the first test using the long duration currents (50 A/200 ms)
did cause ignition.
2. Further tests carried out with the same shunt configuration but with simulated roof
bonding cables added gave inconsistent results. Some ignitions occurred even with
bonding cables added.
3. Bench tests were carried out to study the ability of different roof bonding cable
configurations to carry the long duration currents. A parallel roof bonding cable of
100 mΩ/100 µH effectively extinguished the arc at the shunt within a few tens of
ms.
4. Bench tests were repeated but using higher test levels available in the main facility.
Results showed that the consistent extinction of the arc only occurred if there was
a 0,5-1 mm gap maintained between the shunt and the shell. The arc generally
extinguished when there was 14–40 V drop along the bonding cable. However if the
shunt made intermittent contact with the tank shell then continued arcing appeared
to occur even with lower voltage drop along the roof bonding cable. In this case the
ferocity of the sparking was reduced when a bonding cable was used, as less current
was traversing the arc.

In conclusion:
1. If the shunts make contact against a thinly rusted tank wall the provision of a roof
bonding cable does not extinguish the sparking at high long duration currents (400 A),
but helps reduce the level of sparking. Variations in contact conditions are a big
factor in the level of sparking. For lower level long duration currents (50 A or so), the
bonding cable eliminates the sparking.
2. The voltage at which the arc goes out determines the maximum allowed resistance
of the roof bonding cable system. 14 V is considered a workable minimum voltage,
below which an arc cannot normally be maintained. At low level long duration
currents this was confirmed, but for higher level currents at the shunt (>200 A) the
arc could be maintained as low as 6 V. The 14 V extinction will occur reliably if a
narrow spacer keeps a small (1mm) gap open between the shunt and the shell.
3. To simply retrofit by adding roof bonding cables, it is proposed that two or three
cables are used, having a maximum in parallel resistance of 15 mΩ.
4. With 1 mm spacers incorporated as part of the shunts, a maximum resistance of 35 mΩ
would be acceptable for the bypass cable(s).

Phase 2:4
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

1 INTRODUCTION

This report describes the testing of simulated roof bonding cables, to demonstrate their
capacity to reduce the threat of lightning induced ignition hazards for above ground floating
roof storage tanks.
The potential problem is the arcing and sparking which could occur when lightning
currents flow from the tank floating roof to the shell. Protection against this threat is normally
addressed by the use of shunts around the edge of the roof, and from bonding cables. These
combined devices provide some protection of the tank from separate components of the
strike.

1.1 RIM PROTECTION

This is in the form of sliding metal shunts. These form a low inductance path (as there
are many and they are short), through which the fast component passes. A good metal to
metal contact should minimise creation of arc or sparks. This situation rarely exists as this
would require clean metal and substantial compressive pressure to hold the two conducting
surfaces together. Also hindering the conduction is an often present thin layer of rust and/
or waxy film. As a consequence an arc will form with the production of large numbers of
small fast sparks emitted from the shunt/shell interface. If a flammable vapour/air mixture is
present at the shunt level, the arc may ignite it. However, burning metallic particles may also
fall down behind the secondary seal, where a flammable vapour/air mixture is more likely to
be present.

1.2 ROOF BONDING CABLE

−− Owing to the high inductance of the bonding cables (because of their length) they
do not carry any significant part of the fast component and arcs will form at shunts.
However, if the cable resistance is sufficiently low it will carry the majority of the long
duration components. These current components have slower characteristics so they
are not significantly impeded by the inductance of the cables.
−− If the long duration components pass through the shunts as an arc, the high charge
transfer through the arc causes much more heating of the interface than the fast
component. This arc behaves like a welding arc, and produces the large slow moving
sparks which are known to present a high probability ignition hazard. The purpose
of the roof bonding cable is to short out these components and effectively reduce/
eliminate this potential hazard.

The work carried out for this report aimed to examine the hazards posed by differing lightning
threat levels and identify the maximum inductance/resistance that a roof bonding cable can
have whilst still significantly reducing the ignition hazard.
Note that this is a separate question from that of tank earthing. The current flow
across the shunts will be independent of whether the tank itself is well or poorly earthed.

Phase 2:5
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

2 TEST ARRANGEMENT

The general arrangement for the tests is shown in Figure 1. A segment of a tank shell is
mounted in the arc rig, and connected to the ground output terminal of the lightning bank.
The shunt was clamped to an insulating plate and electrically connected to the bank live
output line. The shunt position on the shell could be adjusted. The set up is visible in the
results photographs presented in section 3.
Where a bonding cable was used it was bonded directly to the output terminals
of the lightning bank close to the shunt/shell interface, so that it provided a parallel path
with the tank shunt/shell path. Several cables of different parameters were available for
evaluation.
Beneath the shunt/shell interface, a trough was located which could be filled with a
defined butane/air mixture. Sparks from the arc dropping into the trough may ignite the gas,
thus helping to quantify the ignition hazards for flammable gases in the region of, or below,
the secondary seal. The flammable gases were not present at the height of the shunt; ignition
of such gases by the arc would have inevitably occurred.

Figure 1 Test arrangement

Roof bonding
cable
Tank shell segment

Shunt

Roof bonding cable

Simulated
secondary seal

Gas trough

2.1 THE GAS TROUGH AND FLAMMABLE GAS MIXTURE

A mixture of butane and air filled a trough below the shunt/shell interface for some of
the tests; at its top a moveable panel simulates a gap between the secondary seal and the
tank wall. Normally such a gap would not be present in a real installation but here a worst

Phase 2:6
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

case such as an ‘out of round’ tank is simulated. The gas was a 1:1 stoichiometric butane/
air mixture that is a good representation of hydrocarbon vapours that might be present
above fuels. The gas was injected into the trough using a carefully designed nozzle to reduce
turbulence and keep it lying within, as opposed to swirling and mixing with the air above,
the secondary seal (see Figure 2 for a schematic of the mixing station). If a spark of sufficient
energy to ignite the fuel falls into the volume of gas it would ignite and shoot flames out of
the trough, indicating an ignition.

Figure 2 Butane/air mixing.

A spark plug within the trough was periodically fired to ensure the gas mixture was ignitable
by a spark. After such an event or any other event in which an ignition occurred, the trough
was allowed to cool and was then filled with gas again.

2.2 SIMULATED ROOF BONDING CABLE

To simulate the self inductance of a bonding cable, a shorter cable wound to have the
representative inductance was used (see Figure 1). Three values were used, representing
lengths of 9 m, 20 m and 50 m (14 µH, 40 µH and 100 µH respectively). These had their own
intrinsic resistance of a few mΩ. However, resistances of this value may be difficult to achieve
in practice so additional resistances were added to simulate worst cases. 2 mΩ, 40 mΩ, 100 mΩ
and 200 mΩ were produced, but only the three higher values were used.

2.3 OPTICAL FIBRES

One fibre looked directly at the arc whilst a second observed the flux of sparks falling into
the gas trough. These were fed into separate photomultipliers whose output was recorded
on a digital oscilloscope.

Phase 2:7
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

The arc light provides a crucial diagnostic, where light will only be produced in an arc
if current is flowing. Therefore by looking at the light emitted from the shunt/shell interface
we get a good indication of the current distribution between the shunt and the bonding
cable.
The bench tests were used to validate this approach (see section 3.1).

2.4 THREAT LEVELS

The test proposal defined the test levels as typically:


Fast component 11 kA at least (fast, high energy current)
Long component 80 A for ~0,5 s (slow, long duration currents)

These tests were to be applied together in representative tests to installations. However, in


order to understand what mechanisms were at work, many series of bench tests were carried
out using different pulse generators.
What is established in these tests is that the fast current inevitably flowed through
the shunts, whilst the long duration currents could be relatively easily diverted through a
bonding cable. When bonding cables are used, an important part of the lightning waveform
for the sparking at shunts is the initial part of the long duration component, which has a
relatively high current over a time scale of a few milliseconds.
We also take account of this type of current in these tests, as the ‘intermediate’
component. The currents and time scales attributable to these aspects of a severe strike are
shown schematically below:

So pulses applied during these tests are not as originally defined, but are intended to
investigate the worst parts of the lightning waveform for shunt sparking.
The main goal is to ensure that the long duration currents are all conducted through
the bonding cable. A value of 400 A is assumed for this component; however, it is not a
parameter which receives much discussion in the specifications for lightning test waveforms.
Some discussion on the test levels, and on the probabilities is given in Annex A. The IEC Level
1 protection for the long duration currents is intended to cover 98 % of strikes and is 200 C
in 0,5 s (i.e. an average current of 400 A).
In Zundl Koordinierte Messsungen von Blitsstroemen und Ihrer elektromagnetischen
Felder an einem Fernmeldeturm measurements on the Peissenberg tower in Germany
suggested long duration peak currents exceeding 1 kA in 10 % of cases, and with the
average current exceeding 500 A also in 10 % of strikes.
By assuming a value of 400 A as the intended protection level this covers perhaps 85 %

Phase 2:8
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

of strikes, rather less than is implied from IEC 1024-1 1990 Protection of structures against
lightning – Part 1: General principles.
The following test levels were typically applied in the different test series:

Table 1 Test levels.

Threat Level Fast Intermediate Long duration


High 24 kA 2 kA 320 A to 0 A
Medium 1 16 kA 4 kA 600 A to 0 A
Medium 2 14 kA 1,5 kA 280 A to 0 A
Low (bench tests) n/a 800 A 40 A square wave

For all but the ‘Low’ threat level in Table 1, the main lightning simulation capacitor bank at
Culham was used to generate the test waveforms. This gave an LR decaying current waveform
for the long duration component. The generator used for the ‘Low’ threat level was a smaller
bench-top generator. This gave a combined intermediate-long duration waveform, in which
the latter part was constant current.

Phase 2:9
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

3 TEST RESULTS AND DISCUSSION

Tests were carried out as follows:


−− Preliminary tests to compare the threat of fast and long duration currents in igniting
gases in the trough (results in 3.2).
−− Further tests with gas trough studying the effects of different simulated roof bonding
cables (results in 3.3).
−− Low level bench tests looking at current distribution between shunts and simulated
roof bonding cable for intermediate and long duration currents (results in 3.1).
−− Further high level tests looking at current distribution effects (results in 3.3).

The following sections discuss the test results in order of increasing severity.

3.1 LOW THREAT LEVEL

The purpose of the low level bench tests was to investigate the current distribution between
the shunt and its parallel roof bonding cable, for a range of bonding cable parameters. They
were also used to show that the light from the arc could be used as an indicator of current
flowing through the shunt/shell path.
It is known that the fast component flows through the shunts even with a roof
bonding cable present. However, the long duration components could be safely conducted
away using a roof bonding cable. Therefore these tests did not include a fast high energy
component but used a waveshape which included the transition between the fast and
the long duration component, as well as the long duration current itself. The waveform
parameters are those which most usefully show how the current distribution between shunts
and roof bonding cable varies with current rise/fall time.
Figure 3 shows a typical waveform from the test series. The red line is the current
through the whole system, and the blue line is the current through the shunt/shell path. In
this case the resistance of the bonding cable path was 110 mΩ. With the current at the point
when the arc is extinguished being 420 A, the voltage across the arc would therefore have
been around 46 V (see 3.5).
At the initial relatively fast rising current up to 100 µs, most of the current flows
through the low inductance shunt. Subsequently, as the current slows, the shunt current falls
away, and by 1,4 ms the arc extinguishes and all of the current flows in the bonding cable.
This mechanism suggests good protection from the long duration currents, because there is
no arc for the long period of 2 ms to 0,5 s.
However at higher current levels the situation and the protection are less well
defined.

