Hazards assessment of vegetable oil storage and processing plants

G. Landucci1, B. Nucci2, M. Pierini1, G. Pannocchia1, L. Pelagagge2, C. Nicolella1*

* c.nicolella@diccism.unipi.it

UNIVERSITA DI PISA

ICheaP10

8-11 May 2011 - Florence, Italy Italy ICheaP10 8-11 May 2011 - Florence,

Introduction (1)

The major part of the global vegetable oil production derives from extraction processes The process is based on the selective oil extraction usually carried out with a volatile organic solvent, typically hexane. After separation from solvent the oil phase is processed in chemical refining The vegetable crude oil usually presents a low content of residual solvent

SAFETY ISSUES PROCESS ISSUES

ICheaP10 8-11 May 2011 - Florence, Italy

Introduction (2)
Source Year Country (A) Location Operation

Accidents involving veg. oil hexane (FACTS, TNO 2007)

Fatalities Injuries Abstract >1 during cleaning of a tank on tanker after unloading cocosoil, some hexane vapours caused at least 1 casualty wear of pumpseal caused hexane release under the space of building, ignition, explosion and 4 casualties fire in extractor of soya factory 1 vapour cloud explosion on vessel in locks caused fire and 1 casualty hexane release at extraction plant for soya bean caused evacuation of the plant & after switch off (un)loaded power supply line an explosion occurred causing 27 casualties a fire in an elevator ignited gas from sunflower pellets in silo setting of an explosion causing more than 2 casualties explosion and fire at oil extraction plant while working in the cooker without the required permit flash fire injured 2 workers A combustible gas-air mixture was generated in an extractor by residual hexane content. Ventilation was not sufficient. Explosion occurred due to static electricity when the repair work started in the extractor. hexane leak caused heavy fire Welding operation inside a pomace oil tank

1975 The Netherlands harbour/port/dock Maintenance (cleaning)

(A) (A) (A)

1976 The Netherlands factory 1977 Germany factory

Processing Processing

1979 The Netherlands Inland navigation Maintenance

(A)

1980 Denmark

Factory

Processing

27

(A) (A) (A)

1983 Argentina 1983 Germany

storage/depot Factory

Storage Processing Processing

1 2

>1 11 2

1984 The Netherlands Factory

(A)

1991 Japan

Factory

Maintenance

>8

(A) (B)

1994 USA 2006 Italy

factory-yard

Processing (application)

11

Factory (refinery) Maintenance (storage)

4 ICheaP10 8-11 May 2011 - Florence, Italy

Introduction (3)

Spoleto, 2006
A catastrophic accident occurred in a vegetable oil refinery in Italy in 2006. During maintenance operations, an explosion occurred in a pomace olive oil- storage tank, causing 4 fatalities among personnel and damaging the surrounding equipment. The cause of the explosion was the accidental ignition of the residual solvent vapour during welding operations An accident with similar features occurred in Spain only few months later

The residual solvent may accumulate in the vapour phase, mixing with air and forming flammable mixtures which could result in the confined explosion of the storage vessel in presence of ignition.
ICheaP10 8-11 May 2011 - Florence, Italy

Introduction (4)

Loss of efficiency in high temperature vacuum operations (such as the drying and bleaching).

ICheaP10 8-11 May 2011 - Florence, Italy

Aims of the work:

Development of an approach able to quantitatively evaluate the hazards due to the residual solvent Safety issues related to storage formation of flammable mixtures inside the tanks Processing of vegetable oils determination of residual solvent accumulation among refining process

ICheaP10 8-11 May 2011 - Florence, Italy

Methodological approach
Estimation of liquid phase composition Thermodynamic model for the estimation of vapor phase composition Model validation against available experimental data

4a

Definition and evaluation of casestudies Determination of hazard indexes

5a

Safety issues in storage tanks

4b

Definition of critical sections for hexane accumulation Control and restoring measures

5b

Safety issues in oil refining

ICheaP10 8-11 May 2011 - Florence, Italy

Model set up (1)

Liquid phase characterization:

Triglycerides mixture (depending on oil type and origin) Free fatty acids (0.5-1 up to few percent) Residual solvent
Crude vegetable oil inside the storage tank

Schematization:

one reference triglyceride, namely LLP (constituted by two Linoleic

groups and one Palmitic group) one free fatty acid, namely oleic acid Thecnical hexane assumed a reference solvent and assimilated as pure n-hexane
ICheaP10 8-11 May 2011 - Florence, Italy

Model set up (2)

A well mixed air-hexane vapour phase is assumed in the top space of the vessel, due to open vent on the tank roof. Stratification is not considered in the present work, which assesses possible formation of flammable mixtures, without accounting for possible sources of ignition and the consequent effects following the ignition The analysis is referred to normal operation, the presence of oil sludge (sediment solids) is not considered in model.

ICheaP10 8-11 May 2011 - Florence, Italy

Model set up (3)

Equilibrium relationship: yi P

SAT i i i

T Pi SAT T

i 1,N c

1 2 3 4 5 6 7 8 9

Where:

yi : molar fraction of the i-th component in vapour phase; P : operative pressure (Pa);
i

: activity coefficient of the i-th component; : fugacity coefficient of the pure i-th component;

xi : molar fraction of the i-th component in liquid phase;

SAT i

Pi SAT : vapour pressure of the pure i-th component; T : operative temperature (K) N c : total number of components.