Phase 2:10
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

Figure 3 Typical waveform.

Applied current

Shunt current

Arc extinguishes

Figure 4 Comparison of current through shunt, and light emitted from arc at shunt/shell
interface.
C ulham Lightning
LIG HTNING panel diagnos tics -C ulham
31-Aug-6 0. 1

-0. 1

Current through shunt/shell path


P lot 1

OIL\E IOILS ~1\P 13\2LAR C _~1.WF T -0. 2

sca ling 1 = x 1.000


V

-0. 3

-0. 4

-0. 5

-0. 6
0 200 400 600 800 1000 1200 1400 1600 1800
Arc extinguishes
Micros econds

250

200
P lot 2

OIL\E IOILS ~1\P 13\3ITR AN~1.WF T


sca ling 2 = x 1.000
150
Photomultiplier output
100 (light from arc)
V

50

-50
0 200 400 600 800 1000 1200 1400 1600 1800
Micros econds

Figure 4 shows a situation similar to that in Figure 3, with a bonding cable of 110 mΩ and
100 µH. In this plot the current through the shunt is compared with the monitored light from
the arc. This light recorded by the photomultiplier can be seen to end very close to or just
after the current ends. Presumably the slight delay is due to some sparks still remaining after

Phase 2:11
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

the arc event and in higher energy situations, the radiated light from hot-spots. A similar
result for a very low resistance bonding cable is shown in Figure 5.

Figure 5 Further comparison of current through shunt, and light emitted from arc.

C ulham Lightning
LIG HTNING panel diagnos tics -C ulham
31-Aug-6 0. 05

-0. 05

P lot 1
-0. 1 Current through shunt/shell path
\OIL\E IOILS ~1\P 4\2LAR C _~1.WF T -0. 15
sca ling 1 = x 1.000
V

-0. 2

-0. 25

-0. 3

-0. 35

-0. 4
-1 0 1 2 3 4
Millis econds

180

160

140
P lot 2

\OIL\E IOILS ~1\P 4\3ITR AN~1.WF T 120

sca ling 2 = x 1.000


100

80
V

60

40
Photomultiplier output (light from arc)
20

-20
-1 0 1 2 3 4
Millis econds

3.2 MEDIUM THREAT LEVEL 1

These tests were initially carried out to compare the incendivity of sparks due to the fast
component and the long duration currents. They demonstrated no ignitions with 10 pulses
of the fast component at only (16 kA) and an ignition the first time the long duration current
(~50 C) was applied. The sparking produced by these types of test is very different, with the
slow components producing the large slow burning particles which were believed to present
more of a threat (see Photograph 1 and 2). This test suggested that this was indeed the
case.

3.3 MEDIUM THREAT LEVEL 2

Most tests were carried out with a combination of all current components as indicated in
Table 1. In the results plots for these tests the fast component is not shown, as it occurs on
too fast a time scale and always produced arcing at the shunt.
An example result is shown in Figure 6. The applied waveform has an intermediate
component of 850 A at 5 ms with a long duration current decaying from an initial value of
300 A. The arc goes out when the total current is 140 A at 100 ms. The roof bonding cable
had a resistance of 110mΩ which gives a voltage across the arc of 15 V. This is close to the
14 V minimum that is normally required to sustain a small arc in air.

Phase 2:12
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

Figure 6 Example of results from medium threat level 2 tests.

Culham Lightning
LIGHTNING panel diagnostics -Culham
30-Aug-6 100

-100

-200
Plot 1
-300
TEST~3\OIL\D2_30\1SLOBA~1.WFT
scaling 1 = x 1.000 V -400 Total current
-500

-600

-700

-800

-900
0 20 40 60 80 100 120 140 160 180
Milliseconds

-0.05

Plot 2
-0.1
TEST~3\OIL\D2_30\2LARC_~1.WFT
scaling 2 = x 1.000 -0.15

-0.2
Arc light at shunt
V

-0.25

-0.3

-0.35

-0.4
0 20 40 60 80 100 120 140 160 180
Milliseconds

Photograph 1 shows the sparks caused by a fast component only (16 kA). This did not have
any roof bonding cable and so all of the current was forced to go through the arc.

Photograph 1 Shunts carrying fast component (16 kA) only.

Phase 2:13
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

Photograph 2 shows the result of applying the slow component only (1,5 kA intermediate
component, followed by a decaying long duration current starting at 280 A). The difference
in the sparking is striking. The faint, straight lines in Photograph 1 clearly show small fast
moving sparks. Such smaller, faster sparks pose less of an ignition hazard than the slower,
larger sparks shown in Photograph 2.

Photograph 2 Shunts carrying slow component (1,5 kA).

Table 2 shows the combinations of resistance and inductance tested for the simulated bonding
cable. It was found that the inductance had little effect upon the arc duration for all but the
highest threat levels; that is, whatever inductance was used, arcing occurred during the fast
component and continued into the intermediate and long. What determined the extinction
of the arc was bonding cable resistance, not inductance. As such, only the largest inductor
was used for the latter tests.

Table 2 Bonding cable parameter combinations for different tests.

  Inductance (µH)
Resistance (mΩ) 14 40 100
2 A - -
6 - B -
10 - - C
40 - - D
110 - - E
210 - - F

Phase 2:14
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

Photograph 3 shows the results of a 24 kA fast component and 2 kA/30 C slow component.
Only the fast type sparks can be seen here and there is no arc during the slow component.
Figure 7 (right hand case) shows the photomultiplier record for this trace, and the left hand
case shows the waveform for the applied intermediate and long duration current. The light
is shown continuing well into the long duration component, but the photomultiplier tube
sensitivity was too high for these first shots, and the light at late time is believed to be
glowing metal rather than arc light. The arc light has driven the tube well into saturation.
For subsequent shots the photomultiplier tube sensitivity is reasonable and remains
at the same setting.

Photograph 3 (Combination A 2 mΩ, 14 µH) High threat level, no spacer.

LIGHTNING panel diagnostics -Culham


100

-100

-200

-300
~1.WFT
-400
V

-500

-600

Figure 7 Applied waveform and photomultiplier record for combination A.


-700

-800

-900
0 20 40 60 80
Milliseconds

LIGHTNING panel diagnostics -Culham


100 0.05

0
0
-100
-0.05
-200
1.WFT -0.1
-300
~1.WFT -0.15
-400
V

-0.2
-500

-600 -0.25

-700 -0.3

-800 -0.35

-900 -0.4
0 20 40 60 80 0 20 40 60 80
Milliseconds Milliseconds

0.05

-0.05

1.WFT -0.1

-0.15
V

-0.2

-0.25

-0.3

-0.35

-0.4
0 20 40 60 80
Milliseconds

Phase 2:15
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

Photograph 4 shows a result with the bonding cable combination B 6 mΩ, 40 µH. The
time base of the plot is much faster than the plot associated with Photograph 3, and the arc
extinguishes at about 300 A (Figure 8).

Photograph 4 (Combination B, 6 mΩ, 40 µH) Medium threat level 2, no spacer.

LIGHTNING panel diagnostics -Culham


100

-100

-200

-300
~1.WFT
-400
V

-500

-600

Figure 8 Applied waveform and photomultiplier record for combination B. -700

-800

-900
0 20 40 60 80
Milliseconds
LIGHTNING panel diagnostics -Culham
100
0.05
0
0
-100
-0.05
-200
1.WFT -0.1
-300
~1.WFT
-400 -0.15
V

-500 -0.2

-600 -0.25

-700 -0.3

-800
-0.35
-900
0 20 40 60 80 -0.4
Milliseconds 0 20 40 60 80
Milliseconds

0.05

-0.05

1.WFT -0.1

-0.15
V

-0.2

-0.25

-0.3

-0.35

-0.4
0 20 40 60 80
Milliseconds

Phase 2:16
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

Figure 9 shows two recordings, the left hand case where no bonding cable is installed, the
right hand case where a 100 mΩ, 100 µH cable is used. There appears to be little difference
between the two. It is assumed that in the left hand case (no bonding cable), a good metal
to metal contact is eventually established through the molten metal at the arc roots, so that
the arc is shorted out. For the right hand case (with bonding cable) the arc does extinguish
at around 12 V, close to the value of 14 V which is generally accepted as the minimum arc
sustaining voltage. Surprisingly, in some cases there appeared to be an arc maintained down
as low as 6 V and the explanation for this is not known, although it could be caused by
intermittent contacts resulting in break arcs.

Figure 9 Recordings for no bonding cable, and 100 mΩ, 100 μH bonding cable.

No bonding cable 100 mŸ/100 µH bonding


cable

So with shunts in reasonable contact with the shell the bonding cable is beneficial, by
reducing statistically the long duration currents flowing through the shunt, but does not
always extinguish the arc completely in the way that had been anticipated.

Phase 2:17
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

3.4 RESULTS WITH SPACER BETWEEN SHUNT AND SHELL (SLOW CURRENTS ONLY)

Having discovered the apparent effect by which the arc appeared to be maintained even
for Voltage <14 V, an artificial gap between the shunt and the tank wall was created usingLIGHTNING panel diagnostics -Culham
100
a small spacer. This ensured that an arc had to be maintained between the shunt and the
0

shell. The modification was made to help understand the arc extinction effects on a pure arc
-100

(rather than what may be an intermittent contact at the shunt/shell with break arcs). No fast
-200

-300
components were applied in these tests.WFT
with spacers. -400

V
-500

-600

-700

Figure 10 Applied waveform and photomultiplier record for combination C. -800

-900
0 20 40 60 80
Milliseconds

LIGHTNING panel diagnostics -Culham


100 0.05

0 0

-100
-0.05
-200 WFT -0.1
-300
.WFT -0.15
-400
V
V

-0.2
-500
-0.25
-600

-700 -0.3

-800 -0.35

-900 -0.4
0 20 40 60 80 0 20 40 60 80
Milliseconds Milliseconds

0.05

-0.05

Photograph
WFT
5 (Combination C, 10 mΩ, 100 µH) Medium threat level 2, 1 mm spacer.
-0.1

-0.15
V

-0.2

-0.25

-0.3

-0.35

-0.4
0 20 40 60 80
Milliseconds

Phase 2:18
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

LIGHTNING panel diagnostics -Culham


100

Increasing the resistance of the cable from 10 mΩ to 40 mΩ slightly lengthened the duration
-100

of the arc and gave a small increase in sparking. -200

-300
WFT
-400

V
-500

-600

Figure 11 Applied waveform and photomultiplier record for combination D. -700

-800

-900
0 20 40 60 80
Milliseconds

LIGHTNING panel diagnostics -Culham


100 0

0 -0.05

-100
-0.1
-200 WFT
-0.15
-300
WFT
-0.2

V
-400
V

-500 -0.25

-600
-0.3
-700
-0.35
-800

-900 -0.4
0 20 40 60 80 0 20 40 60 80
Milliseconds Milliseconds

-0.05

-0.1
Photograph
WFT
-0.15
6 (Combination D, 40 mΩ, 100 µH) Medium threat level 2, 1 mm spacer.
-0.2
V

-0.25

-0.3

-0.35

-0.4
0 20 40 60 80
Milliseconds

Phase 2:19
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2
LIGHTNING panel diagnostics -Culham
100

-100

Increasing the resistance again from 40 mΩ to 110 mΩ also lengthened the duration of the
-200

arc and gave a further small increase


WFTin sparking.
-300

-400

V
-500

-600

-700

Figure 12 Applied waveform and photomultiplier record for combination E. -800

-900
0 20 40 60 80
Milliseconds

LIGHTNING panel diagnostics -Culham


100 0

0
-0.02
-100

-0.04
-200 WFT
-300
WFT -0.06
-400

V
V

-0.08
-500

-600 -0.1

-700
-0.12
-800

-900 -0.14
0 20 40 60 80 0 20 40 60 80
Milliseconds Milliseconds

-0.02

Photograph
WFT
-0.04
7 (Combination E, 110 mΩ, 100 µH) Medium threat level 2, 1 mm spacer.
-0.06
V

-0.08

-0.1

-0.12

-0.14
0 20 40 60 80
Milliseconds

Phase 2:20
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

LIGHTNING panel diagnostics -Culham


100

The final increase from 110 mΩ to 210 mΩ increased the duration of the arc, which continued0

until the applied current fell below 67 A, up to which point (due to the high resistance of the
-100

bonding cable) there would have been an arc voltage of >14 V.


-200

-300
WFT
-400

V
-500

-600

Figure 13 Applied waveform photomultiplier record for combination F. -700

-800

-900
0 20 40 60 80 100 120
Milliseconds

LIGHTNING panel diagnostics -Culham


100 0.05

0
0
-100
-0.05
-200
WFT -0.1
-300
WFT -0.15
-400
V

V
-0.2
-500

-600 -0.25

-700 -0.3

-800 -0.35

-900 -0.4
0 20 40 60 80 100 120 0 20 40 60 80 100 120
Milliseconds Milliseconds

0.05

-0.05

WFT -0.1
Photograph 8 (Combination F, 210 mΩ, 100 µH) Medium threat level 2, 1 mm spacer.
-0.15
V

-0.2

-0.25

-0.3

-0.35

-0.4
0 20 40 60 80 100 120
Milliseconds

Phase 2:21
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

As can be seen in Photograph 9, the slower sparks can ignite flammable gases if present. It
was found that generally, whenever the slower arcing sparks were formed and flammable
gases were present within the trough, an ignition occurred. After this was understood the
gas was no longer used as it impeded the visibility of other events within the test, as well
as limiting the number of tests which could be performed. Even though the long duration
current was the main cause of such sparks, the faster current waveform did cause some
(but far fewer) ignitions. It appears that incendive sparks can still be produced by the fast
components, although more rarely.

Photograph 9 An ignition caused by the slower sparks.

3.5 HIGH THREAT LEVEL

The higher threat level tests were all carried out with the flammable butane/air mixture in the
trough. All of the tests included the fast component but had no gap. As with the other threat
levels, several combinations of bonding cable resistance and inductance were used.
Table 3 shows a summary of the results from these higher level tests. The inductance
and resistance values are those of the bonding cable and the ignition % is the percentage of
test shots that resulted in ignition of the flammable mixture below the simulated secondary
seal. With more testing, the ignition statistics would form a smoother trend but the inferred
trend is nonetheless clear: that the higher the resistance of the bonding cable, the more likely
an ignition would occur.

Phase 2:22
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

Table 3 Summary of results.

Inductance Resistance Gas test shots Ignitions (%)


40 µH 6 mΩ 10 0%
102 µH 10 mΩ 5 20 %
102 µH 40 mΩ 7 28 %
102 µH 110 mΩ 6 16 %
102 µH 210 mΩ 1 100 %

3.6 RELEVANCE OF THE ARC EXTINCTION VOLTAGE

The voltage across the arc at the moment it is extinguished has been noted several times thus
far. In previous tests, as well as in the reported literature, it is noted that whilst you need a
relatively high voltage to initiate a spark (>300 V), you require only a voltage >14 V to maintain
one once formed. Arc welders therefore operate, for example, at voltages of typically 20-25 V. It
would be expected that an arc goes out once the voltage across it drops below such a value,
and this allows us to calculate the maximum resistance of the bonding cable. In short, the
resistance should be low enough to allow the arc to extinguish the moment the current falls
to that of the assumed severe level for the slow, long duration component (400 A).
Generally, the arc extinction voltage was found to vary, and the higher the threat
level, the lower the arc extinction voltage was. This is presumably due to the hotter and
larger region of plasma in a higher current arc, which makes it more stable. Although
generally the extinction voltages were in the region of 14-40 V, some cases occurred where
the arc extinction voltage was as low as 6 V. This is not understood, and may actually be a
consequence of intermittent make-break contacts.
Some of the testing at the medium threat levels was carried out with a 1 mm spacer,
which means that arcs of this length have to be maintained across the shunt/shell interface,
and reconnections via fortuitous contact or molten material cannot occur. With such spacers
in place the arc always extinguished before the available voltage dropped below 14 V. Often
the arc extinction voltage was higher but it is useful to have the minimum value to work
from.
To effectively protect against a large portion of the continuing current passing
through the shunt/shell interface, the roof bonding cable must be of a low enough resistance
so that the voltage drop along it does not exceed this critical minimum. So taking 14 V
as the minimum and a value of 400 A for the slow component, the maximum allowable
resistance of the bonding cable and its connections is 35 mΩ. However, to allow for the
possibility of higher current arcs being maintained at lower voltages down as low as 6 V,
then a resistance of 15 mΩ would be appropriate. This would not only cover long duration
currents to IEC Level I (98 %), but to a lower (~85 %) of strikes according to test data in Zundl
Koordinierte Messsungen von Blitsstroemen und Ihrer elektromagnetischen Felder an einem
Fernmeldeturm.
It is proposed that two bonding cables, each with a resistance including that of the
connections not exceeding 25 mΩ each (12 mΩ combined) are used. This would provide a
high level of protection against the slow (and more hazardous) component.
If a 1 mm spacer is incorporated into the shunt design, the 14 V extinction will
occur and the roof bonding cable protection is enhanced, and 35 mΩ resistance would be
adequate.

Phase 2:23
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

3.7 COMPARISON WITH REAL TANK DESIGNS

Some comments are made here on typical installations, and the feasibility of implementing
protection against the spark particles falling behind or close to the secondary seals.
Tanks examined in phase 1 of the investigation were typically protected by shunts
either with one roof bonding cable lying upon the roof or by two retractable bonding
cables.

3.7.1 Resistances of typical roof bonding cables

Resistance values were not available for the designs assessed in phase 1 of the investigation,
so this should not be viewed as a comment on any particular manufacturer’s products.
However the test results in this report show that resistance clearly is the crucial factor and
effort should be directed to keeping the resistance value down.
For the bonding cable it would be possible to use, for example, AWG0 gauge copper
wire (8,3 mm diameter), which has a rated resistance of 0,29 Ω per kilometre. A 50 m long roof
bonding cable would then have a resistance of about 15 mΩ. Using two or more cables will
of course reduce the resistance accordingly. It will also help current distribution if bonds are
made from the roof to two or more well separated locations on the shell rim.
Clearly the resistance of any bonding connection must be kept low. It is practical to
achieve values of well below 1 mΩ, but care must be taken to ensure that the joint is well
sealed and does not corrode.
The material from which the cable is constructed is not relevant, as long as the
appropriate resistance can be achieved and maintained without corrosion issues.

3.7.2 Inductances of typical roof bonding cables

Inductance is addressed in this report, but for values up to 100 µH it does not appear to
be a significant parameter. The inductance of a 50 m wire in free space is about 50 µH. As
demonstrated during the test series, an inductance of this order does not hamper the
protection the bonding cable offers from the slow component. The high energy component
is so fast that we should not expect any significant part of this component to flow through
the roof bonding cable.

3.7.3 Spooled roof bonding cables

Some installations use flat roof bonding cables which self spool, and which may have no
external insulation. It is difficult to comment on these without seeing one after some time in
operation. In principle the device can have the advantage of a low resistance and inductance
if current can freely flow between the adjacent spooled layers. However, contamination with
wax or oxide is an effect which would impede such current flow and should be considered,
especially as the resistance particularly needs to be relatively low. With several such devices
installed, the resistance of <10 mΩ should be achievable.
The conducting bearing/bushing must also be included in an assessment of the
conductive path. It would be a good idea to simply measure the inductance (L) and resistance
(R) of such devices in the condition in which they are normally found, and also in the retracted
and extended mode.
Such devices would have a minimal effect on conducting the fast, high energy
components, even with several around the rim. Therefore they would need to be used in
conjunction with shunts.

Phase 2:24
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

3.7.4 Ladders

Ladders may help protect the structure from the effects of lightning, and will present a fairly
low inductance, but due to the nature of the contacts (either sliding, with wheels or hinges)
the resistance of these interfaces is likely be high, or at best ill-defined with the possibility of
arcing, and so it is appropriate not to assume credit for such a path.

3.8 PROBABILITIES OF IGNITION

The results are statistical in nature, and no configuration eliminated completely the threat of
ignition. The random nature of the spark showers created in the tests makes the results of the
gas tests variable. For this reason an attempt was made to also measure and use the concept
of arc extinction voltage in support of design improvements (see 3.4).
It should be noted that:

a) Flammable gases present at shunt level


For a strike to the roof, sparking will occur at the shunts, and if flammable gases
are present at shunt level an ignition should be expected. An ignition would also
be expected under these circumstances for a strike to the tank rim when the roof is
high. Here, any shunts local to the strike would see large inductive voltages and most
likely spark.

b) Flammable gases present at the secondary seal but not at shunt level
i) If no roof bonding cable is present
For a strike to the roof, test evidence suggests that severe sparking would
occur at these levels and an ignition would be probable. For a strike to the
rim the long duration currents would be conducted through the steel, and
an ignition is less probable.
ii) If a roof bonding cable is present
For a strike to the roof test evidence suggests that the ignition probability is
significantly reduced by the use of a roof bonding cable. To protect against
the long duration currents a 15 mΩ cable should cover ~85 % of strikes.

It should be noted that positive strikes are of long duration and high charge transfer, and
this waveform threat has not been explicitly covered by these tests. Based on the test data
one would anticipate that this threat would generate hazardous sparks even in the presence
of a roof bonding cable. Such positive strikes account for ~10 % of all ground strikes
worldwide.

Phase 2:25
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

ANNEX A
NOTES ON IGNITION PROBABILITIES ARISING FROM
LIGHTNING STRIKES TO TANKS

A.1 STRIKE RATE

Roughly speaking, the lightning ground flash density is related to thunderstorm days by:

(1) Ng = 0,04 · Td1,25 per km² per year

Ng = Ground flash density


Td = Thunderstorm days

BS 6651:1999 Code of practice for protection of buildings against lightning gives 10


thunderstorm days per year in Great Britain. This equates to 0,71 ground flashes per km per
year. In Texas for example, the rate is around 1,5 ground flashes per km per year.

The effective area of a spherical tank (assuming on flat ground with no shadowing structures)
is:

(2) Aeff = π · (D + 3H)² / 4

Aeff = Effective area where D is the diameter and H is the height

The 4H term arises from the structure's height shielding the local vicinity and thus enhancing
the structure's catchment diameter by an effective twice the tank's height; assuming that no
other tall structures occur within this enhanced area.

For a 60 m diameter and 15 m high tank, the effective area is 0,011 square km. Note: this is
for a tank that is unobstructed by any local structures.

So the strike rate is:

(3) Φ = Ng · Aeff (3) = (1) x (2)

Φ = Strike rate
Following the example of a 60 m tank in Great Britain, the strike rate is approximately 0,00781
strikes per tank per year; or once every 128 years.