A modified version of the UNIFAC (UNIversal Functional Activity Coefficient) model was used to assess the activity coefficient

ICheaP10 8-11 May 2011 - Florence, Italy

Model validation

In order to test the validity of the model, a comparison with available experimental data was carried out. Available data ONLY at high hexane concentration (Extraction studies) The only available data for diluted solutions are reported by Smith & Wechter (J. of the American Oil Chemists Society 28, 381-382, 1951).
30

Experimental values
Hexane partial pressure (kPa)
25 20 15 10

Model estimates

RSC = 1.32% RSC = 1.08% RSC = 0.85%

Extrapolation

RSC = 0.62% RSC = 0.41%

5 0 15 30 45 60 75 90 105

RSC = 0.20%

120

Temperature (C) Temperature (C)

ICheaP10 8-11 May 2011 - Florence, Italy

Analysis of vegetable oil storage (1)

Vegetable oil refinery of SALOV S.p.A., located in Massarosa (Italy) The tank farm of the plant comprises 28 vessels with volumes ranging between 85 and 750 m3, with a total capacity of about 8500 tons.

The model was applied to the analysis of an industrial storage plant. The refinery processes sunflower, olive, peanut and corn oil with a maximum residual solvent content RSC of 0.1% by weight Higher RSC values were taken into account (e.g., 0.38% and 0.75% by weight). The tank farm is equipped with a heating system which prevents the vessel cooling to temperatures lower than about 18C in presence of cold weather. For temperatures higher than 18C the operative temperature can be assumed equal to the average ambient temperature. Typical weather conditions of the zone of the oil refinery were considered in the study

ICheaP10 8-11 May 2011 - Florence, Italy

Analysis of vegetable oil storage (2)

Estimation of vapor pressure in the tank according to external meteo conditions Possibility of having flammable mixtures for high RSC

ICheaP10 8-11 May 2011 - Florence, Italy

Analysis of vegetable oil refining (1)

Typical refining process conditions: purification of a sunflower oil assuming a free fatty acid content (FFAC) of 4 %wt. to be reduced up to 0.04 %wt.

Equipment: R: reactor; P: pump; CS: centrifugal separator; E: heat exchanger; S: flash separator; EJ : steam ejector; F: filter; C: column. Material streams : S_MP: medium pressure steam; S_LP: low pressure steam; EA: earths & activated carbons; W: water; SH: sodium hydroxide; HP: phosphoric acid; FO: feedstock oil; SW: soaps & waxes; SE: spent earths; FFA: free fatty acids; RO: refined oil. ICheaP10 8-11 May 2011 - Florence, Italy

Analysis of vegetable oil refining (2)

Process section Operative temperature ( C) Operative pressure (mbar)

Process description

Neutralization 20
Degumming 20

1000
1000

Adding sodium hydroxide to reach an intermediate grade of acidity

adding phosphoric acid and subsequently oil is sent to centrifugal separation to split the oil fraction from the solid waste.

Washing
Dehydration Discoloration

90
90 95

1000
50 60 2

The oil is washed with water

Oil is dehydrated in a flash separator under vacuum conditions Adding bleaching earth and activated carbon in a stirred vessel under vacuum conditions, then oil filtered physical neutralization with water steam at high temperature under vacuum conditions. The FFAC is stripped and condensed in a direct contact condenser, while steam with non condensable vapor are sent to the ejectors section.

Deodorization 230

ICheaP10 8-11 May 2011 - Florence, Italy

Simulation of vegetable oil refining (1)

Process simulations were performed using HYSYS V7.1 (by ASPEN Tech, Cambridge MA) simulation software.

The crude vegetable oil, including FFAC and hexane, was schematized according to the described thermodynamic model A residual hexane content of 0.1%wt. was considered as process input. The aim of the process simulation was to trace the hexane flow through each part of the process, with particular attention to units working at high temperature and low pressure, in which hexane vaporization is favoured (thus issuing safety problems in shut down and maintenance operations)
ICheaP10 8-11 May 2011 - Florence, Italy

Simulation of vegetable oil refining (2)

Sankey diagram identifying the critical nodes in which the accumulation of hexane is predicted The analysis evidenced the need to focus on the operative problems related to dehydration, discoloration and deodorization

The vacuum conditions are kept in these units by steam ejectors. The steam ejector model was set up on the basis of actual plant field data, based on typical operative conditions
ICheaP10 8-11 May 2011 - Florence, Italy

Sensitivity analysis

Determine the effect of the variation of a process variable to the refining process performance. Increasing content of hexane in feedstock oil was analyzed (up to 0.5%) Effect on the ejector system aimed at keeping low pressures in high temperature units. The process performance was analyzed through the dehydration efficiency

1,00

Dehydration efficency

0,95 0,90 0,85 0,80 0,75 0,70 0,000 0,001 0,002 0,003 0,004 0,005

Hexane weight fraction (w/w)

Potential decrease of 30% in process efficiency related to ejectors, thus evidencing the criticality of the residual solvent content in input feedstock.

ICheaP10 8-11 May 2011 - Florence, Italy

Conclusions

A quantitative methodology for the evaluation of residual volatile solvent content effect on storage and processing of edible extracted oils. A thermodynamic model was implemented and validated in order to reproduce the liquid-vapor equilibrium of crude vegetable oil - residual solvent system. Application to industrial case studies:

Identification of potential hazards due to formation of flammable mixtures inside the storage tanks The critical nodes of solvent accumulation were identified among the process, evaluating the influence on the global efficiency.

Acknowledgments

This research has received funding from Regione Toscana (Bando Unico R&S n.2009DUA/526090469/1)
ICheaP10 8-11 May 2011 - Florence, Italy

Maintenance operations on a crude vegetable oil tank

Thank you.
ICheaP10 8-11 May 2011 - Florence, Italy