Phase 2:26
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

A.2 RISK OF A PARTICULAR THREAT LEVEL

The following table shows threat levels from IEC 1024-1-1 1990 Protection of structures
against lightning – Part 1: General principles

Protection level I II III-IV


Percentile % 98 95 90
Current peak value I (kA) 200 150 100
Total charge Qtotal (C) 300 225 150
Impulse charge Qimpulse(C) 100 75 50
Specific energy W/R (kJ/Ω) 10 000 5 600 2 500
Average steepness di/dt30-90 % kA/µs 200 150 100

(4) λ = 1 – Percentile (as decimal)



λ = Risk of threat level

This is the chance that a strike will exceed the particular threat level.

Note that the long duration current component is not explicitly given in this table, and in fact
it is a parameter which is rarely discussed in specifications. The IEC Level 1 protection for the
long duration currents exceeds 98 % of strikes and is 200 C in 0,5 s (i.e. an average current
of 400 A).
In Zundl Koordinierte Messsungen von Blitsstroemen und Ihrer elektromagnetischen
Felder an einem Fernmeldeturm, measurements on the Peissenberg tower in Germany
suggested long duration peak currents exceeding 1 kA in 10 % of cases, but with the average
current exceeding 500 A in 10 % of strikes.
By assuming a value of 400 A as the intended protection level these test data imply
that perhaps 85 % of strikes are covered, rather less than is implied from IEC 1024-1.

A.3 CHANCE OF DAMAGE AT PARTICULAR THREAT LEVEL

This is a rather complicated statistical problem. The best approach to this is to pick a desirable
probability of survival, and test structures to find the level at which they survive with said
probability.
For example, you want to know the level at which there is an 80 % chance of survival.
Hypothetically we conclude that 10 out of 10 tests of Level III give a pass. For a sample size
of 10, 10 passes only makes us 80 % sure that any proceeding test would pass.

(5) H = 1 – Confidence at tested threat level (as decimal)

η = 1 – 0,8 = 0,2
λ = 1 – 0,9 = 0,1

Phase 2:27
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

A.4 CHANCE OF EXCEEDANCE

This is calculated as the product of the strike rate, risk of a particular level and chance of
damage at that particular level.

(6) Є = Φ · λ · η (6) = (3) x (4) x (5)

Є = Chance of exceedance

This is the chance of extensive damage occurring within any one year.
If it has been found that the hypothetical tank has an 80 % chance of surviving a
level III threat, then the chance of exceedance is:

Є = 0,00781 x (1 – 0,95) x (1 – 0,8) = 0,0000781

Or, once in every 12 804 years.

For the same tank in Texas, using the above calculation increased rate of ground flashes by
a factor of 2,1 a destructive event could be expected every 6 402 years for our hypothetical
tank.

A.5 REALITY

It should be noted that both λ and η are complicated variables. For instance, if the roof is high
and flammable gases are present around the shunts, the effective chance of exceedance is
greatly increased. This is because any strike to the tank and especially the roof would cause
some sparking at the shunts as these are the least inductive current path. Having flammable
gases present amongst these sparking contacts will almost certainly lead to a fire and a
greatly increased value of η. Therefore, with the above example, if the roof was raised and
flammable gases present, the probability it will be ignited by a lightning strike is reduced to
once every 128 years – more than a 100 fold increase in risk.
A second complication arises from the nature of lightning. There are two types:
firstly negative lightning (which accounts for roughly 90 % of strikes, has fast rise times but
relatively low specific energy and charge transfer); secondly, positive strikes (which are of
longer duration and carry significantly more energy and impulse charge than the negative
strikes). This complicates the situation for roof bonding cables, since a design to divert the
400 to 600 A trailing currents from a negative strike will be too resistive to protect against the
50 kA or so positive strike. For a positive strike it would therefore be very difficult to prevent
hazardous sparking in shunts.

A.6 SUGGESTED CHANCE OF EXCEEDANCE FROM TESTING SO FAR

Testing thus far has shown that arcing on the shunts from the slow component is more
hazardous than that from the fast component.
So for example with 0/10 ignitions caused with a 16 kA fast component, 4 kA
intermediate component and 600 A peak (300 A average) slow component we can be 80 %
confident that this threat level will not cause an ignition. The long duration current is the
critical component, and we know that a 400 A average current covers about 85 % of strikes.

Phase 2:28
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

This current pulse was through a 6 mΩ, 40 µH bonding cable. This reflects a Level II
threat assuming that impulse current can divide between 10 of the shunts, but that the long
duration currents only divide between a single shunt and the bonding cable. If this is the only
source of an ignition, then the chance of exceedance for a single tank in Great Britain is once
in every 6 410 years, and in Texas once in every 3 052 years.

Є(Texas) = 0,01562 x (1 – 0,95) x (1 – 0,8) = 0,0001562

Or, once in every 6 402 years.

Phase 2:29
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

CUL/LT-0401

INVESTIGATIVE TESTS ON THE LIGHTNING PROTECTION OF


SUBMERGED SHUNTS WITH PARALLEL ROOF BONDING CABLES

CONTENTS

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

2 Test arrangement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.1 The bench tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.2 The high level tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.3 Threat levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

3 Test results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40


3.1 Air (2 kA peak arc, no bonding cable) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.2 Air (2 kA peak arc, 30 mΩ bonding cable) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.3 Air (2 kA peak arc 10 mΩ bonding cable) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3.4 Water (20 mm, 2 kA peak arc, no bonding cable) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3.5 Water (20 mm, 2 kA peak arc, 30 mΩ bonding cable) . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.6 Water (20 mm, 2 kA peak arc, 10 mΩ bonding cable) . . . . . . . . . . . . . . . . . . . . . . . . . 50
3.7 Oil (20 mm, 2 kA peak arc, no bonding cable) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.8 Oil shunts at greater depths (40 mm-450 mm, 2 kA peak, no bonding cable) . . . . . . . 53
3.9 Oil (20 mm, 2 kA peak arc, 30 mΩ bonding cable) . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.10 Oil (20 mm, 2 kA peak arc, 10 mΩ bonding cable) . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.11 Oil shunts at various depths (40-450 mm, 2 kA peak, 30 mΩ bonding cable) . . . . . . . . 57
3.12 Oil shunts at various depths (40-450 mm, high energy current component ,10 kA- 44 kA) . . . . . 59
3.13 Tests to partially submerged shunts in oil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
4.1 Use of bonding cables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
4.2 Submerged shunts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
4.3 Partially submerged contacts (primary seal straddling fluid surface) . . . . . . . . . . . . . . . 62
4.4 Comparison with real tank designs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
5.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
5.2 External shunts (not submerged) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
5.3 Submerged shunts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

Annexes:

Annex A Results from test series to deeper shunts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
Annex B Gases emitted by submerged arcs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
Annex C Facility and calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

Phase 2:31
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

Figures:
Figure 1 Bench test set up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Figure 2 Measurements made during arcing under water. . . . . . . . . . . . . . . . . . . . . . . . . . 37
Figure 3 Measurements made during arcing in air . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Figure 4 High level test set up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
Figure 5 Typical components of a lightning strike . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
Figure 6 Measurements made during test (air, 2 kA peak arc, no bonding cable) . . . . . . . . 42
Figures 7a-d Measurements made during test (air, 2 kA peak arc, 30 mΩ bonding cable) . . . . . 44
Figures 8a-c Measurements made during test (air, 2 kA peak arc, 10 mΩ bonding cable) . . . . . 45
Figure 9 Measurements made during test 15 (water, 20 mm 2 kA peak arc,
no bonding cable) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
Figure 10 Measurements made during test 18 (water, 20 mm 2 kA peak arc,
no bonding cable) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
Figures 11a-c Measurements made during test 19 (water, 20 mm 2 kA peak arc,
30 mΩ bonding cable) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
Figures 12a-c Measurements made during test 20 (water, 20 mm 2 kA peak arc,
10 mΩ bonding cable) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
Figure 13 Measurements made during test 25 (oil, 20 mm 2 kA peak arc, .
no bonding cable) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
Figures 14a-c Measurements made during test 29 (oil, 20 mm 2 kA peak arc, .
30 mΩ bonding cable) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
Figures 15a-c Measurements made during test 31 (oil, 20 mm 2 kA peak arc, .
10 mΩ bonding cable) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
Figure 16 Floating roof with pusher plate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
Figure 17 1st test set up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
Figure 18 2nd test set up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
Figures 19a-c 2nd test current configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
Figures 20a-c Secondary seal configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

Photographs:
Photograph 1 Layout of bench test set up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Photograph 2 In air . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Photograph 3 Under several mm of water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Photograph 4 Under ~1 mm of water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Photograph 5 Shunt/shell interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
Photograph 6 Test in air. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
Photographs 7a-7e Test 18 sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Photographs 8a-8d Test 19 sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
Photograph 9 Test 20 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
Photographs 10a-c Test 25 sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
Photograph 11 40 mm deep shunt 2 kA peak, no bonding cable. . . . . . . . . . . . . . . . . . . 53
Photographs 12a-d Test 29 sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
Photograph 13 Test 24 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
Photographs 14a-c Test 23 sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
Photograph 15 40 mm deep shunt 2 kA peak, no bonding cable . . . . . . . . . . . . . . . . . . . 58
Photograph 16 100 mm deep shunt 2 kA peak, 30 mΩ bonding cable . . . . . . . . . . . . . . . 58
Photograph 17 Metal pusher plate sparking at oil surface with roof bonding cable in place . 61

Tables:
Table 1 Results from first series of tests all at 2 kA, 200 C to 250 C. . . . . . . . . . . . . . . . . . . . . 41
Table 2 Results from second series of tests with deeper submersion of the shunts . . . . . . . . . . 41

Phase 2:32
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

SUMMARY

This report describes the testing which has been performed under Phase 2 of API/EI Research
report: Verification of lightning protection requirements for above ground hydrocarbon
storage tanks. This comprised:
1) Looking at sparking effect at shunts submerged under water and representative
hydrocarbons, on the bench with current levels of 40-400 A. Determine if there is any
appreciable pressure effect from the long duration arcs which might lead to eruption
of fluid. Determine if sparking or plasma can escape from fluid, using flammable gas
and optical methods.
2) Investigating how such sparking is affected by the presence of additional bonding
cables.
3) Performing simple tests on the main bank using high energy waveforms and looking
at fluid eruption effects with water, and arc quenching effects using a parallel bonding
cable.

A development for this phase of testing was the measurement of the voltage across the
shunt/shell interface during such tests. This measurement gives a good indication for
occurrence of arcing, and provided information on arc voltage drop which supersedes
some of the findings reported in CUL/LT-0373 Lightning simulation testing to determine
the required characteristics for roof bonding cables on external floating roof above ground
storage tanks.
The findings are summarised as follows:
Bench tests were carried out with arc currents of up to 100 A, in water and air, mainly
to understand arc behaviour in these media.
Arcs lasting for up to 0,2 s duration could be formed, as long as sufficient voltage is
available to maintain the arc, and if the geometry of the contact does not result in a weld.
−− The voltage drop across an arc in air was relatively constant, and never less than 14 V;
the arc would extinguish when the available voltage dropped below this.
−− The voltage drop across arcs under water or oil was similar. It was generally a little
higher and more variable than in air. Arcs under these fluids would extinguish when
the available voltage dropped below about 20 V.

In the main bank tests, arcs were formed both in air and from 20 mm to 450 mm below the
surface of the fluid. The voltage measurements confirmed the findings in the bench tests,
that a voltage of at least 14 V was required to prevent the arc extinguishing.
Most tests were performed with the long duration currents but eventually tests were
carried out with both the fast and long duration components.
−− With shallow (20 mm) oil, an eruption of fluid and bubbling occurred, and sometimes
sparks and a jet of sooty flame. Flames also occurred testing shunts under 20 mm
water.
−− With submersion of 300 mm or greater under oil no ignition hazards were observed
(even without bonding cables). This was even the case with the fast component
applied to the shunt.

The magnitude of sparking and other effects was clearly reduced by having a parallel
bonding cable. The effectiveness of the cable is determined by its resistance, since any arcs
will extinguish if the voltage drop along the bonding cable is <14 V.
However it appears that in at least some tank designs there are pusher plates which
hold the primary seal against the shell, which may well provide additional metallic current
paths. If arcing occurs at these plates or primary seals it will be predominantly in the vapour
space, or at the liquid level line, and so potentially would be hazardous.

Phase 2:33
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

1 INTRODUCTION

This report describes investigative tests to understand how effective submerged shunts might
be in protecting storage tanks from lightning induced tank fires. Shunts are metallic sprung
plates generally used to provide electrical bonding for both lightning and static electricity
between the floating roof and the shell (wall) of hydrocarbon storage tanks.
The potential problem is the arcing and sparking which could occur when lightning
currents flow from the tank floating roof to the shell. Protection against this threat is
normally addressed by the use of shunts above and around the edge of the roof, and from
roof bonding cables. These combined devices provide some protection of the tank from the
separate components of a lightning strike.
It has been suggested that the use of submerged shunts would greatly increase
safety, since the arc would not be in a flammable atmosphere, and there would be no ignition
hazard unless sparks could erupt from the surface of the liquid.
Part of the aim of these tests is to see whether such eruptions could occur.
The tests also look at the additional effectiveness provided by roof bonding cables,
a study which was reported in CUL/LT-0373 Lightning simulation testing to determine the
required characteristics for roof bonding cables on external floating roof above ground
storage tanks. These cables can reduce the threat of arcing at the shunt by conducting a large
proportion of the long duration currents directly from the roof to the shell of the tank.
The tests reported here were carried out in two phases:
1) Bench tests to study voltage drops across arcs in air and water, and any tendency for
fluid to erupt above sites of any arcing. Tests were performed with arc currents up to
100 A.
2) High level tests using the main capacitor bank. This is in part to determine whether
the arc behaviour from the bench tests would be the same at high level. It also
provides a more representative installation to study current distribution and effects
of sparking under both water and oil.

Phase 2:34
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

2 TEST ARRANGEMENT

2.1 THE BENCH TESTS

The general arrangement for the tests is shown in Figure 1. A DC power supply or battery
provides a source of 0-40 V which can drive up to 250 A through a lightly held steel-steel
interface. This voltage range is what would be expected at a shunt when a roof bonding
cable is installed.
One small steel plate is mounted onto a threaded brass rod, and a second steel disc
is held lightly against the plate using a plastic spring. A heavy but flexible braid provides
the current path into the disc and the threaded rod the current path out of the small steel
plate.

Figure 1 Bench test set up.

40 mΩ DC power supply or batteries


0-40 V up to 250 A

Steel The voltage across the 40 mΩ resistor


Steel is monitored to give a current
plate disk measurement

For these particular tests measurements were made as follows:


1) Applied current (as voltage across 40 mΩ resistor).
2) Voltage drop across the interface.
3) Light emitted from the interface region.

Photograph 1 shows the layout as set up for use with the batteries. Switching of the current
is made by physically bringing together two contacts, and an in-circuit fuse mounted on the
batteries limits the duration of the high current.
The cables run off at left to a digital storage oscilloscope.
Some typical examples of the measurements and results are given next.

Phase 2:35
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

Photograph 1 Layout of bench test set up.

40 mŸ resistor

Arc contacts

Photographs 2, 3 and 4 are representative photographs of arcs of ~100 A. Very little disruption
of the water occurs for submerged arcs at this current level, but sparks could escape through
very shallow water (~1 mm). The disc is 25 mm (1”) in diameter.

Photograph 2 Photograph 3 Photograph 4


In air. Under several mm water. Under ~1 mm water.

Phase 2:36
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

The upper case in Figure 2 is the current, and the lower case the voltage across the contacts.
The contacts here intermittently closed and opened apparently at random, with an arc usually
present when the contacts are open. Note that the arc voltage is close to 15 V. A similar
measurement on arcing in air is given in Figure 3.

Figure 2 Measurements made during arcing under water.

Note: The dashed lines below the plot show the changes in contact configuration during the
test, which can be deduced from the measurements.

C ulham Lightning
LIG HTNING panel diagnos tics -C ulham
16-F eb-7 16

14

12

P lot 1
10

1F E B ~1\WAV E 0047.WF T 8
sca ling 1 = x 1.000
Current
V

25 A/V 4

-2
0 50 100 150
Millis econds

30

25

P lot 2
20
1F E B ~1\WAV E 0048.WF T
sca ling 2 = x1000.000

Voltage V 15
V

10

-5
0 50 100 150
closed contact Millis econds

open arcing contact


open contact

Figure 3 Measurements made during arcing in air.

Current
25 A/V

Voltage V

Phase 2:37
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

The bench tests were carried out over several days. Their primary purpose was to understand
the mechanisms behind the arcing, both in air and under fluid, and to develop measurement
techniques. The measurement techniques were then used in the high level tests which had a
more representative set-up.

2.2 THE HIGH LEVEL TESTS

The general arrangement for the tests is shown schematically in Figure 4. A segment of a
tank shell is mounted in the arc rig, and connected to the ground output terminal of the
lightning bank. The shunt was connected by a bonding cable to the bank live output line.
The shunt position on the shell could be adjusted.
To simulate a roof bonding cable a cable of 100 µH inductance was used, its resistance
was ~10 mΩ, but was increased to 30 mΩ for some tests by adding an additional resistor.
This simulates the typical range of parameters which could be achieved for a cable over 50 m
long.
For these particular tests measurements were made as follows:
1) Total injected current.
2) Current flow through the shunt (mainly indicative to look for make/breaks).
3) Voltage drop across the shunt/shell interface.
4) Light emitted from the shunt/shell interface region.

In addition many of the tests were recorded with a video camera and DVD recorder.
The shunt/shell interface as used in the first series of tests is shown in Photograph 5.
The wire leading from the shunt up against the shell is a diagnostic wire for measuring shunt/
shell voltage. Also visible is the optical fibre pointing downwards towards the shunt, and a
pipe for directing butane/air mixture against the shell.

Figure 4 High level test set up. Photograph 5 Shunt/shell interface.

Roof bonding cable of 100 µH


inductance, resistance 10 mΩ or 30 mΩ

Shell
Current in
Shell
Shunt

Shunt

In the first series of tests, shallow depths (~20 mm) of water or oil were investigated. Further
consideration led to a series of tests where the simulated shunt was moved deeper to
investigate the depth at which sparking would no longer pose an ignition hazard. Here all
tests were carried out with oil rather than with water.
The results from these tests provide the main content of section 3.

Phase 2:38
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

2.3 THREAT LEVELS

The currents and time scales attributable to a severe strike which combines all the severe
parameters is shown schematically in Figure 5.

Figure 5 Typical components of a lightning strike.

It is known from earlier tests that the fast currents inevitably flow through the shunts and will
divide between many such shunts. Hence, 11 kA was originally proposed in Phase 1 as being
a reasonable test level for a single shunt.
The long duration currents can be largely diverted through a roof bonding cable,
although some can flow through the shunts (depending on the conditions) and it is the
purpose of these tests to determine this dependency.
The testing assumes that all of the intermediate and long duration current can flow
through a roof bonding cable and a single shunt. The bench tests were at relatively low level,
but the main bank tests reach a peak current of 2 kA (intermediate) with a long duration
current of 1 500 A decaying to zero in 0,3 s, with a typical total charge transfer of 250 C.
Tests with the fast component only or with a combined (fast and long duration) pulse
are discussed in Section 3.12 of this report and in EI-EN2-04 Lightning tests to tank shell/shunt
samples. At high levels of 44 kA considerable eruption of fluid occurs, although no sparks or
burning gas occurs.

Phase 2:39
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

3 TEST RESULTS AND DISCUSSION

Bench tests were carried out with applied currents of up to 250 A, with currents falling to
100 A when contacts break and an arc is formed. These tests were performed in water and
air, mainly to understand arc behaviour in these media. Arcs lasting for up to 0,2 s duration
could be formed, as long as sufficient voltage is available to maintain the arc, and if the
geometry of the contact does not result in a weld. The main findings from these bench tests
were:
1) A voltage drop across the arc which was largely independent of current and exceeded
14 V, being generally in the range 14-25 V (Figures 2 and 3).
2) A clear visible difference in arcs formed in air and under water, in that the showers
of bright sparks emitted in air were not evident under water. However, with very
shallow liquids burning metallic particles could still escape (Photographs 2, 3, 4).

The supply voltage in these bench tests did not exceed 40 V, so they were only simulating
cases where the shunt currents and arcs were driven by the voltage drop across bonding
cables (without bonding cables present a limitless voltage would be available to drive current
across shunts, as the lightning is essentially a constant current source).
High level tests were carried out to determine whether these findings were still valid
at higher currents, and also to test more representative combinations of bonding cables and
shunts. In the high level tests the following bonding cable configurations were investigated:
1) No bonding cable (all of current forced through the shunt-shell).
2) 10 mΩ bonding cable of 100 µH inductance.
3) 30 mΩ bonding cable of 100 µH inductance.

These tests were carried out for arcs in air, under ~20 mm of water, and under ~20 mm to
450 mm of oil. The relatively shallow immersion of 20 mm was used as an assumed worst
case, although favourable results were still expected. Results were in fact less favourable than
expected in terms of ejected sparks and flame. Only depths greater than 300 mm were found
to produce no ignition source at the surface of the oil.
Results of these high level tests are given in detail in this section. In summary it was
found that the arc electrical behaviour remained essentially the same at these high current
levels as it was during the bench tests i.e:
−− The voltage drop across an arc in air was relatively constant, and never less than 14 V.
The arc would extinguish when the available voltage dropped below this.
−− The voltage drop across arcs under water or oil was similar, but a little higher and
more variable than in air. Arcs under these fluids would extinguish when the available
voltage dropped below ~20 V.

However, the arcing was more violent during higher level tests and resulted in some eruption
of fluid. The magnitude of sparking and eruption of fluid is visibly reduced by having a
parallel roof bonding cable. The effectiveness of the cable is determined by its resistance,
since arcs will extinguish if the voltage drop along the cable is <14 V. The main test results
are shown in Table 1.

Phase 2:40
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

Table 1 Results from first series of tests all at 2 kA, 200 C to 250 C. Unless stated otherwise
all shunts were submerged by 20 mm only.

Medium Bonding configuration Test No. Comments Butane Results


ignition in
Y/N section
Air No bonding cable 16 Spark shower Y 3.1
Air 30 mΩ bonding cable 12 Spark shower Y 3.2
Air 10 mΩ bonding cable 14 Spark shower Y 3.3
Water No bonding cable 18 Sparks and burning gas from N/A 3.4
surface
22 Sparks and burning gas from N
surface
Water 30 mΩ bonding cable 19, 21 Several sparks on surface N 3.5
Water 10 mΩ bonding cable 20 Few sparks N 3.6
Oil No bonding cable 25 Plume of burning vapour and N 3.7
particles
Oil No bonding cable Varying submersion tests, see Table 2 3.8
Oil 30 mΩ bonding cable 29 Plume of burning gas and particles Y 3.9
24, 28 Smoke and surface burning N
particles
Oil 10 mΩ bonding cable 23, 26, 27, Burps of sooty smoke and N 3.10
30, 31 occasionally one or two burning
surface particles
Oil 30 mΩ bonding cable Varying submersion tests for fast 3.11
components, see Table 2
Oil Partially submerged shunt 32, 33, 34 Arcs in air or air/oil interface N/A 3.13

Table 2 Results from second series of tests with deeper submersion of the shunts (more
detailed results see Annex A).

300 mΩ 100 µH bonding cable No bonding cable


Depth under oil No. of tests Sparking? No. of tests Sparking?
Long duration currents (~1,8 kA, 110 C) (Section 3.8 and 3.11)
40 mm 1 1 sparking 2 1 sparking
100 mm 3 No sparking 1 1 sparking
300 mm 2 No sparking 6 No sparking
450 mm 2 No sparking 2 No sparking
Fast component currents (from 11 kA to 44 kA) (Section 3.12)
100 mm 6 No sparking
Primary seal straddling fluid line (Section 3.13)
Crossing oil level 5 5 sparking 3 3 sparking

Phase 2:41
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

The following points about the results were observed:


−− With water at shallow depths (~20 mm), an eruption of fluid occurred, and sparks
and flame were sometimes emitted too.
−− With oil at 100 mm or less, an eruption of fluid occurred, and sometimes sparks and
a jet of flame. The long duration currents were more significant for ignition hazards,
although the fast components disturbed the liquid much more.
−− The tests with shunts submerged to 300 mm or greater had no eruption of hazardous
sparks, even if no bonding cable is used. The fast components produce some eruption
of fluid but no sparks, even at high level and shallow submersion (44 kA–100 mm).

However, the primary seal will spark above the fluid level if it can carry current.

3.1 AIR (2 kA PEAK ARC, NO BONDING CABLE)

In this test (Test 16) the arc continues for 100 ms, at which time the shunt welds to the shell
and the measured arc voltage drops to zero.

Figure 6 Measurements made during test (air, 2 kA peak arc, no bonding cable).
A
Current

Milliseconds
V
Shunt/shell voltage

15 V

Milliseconds

Phase 2:42
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

Photograph 6 Test in air. All video sequences of arcs in air showed very intense spark
showers. They were difficult to compare for the different configurations so only this
example is given.

Phase 2:43
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

3.2 AIR (2 kA PEAK ARC, 30 mΩ BONDING CABLE)

In this test (Test 12) the arc continues for about 80 ms, at which time the arc goes out. The
voltage across the shunt/shell at this time is measured as ~20 V (see Figure 7b). Note that
this voltage is equal to the bonding cable resistance times the current flowing in it, which
is ~30 mΩ x 600 A = ~18 V.
Note in Figures 7b, 7c, 7d regions where the shunt contacts the shell or briefly welds
to it (zero voltage and higher current). Also where the arc extinguishes briefly or permanently
(zero current). In both cases there is no arc and the light signal reduces sharply.

Figures 7a-d Measurements made during test (air, 2 kA peak arc, 30 mΩ bonding
cable).
A

600 A
Current

Figure 7a
V
Shunt/shell voltage

Figure 7b

15V

Note in Figures 7b, 7c, 7d


~950A
regions where the shunt
contacts the shell or briefly
A

welds to it (zero voltage and


Shunt/shell current

higher current).Figure 7c

Also where the arc


extinguishes briefly or
permanently (zero current).

In both cases there is no arc


and the light signal reduces
sharply.
Light signal (-ve going)

Figure 7d

Phase 2:44
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

3.3 AIR (2 kA PEAK ARC, 10 mΩ BONDING CABLE)

In this test (Test 14) the arc continues for only 30 ms, at which time the arc goes out as
the shunt contacts or welds to the shell. The voltage across the shunt/shell at this time is
measured as ~15 V (see Figure 8b). Note that this voltage is equal to the bonding cable
resistance times the current flowing in it, which is ~10 mΩ x 1 400 A = ~14 V.
The lower resistance bonding cable helps the arc to extinguish earlier.

Figures 8a-c Measurements made during test (air, 2 kA peak arc, 10 mΩ bonding
cable).
A
Current

Figure 8a
V
Shunt/shell voltage

Figure 8b 15 V

~1 100 A
A
Shunt/shell current

Figure 8c

3.4 WATER (20 MM, 2 kA PEAK ARC, NO BONDING CABLE)

In Photograph 7 a plume of burning gas is clearly visible erupting from the water, although
the water itself is not much disturbed. There is no splashing onto the shell, which remains
dry.
No butane was applied for this test, the plume of burning gas arose from the arc and
its interaction with the water.
Metallic particles continue burning, skating along the water surface, but none were
seen thrown out of the water (Test 18).
This sequence is <0,5 s, as are subsequent sequences for other configurations. For

Phase 2:45
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

the tests in this section the shunt/shell interface is submerged approximately 20 mm below
the water surface. In Test 15 (Figure 9) a fine fuse was used to initiate an arc 1 mm to 2 mm
long, and the voltage drop across the arc is relatively high, eventually going off scale at >50
V. In Test 18 (Figure 10) the arc is very short, as the shunt was in glancing contact with the
shell. The arc voltage is fairly constant at 20–25 V for over 200 ms.

Figure 9 Measurements made during test 15 (water, 20 mm 2 kA peak arc, no bonding


cable).
V A Current
Current A voltage
voltage
Shunt/shell
Shunt/shell V

15 V

15 V

Figure 10 Measurements made during test 18 (water, 20 mm 2 kA peak arc, no


bonding cable).
A VCurrent
Current A voltage
voltage V

15 V
Shunt/shell

15 V
Shunt/shell

Phase 2:46
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

Photographs 7a-7e Test 18 sequence.

Photograph 7a

Photograph 7b

Photograph 7c

Photograph 7d

Photograph 7e

Phase 2:47
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

3.5 WATER (20 MM, 2 kA PEAK ARC, 30 mΩ BONDING CABLE)

In this test (Test 19) the arc continues until the voltage across the shunt/shell is ~22 V (Figure
11b) at which point it extinguishes leaving a gap, and all the current subsequently flows
through the bonding cable.

Figures 11a-c Measurements made during test 19 (water, 20 mm 2 kA peak arc, 30


mΩ bonding cable).
A

Figure 11a
Current
V
Shunt/shell voltage

Figure 11b 15V


15 V

~1 300 A
A
Shunt/shell current

Figure 11c

In this sequence (test 19) local arc light at the shunt is visible, as are some burning metallic
particles which move on the surface, but there is no plume of burning gas.

Phase 2:48
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

Photographs 8a-8d Test 19 sequence.

Photograph 8a

Photograph 8b

Photograph 8c

Photograph 8d

Phase 2:49
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

3.6 WATER (20 MM, 2 kA PEAK ARC,10 mΩ BONDING CABLE)

In this test (Test 20) the arc continues only for 10 ms and then extinguishes leaving a gap,
so that all the current subsequently flows through the bonding cable. Figure 12c shows the
shunt/shell current clearly falling to zero on extinction of the arc.

Figures 12a-c Measurements made during test 20 (water, 20 mm 2 kA peak arc, 10 mΩ


bonding cable).
A
Current

Figure 12a
V
Shunt/shell voltage

15 V

Figure 12b
A
Shunt/shell current

Figure 12c

Apart from the initial brief flash of the arc, only the slight glow shown here is visible. One
particle subsequently appears to reach the surface along with a couple of gas bubbles,
understood to be mainly hydrogen (Test 20).

Photograph 9 Test 20.

Phase 2:50
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

3.7 OIL (20 MM, 2 kA PEAK ARC, NO BONDING CABLE)

In these tests with oil the glancing shunt/shell interface is submerged approximately 20mm
below the surface.
Initially with a depth of 20 mm, particles and burning gas are quickly ejected from
the surface (Photograph 10a).
A large plume of burning gas erupts from the surface, although the oil itself is not
much disturbed. There is no splashing onto the shell, which remains dry (Photograph 10b).
Metallic particles continue burning, and skating along the surface of the oil (Test 25,
Photograph 10c).
In this test (Test 25) the arc voltage is fairly constant at 20–30 V for 230 ms before
the interface welds together.

Figure 13 Measurements made during test 25 (oil, 20 mm 2 kA peak arc, no bonding


cable).
VA
Current
Shunt/shell Voltage

15V

Phase 2:51
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

Photographs 10a-c Test 25 sequence.

Photograph 10a

Photograph 10b

Photograph 10c

Phase 2:52
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

3.8 OIL SHUNTS AT GREATER DEPTHS (40 MM–450 MM, 2 kA PEAK, NO BONDING
CABLE)

In the second series of tests, with the long duration current components, flames only occurred
when depths of 100 mm or shallower were tested. Here, the oil was shallow enough for the
arc pressure to blow sparks and burning vapour out of the surface (Photograph 11).

Photograph 11 40 mm deep shunt 2 kA peak, no bonding cable.

At depths greater than 40 mm, the tendency for this to happen was reduced, and only one
other such event was observed, for a submersion depth of 100 mm. However, in almost all
tests bubbles came to the surface carrying a sooty vapour. The deeper the shunt contact, the
less vigorous the bubbles, presumably afforded by the oil cooling the gas as it rises.

3.9 OIL (20 MM, 2 kA PEAK ARC, 30 mΩ BONDING CABLE)

Here, two of the result types are shown, as they were quite different.
In Test 29 (Photographs 12a – 12d), with a depth of 20 mm, there was an eruption
of flame, as well as burning metallic particles on the oil surface.
In Test 24 (Photograph 13), and Test 28 (which were of the same configuration as
Test 29), burning metallic particles also ran on the surface, and bubbles of gas and 'smoke'
were ejected, but did not burn.
In Test 29 butane/air mixture was being played over the surface, and this may have
helped the gases to ignite. In tests under oil without the bonding cable these gases also
ignited, without the presence of the butane/air mixture.
In Figures 14a – 14c (Test 29) the arc voltage is fairly constant at 20–30 V, although
the contact intermittently opened and closed. The arc finally goes out at 100 ms, leaving the
contact open. At arc extinction the voltage across the interface is around 25 V.

Phase 2:53
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

Figures 14a-c Measurements made during test 29 (oil, 20 mm 2 kA peak arc, 30 mΩ


bonding cable).

Figure 14a
A
Current
V

FIgure 14b
Shunt/shell voltage

1515V
V
A
Shunt/shell current

Figure 14c

Phase 2:54
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

Photographs 12a-d Test 29 sequence.

Photograph 12a

Photograph 12b

Photograph 12c

Photograph 12d

Photograph 13 Test 24.

Photograph 13

Phase 2:55
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

3.10 OIL (20 MM, 2 kA PEAK ARC, 10 mΩ BONDING CABLE)

In the tests with 10 mΩ bonding cables, light from the initial arc (Photograph 14a), and
from a few burning metallic particles on the surface (Photograph 14b) were seen. Small
bubbles of gas and 'smoke' were ejected but did not burn (Photograph 14c). This sequence
of photographs is from Test 23.
In Test 31 (Figures 15a-c) the arc voltage is fairly constant at 20–30 V, although
the contact intermittently closes and opens, and the arc finally goes out at 100 ms, leaving
the contact open. At arc extinction the voltage across the interface is around 25 V, but the
voltage drop across the bonding cable is much lower at 15 V. This is noticeable as a step drop
in voltage from 25 V to 15 V as the arc goes out.
This difference is due to the stored energy in the 100 µH inductance of the roof
bonding cable increasing the voltage as the current falls (V=L di/dt). Essentially this energy
helps feed the arc and sustain it, and is a more noticeable effect when the arcs extinguish at
high currents, here 1 200 A. Whilst interesting as an effect it does not significantly impact
upon the conclusions of this investigation.

Figures 15a-c Measurements made during test 31 (oil, 20 mm 2 kA peak arc,10 mΩ


bonding cable).
A

Figure 15a
Current
Shunt/shell voltage V

Figure 15b

15 V
A
Shunt/shell current

Figure 15c

Phase 2:56
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

Photographs 14a-c Test 23 sequence.

Photograph 14a

Photograph 14b

Photograph 14c

3.11 OIL SHUNTS AT VARIOUS DEPTHS (40-450 MM, 2 kA PEAK, 30 mΩ BONDING CABLE)

This section covers tests with the shunt contacting the tank shell at various depths, and using
only long duration currents. Section 3.12 covers fast or combined pulses for the same ranges
of depths. Tests are summarised in Table 2, and listed in detail in Annex A.

Phase 2:57
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

Depths tested were 40 mm, 100 mm, 300 mm and 450 mm. If a bonding cable was
used in parallel sparks and flames were emitted only at 40 mm depth. With no bonding cable
the effect was more dramatic, as shown in Photograph 15 (see also 3.8).
At a depth of 100 mm only one such event was observed, in a test when no bonding
cable was used (see Section 3.8). Although at deeper submersions the emission of flames
could be eliminated, in almost all tests bubbles came to the surface carrying a sooty vapour.
The deeper the shunt contact, the less vigorous such bubbles, presumably afforded by the oil
cooling the vapour as it rises.
There was a notable reduction in current that traversed from the shunt to shell when
a simulated roof bonding cable (100 µH, 40 mΩ) was put in place. This was accompanied
by a reduction in the amount of smoke produced by the long duration components but did
not prevent the 40 mm deep shunt from emitting a flame. In tests at 100 mm, bubbles of
'smoke' occurred, but no ignitions when the bonding cable was used. As Photograph 16
shows, smoke is barely visible, and there is no sign of sparks or flame.

Photograph 15 40 mm deep shunt 2 kA peak, no bonding cable.

Photograph 16 100 mm deep shunt 2 kA peak, 30 mΩ bonding cable.

Phase 2:58
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

3.12 OIL SHUNTS AT VARIOUS DEPTHS (40-450 MM, HIGH ENERGY CURRENT
COMPONENT, 10 kA- 44 kA)

The fast, high energy components caused little or no smoke to be emitted from the oil
surface. At 10 kA, which is close to the agreed threat level, the surface of the oil is caused to
ripple gently. Levels were increased to a severe 44 kA, at which oil was caused to erupt from
the tank vigorously, but with no flames or ignition (again this test was carried out at a depth
of 100 mm). It appears that the fast, high energy components do not present a threat if the
shunts can be well submerged.

3.13 TESTS TO PARTIALLY SUBMERGED SHUNTS IN OIL

From discussions it became apparent that in at least some types of installation the floating
roof has pusher plates which hold the roof at a determined distance from the tank shell.
These metallic plates are essentially under a high sprung load and would be expected to
present an electrical path into the primary seal and thence into the shell.
This is shown schematically in Figure 16.

Figure 16 Floating roof with pusher plate.

Secondary seal

Floating roof
Shell

Pusher plate

Primary seal

In this case the primary seal passes through the oil/air surface. It is to be expected that if
sparking were to occur it would be in the air, as air is considerably easier to break down
electrically.
In the first series of testing, some simple tests were carried out to study this, as
shown in Figure 17 and Figure 18. A flat stainless steel plate was held against the shell, and
currents of the same waveshape as above (i.e. 2 kA peak long duration current) were injected
through the interface. Most of the interface was outside of the fluid as shown in Figure 17,
where the tests both gave breakdown in the air. In the second test configuration (Figure
18) the initial breakdown was forced to take place 20 mm under the surface. However, the
arcing damage mainly occurred at the level of the oil/air interface, and it is clear that the arc
moved up to this preferred level. However, the fast impulse, had it been applied, would have

Phase 2:59
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

occurred at the point of metal to metal contact, and if this were under the fluid it would not
have been expected to move position, because of its short duration.
In Figures 17 and 18, the current flow is indicated by arrows, electrical connections
were made by bolting or clamping cables to these items. A bonding cable did not make a
noticeable difference to the vigour of the sparks issued from the arc. This is presumably as
they are emitted before current has started flowing through the inductive bonding cable as
the arc burns away a layer of rust and tar.

Figure 17 1st test set up. Figure 18 2nd test set up.

Angled plate initiates arc


under oil, but arc moves
onto the surface and most
current flows here
Plate held flat
against shell.
Arcs occur in air
Figure 18
Oil

The concern here is that the configuration of pusher plate and metallic primary seal is difficult
to protect. Even if shunts are installed, a significant fraction of the current is likely to flow
through the primary seal. This is because the shunts provide a similar electrical contact to
the shell as (it is believed) does the primary seal. This is a glancing contact against a surface
which is uneven, and either rusted or with a tarry layer. So breakdown will occur at several
locations where the gap is smallest and the tarry layer or rust is easiest to break through. If
such small gaps occur at the primary seal/shell interface, then that is where the breakdowns
will also occur.
In the second series of tests, a metal plate simulating a pusher plate or primary seal
was held in contact with the tank wall sample so that some of it was above the oil surface
and the rest beneath. Three configurations were tested, each with the same result, in that
sparks were emitted in quantity. The configurations tested were with the current conducted
to the top of the plate (Figure 19a), then the base (Figure 19b) and finally via an insulated
bolt at the top of the plate to the base (Figure 19c).

Figures 19a-c 2nd test current configurations.

Figure 19a Figure 19b Figure 19c

A B

Phase 2:60
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

Most of the tests resulted in a weld at or above the oil surface, and either case resulted in a
source of ignition.
An example of typical sparking at or above the fluid surface in such tests is shown in
Photograph 17.
For some hydrocarbons the flammability of the vapours in this region of the primary
seal would be greater than in the region of the secondary seal/shunts, so the consequences
may be more severe.

Photograph 17 Metal pusher plate sparking at oil surface with roof bonding cable
in place.

Phase 2:61
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

4 DISCUSSION

4.1 USE OF BONDING CABLES

The effect of the protection provided by the shunts has become clearer in this investigation,
mainly through the ability to measure the voltage drop across the shunt/shell interface during
the tests.
A voltage of 14 V is the significant value for arc extinction. So if the roof bonding
cable has a resistance of 15 mΩ then it can carry 1 kA of long duration current without
presenting problems of arcing at the shunts. An average value of long duration current is
considered to be 400 A.
The fast, high energy component and the intermediate components of lightning may
still create arcing at the shunts, and so this approach does not eliminate the threat. However
it does reduce the level and duration of the most hazardous arcing.

4.2 SUBMERGED SHUNTS

If shunts are fully submerged below 300 mm, it is unlikely that a lightning strike will create
an ignition source at this interface. Currents tested for this depth were severe (up to 44 kA
fast high energy component, 2 kA/200 C long duration component). At shallower depths,
the arc itself is no longer exposed as an ignition source, but burning particles and eruptions
of flame can still be ejected as a result of long duration currents.
For arcs under the water surface the eruption of gas is believed to be mainly hydrogen
(see Annex B), but for arcs under the oil surface it is unclear, except that it seems to include
smoke and condensed oil vapour, and is also flammable.

4.3 PARTIALLY SUBMERGED CONTACTS (PRIMARY SEAL STRADDLING FLUID SURFACE)

Some designs of primary shoe seals and their scissors and pusher plates will compete with
the shunts and carry current across the shoe/shell at points of closest contact. Design
improvements could be to:
−− Insert an isolating barrier in their support.
−− Ensure that the contact between the shoe and shell is submerged.
−− Modify the design of the shoe to minimise the threat from sparking.

These ideas are shown in the following figures. If these ideas cannot be implemented,
companies may wish to propose alternatives which have the same effect.
Division of current in existing designs can promote sparking not only at the external
shunts, but at interfaces within the secondary seal (Figure 20a). The sparking in this region
may be eliminated by breaking the possible current paths with isolation (Figure 20b). The
primary shoe may also be modified so that sparking locations are not so exposed and ignitions
would quench (Figure 20c).

Phase 2:62
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

Figures 20a-c Secondary seal configurations.

Figure 20a Figure 20b Figure 20c

Secondary seal Secondary seal


Secondary seal

Shell
Floating roof
Shell

Shell
Isolating block between
pusher plate and primary Reshaped primary
shoe, or added
During roof or rim strike current will flow as shoe
upper rim
shown, and we would expect sparking at
several locations.

The shaped shoe modification would require a 1 mm to 2 mm step in the shoe for the upper
200 mm or so. Providing that the part maintains the gap mentioned, sparking that occurs
lower down will be quenched in the narrow gap.

4.4 COMPARISON WITH REAL TANK DESIGNS¹

Some comments are made here on typical installations, and the feasibility of implementing
adequate protection against the spark particles falling behind or close to the secondary
seals.
Tanks examined in phase 1 of the investigation were seen to be typically protected by
shunts either with one roof bonding cable lying upon the roof or by two retractable bonding
cables.

4.4.1 Resistances of typical bonding cables

Resistance values were not available for the designs assessed in phase 1 of the investigation,
so this should not be viewed as a comment on any particular manufacturer’s products.
However, the test results in this report show that resistance clearly is the crucial factor and
effort should be directed to keeping the resistance value down.
For the bonding cable it would be possible to use, for example, AWG0 gauge copper
wire (8,3 mm diameter), which has a rated resistance of 0,29 Ω per kilometre. A 50 m long
roof bonding cable would then have a resistance of about 15 mΩ. Using two or more cables
would reduce the resistance accordingly. It will also help current distribution if bonds are
made from the tank roof to two or more well separated locations on the shell rim.

1 This section is taken from Cul/LT-0373 Lightning simulation testing to determine the required characteristics
for roof bonding cables on external floating roof above ground storage tanks.

Phase 2:63
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

Clearly the resistance of any bonding connection must be kept low. It is practical to
achieve values of well below 1 mΩ, but care must be taken to ensure that the joint is well
sealed and does not corrode.
The material from which the cable is constructed is not relevant, as long as the
appropriate resistance can be achieved and maintained without corrosion issues.

4.4.2 Inductances of typical bonding cables

For values up to 100 µH it does not appear to be a significant parameter. The inductance
of a 50 m wire in free space is about 50 µH. As demonstrated in Cul/LT-0373 Lightning
simulation testing to determine the required characteristics for roof bonding cables on
external floating roof above ground storage tanks, an inductance of this order does not
hamper the protection the bonding cable offers from the slow component. The high energy
component is so fast that we should not expect any significant part of this component to
flow through the bonding cable.

4.4.3 Spooled bonding cables

Some installations use flat roof bonding cables which self spool, and which may have no
external insulation. It is difficult to comment on these without seeing one after some time in
operation. In principle the device can have the advantage of a low resistance and inductance
if current can freely flow between the adjacent spooled layers. However, contamination with
wax or oxide is an effect which would impede such current flow and should be considered,
especially as the resistance particularly needs to be relatively low. Of course with several such
devices installed the resistance of <10 mΩ should be achievable.
The conducting bearing/bushing should also be included in an assessment of the
conductive path. It would be a good idea to simply measure the inductance (L) and resistance
(R) of such devices in the condition in which they are normally found, and also in the retracted
and extended mode.
Such devices would have a minimal effect on conducting the fast, high energy
components, even with several around the rim. Therefore they would need to be used in
conjunction with shunts.

4.4.4 Ladders

Ladders may help protect the structure from the effects of lightning, and will present a fairly
low inductance, but due to the nature of the contacts (either sliding, with wheels or hinges)
the resistance of these interfaces is likely be high, or at best ill-defined with the possibility of
arcing. It is better not to assume credit for such a path.

Phase 2:64
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

5 CONCLUSIONS

5.1 GENERAL

From this investigation, there is now reasonable understanding of the mechanisms of


sparking. Sparking cannot be eliminated at shunts, but its intensity can be reduced by low
resistance roof bonding cables. Submerging shunts beneath the fluid surface can also provide
additional protection.

5.2 EXTERNAL SHUNTS (NOT SUBMERGED)

For external shunts (not submerged) arcing caused by the fast component is unavoidable and
is not significantly attenuated by a roof bonding cable. However, the fast component arcing
is primarily a hazard near its location at the shunt/shell interface.
Arcing caused by the slow, long duration components is more hazardous as showers
of burning particles are ejected, and these can fall down onto or behind the secondary seal.
This particular arcing threat may be reduced if low resistance roof bonding cable(s) are fitted.
The arcing is extinguished if the resistive voltage drop down the cable(s) is <14 V. Since the
average value of long duration current is considered to be 400 A (See CUL/LT-0235 Review of
tank base earthing and test current recommendations) then this would require the bonding
cable(s) to have a combined resistance of <35 mΩ, and it would be prudent to aim for 20
mΩ as a maximum (see section 4). The common industry practice of using more than one
bonding cable (for redundancy) will also help in this regard.

5.3 SUBMERGED SHUNTS

If shunts can be submerged, and if the contact point is submerged below 300 mm, an
ignition source at this interface is less likely, even without roof bonding cables. High level
tests were carried out in this configuration (using 44kA fast component, 2 kA slow, long
duration component) without ignition. Submerged shunts at shallower submersion depths
may result in burning particles and flame being ejected by slow current arcs, but the use of
roof bonding cables reduces this threat. So in summary:

Fast currents submersion at >100 mm gives improved protection.
Slow currents submersion at >300 mm gives improved protection.
submersion at 100 mm to 300 mm would require bonding cables.

In some installations there is a possibility that the shunts will not normally be the only current
path, so that some other fortuitous contacts may exist near or above the fluid surface. In
these cases any arcing is much more likely to occur outside the fluid.
For example in those installations with submerged shunts, but in which the scissor
mechanism is not isolated, shunts will not provide any means of protection as the current will
flow through the primary seal.

Phase 2:65
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

ANNEX A
RESULTS FROM TEST SERIES TO DEEPER SHUNTS

Test Configuration Depth Bonding cable? Flames Sparks Smoke I(kA) Q(C) I(kA) AI(A²s)
1 Setup shot n/a Yes n/a n/a n/a 1,85 108 - -
2 Setup shot n/a Yes n/a n/a n/a 1,86 115 - -
3 Shunt 40 mm Yes Yes Yes Yes 1,83 110 - -
4 Shunt 40 mm - Yes Yes Yes 1,85 109 - -
5 Shunt 40 mm - - - Yes 1,85 103 - -
6 Shunt 100 mm - Yes - Yes 1,85 106 - -
7 Shunt 100 mm Yes - - - 1,84 107 - -
8 Shunt 100 mm Yes - - Yes 1,85 108 - -
9 Shunt 100 mm Yes - - Yes 1,85 109 - -
10 Shunt 300 mm Yes - - - 1,74 107 - -
11 Shunt 300 mm Yes - - Yes 1,85 111 - -
12 Shunt 300 mm - - - Yes 1,85 102 - -
13 Shunt 300 mm - - - Yes 1,84 101 - -
14 Shunt 300 mm - - - Yes 1,84 99 - -
15 Shunt 300 mm - - - Yes 1,84 97 - -
16 Shunt 300 mm - - - Yes 1,85 103 - -
17 Shunt 300 mm - - - Yes 1,85 105 - -
18 Pusher (A) 300 mm - - Yes Yes 1,85 112 - -
19 Pusher (A) 300 mm - - Yes Yes 1,85 110 - -
20 Pusher (A) 300 mm - - Yes Yes 1,85 112 - -
21 Pusher (A) 300 mm Yes - Yes Yes 1,81 110 - -
22 Pusher (A) 300 mm Yes - Yes Yes 1,85 114 - -
23 Pusher (A) 300 mm Yes - Yes Yes 1,85 115 - -
24 Pusher (B) 300 mm Yes - Yes Yes 1,84 113 - -
25 Pusher (C) 300 mm Yes - Yes Yes 1,85 114 - -
26 Shunt 450 mm Yes - - Yes 1,84 108 - -
27 Shunt 450 mm Yes - - Yes 1,85 117 - -
28 Shunt 450 mm - - - Yes 1,85 101 - -
Shunt with
29 isolated cable 450 mm - - - Yes 1,84 108 - -
connector
30 Shunt 450 mm - - - Yes - - 6 1 250
31 Shunt 450 mm - - - Yes - - 5 1 250
32 Shunt 450 mm - - - Yes - - 11 5 877
33 Shunt 450 mm - - - Yes - - 21 23 610
34 Shunt 450 mm - - - Yes - - 21 23 723

Phase 2:66
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

Test Configuration Depth Bonding cable? Flames Sparks Smoke I(kA) Q(C) I(kA) AI(A²s)
35 Shunt 100 mm - - - Yes - - 11 6 205
36 Shunt 100 mm - - - Yes - - 15 12 432
37 Shunt 100 mm - - - Yes - - 15 17 991
38 Shunt 100 mm - - - Yes - - 32 53 305
39 Shunt 100 mm - - - Yes - - 44 103 500
40 Shunt 100 mm - - - Yes 1,84 92 14 11 230

Phase 2:67
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

ANNEX B
GASES EMITTED BY SUBMERGED ARCS

The following is an extract from Santilli Durable and efficient equipment for the production
of a combustible and non-pollutant gas from underwater arcs and method therefore² which
describes succinctly what is also discussed elsewhere on this topic.

“...The arc is generally produced by a DC power unit, such as a welder, operating at low
voltage (25-35 V) and high current (300 A to 3 000 A) depending on available Kwh. The high
value of the current brings to incandescence the tip of the carbon electrode in the cathode,
with consequential disintegration of the carbon crystal, and release of highly ionized carbon
atoms to the arc.
Jointly, the arc separates the water into highly ionized atoms of Hydrogen and
Oxygen. This causes in the immediate surrounding of the arc a high temperature plasma of
about 7 000 F, which is composed by highly ionized H, O and C atoms. A number of chemical
reactions then occur within or near said plasma, such as: the formation of the H2O2 molecule;
the burning of H and O into H2O; the burning of C and O into CO; the burning of CO and
O into CO2; and other reactions. Since all these reactions are highly exothermic, they result
in the typical, very intense glow of the arc within water, which is bigger than that of the
same arc in air. The resulting gases cool down in the water surrounding the discharge, and
bubble to the surface, where they are collected with various means. According to numerous
measurements conducted at various independent laboratories, the combustible gas produced
with the above process essentially consists of 45 %-48 % H.sub.2, 36 %-38 % CO, 8 %-10
% CO2, and 1 %-2 % O2, the remaining gas consisting of parts per million of more complex
molecules composed by H, O and C…”

2 Further information can be found at http://www.rexresearch.com/santilli/santilli.htm.

Phase 2:68
VERIFICATION OF LIGHTNING PROTECTION REQUIREMENTS FOR ABOVE GROUND HYDROCARBON STORAGE TANKS: PHASE 2

ANNEX C
FACILITY AND CALIBRATION

The long duration currents applied in this test series were generated using the Culham
Slow Bank. The output waveform of the Culham Slow Bank is monitored by a coaxial shunt
mounted in the return (low potential) transmission line.
Screened balanced twin cable was used to transmit the output from the coaxial
shunt to a screened diagnostic room where the signals were digitally processed, stored and
displayed. Further digital processing produced the values of peak current and charge transfer
associated with each test.

Culham A/D Bank


Transducer Dedicated Culham-manufactured Rogowski coil
Connection Screened coaxial transmission line
Conditioning Active integrator, serial number 95.8377-1/6
Acquisition W+W Signal Memory Recorder, channel 3
Model 108.245 HS
Serial number 1610528
Scaling Factor 43,29 kA/V
Latest System Confidence Check 01/12/06, See Calibration Certificate, Book 2/7

Culham Slow Bank


Transducer Culham-manufactured 1mΩ resistive shunt
Serial number 981012
Connection Screened coaxial transmission line
Conditioning Resistive termination circuit with differential output
Acquisition W+W Signal Memory Recorder, channel 2
Model 108.245 HS.
Serial number 1610528
Scaling Factor 1,029 kA/V
Latest System Confidence Check 01/12/06, See Calibration Certificate, Book 2/7

The voltage and currents measured during the tests were also recorded on the Nicolet Sigma
oscilloscope.
The current measurements are made by recording the voltage across a resistance of
approximately 0,25 Ω. The resistance was verified by cross calibrating with the main slow
bank measurement, although absolute current measurements through the shunt are not
quoted here, only the waveshapes and cut-off times are of interest.
Voltage measurements are made using voltage dividers and through fibre optic links;
the whole system of dividers and links was checked by comparing an input signal on one
Sigma scope channel with the same signal measured through the diagnostic chain. Based on
this a scaling factor of x 821 is applied to this voltage measurement.
These cross calibrations were carried out immediately prior to the main tests.

Phase 2:69
This publication has been produced as a result of
standards development work within the Subcommittees
on Electrical Equipment (SOEE) and Aboveground Storage
Tanks (SCAST) of the American Petroleum Institute.
1220 L Street Northwest SOEE and SCAST provide leadership in developing and
maintaining state-of-art electrical and aboveground
Washington, D.C.
storage tank standards use around the world.
20005-4070
Additional copies are available through IHS.
USA
Phone orders:
+1 202-682-8000 1-800-854-7179 (Toll-free in the U.S. and Canada)
+1 303-397-7956 (Local and International)
Fax Orders: +1 303-397-2740
API Product No: CLP2009 Online Orders: global.ihs.com
Information about API Publications, programs and services
is available at: www.api.org

This publication has been produced as a result of work


carried out within the Technical Team of the Energy
Institute (EI), funded by the EI’s Technical Partners. The
EI’s Technical Work Programme provides industry with
cost effective, value adding knowledge on key current
Energy Institute
and future issues affecting those operating in the energy
61 New Cavendish Street sector, both in the UK and beyond.
London W1G 7AR, UK
Additional copies of this title and other EI publications are
available online from:
t: +44 (0) 20 7467 7100 www.energyinstpubs.org.uk
f: +44 (0) 20 7255 1472 or the EI’s book distributors, Portland Customer Services:
t: +44 (0)1206 796 351
e: pubs@energyinst.org
e: sales@portland-services.com
www.energyinst.org

ISBN 978 0 85293 376 3