M
echanical seals play a pivotal role in the chemical Radial space available Shaft Tap depth
process industries (CPI). As the sealing interface around shaft center line or stud
Bolt/stud to top of length
between rotating process equipment and the Bearing bearing
circle
materials contained inside, these devices prevent diameter Sleeve length
product leakage and fugitive emissions, while per- from mounting
mitting process machines, pumps and compressors surface
to operate at ever-increasing pressures, vacuums,
temperatures and speeds. Even the most basic Shaft/sleeve dia.
Recess
mechanical seal is a complex assembly. diameter Shaft length Vessel
Correct measurement of the machinery for (if any) from mounting I.D.
replacement seals can be an underappreciated Bearing surface
aspect of maintenance planning. shelf
to shaft Distance to Recess
The following steps outline basic procedures and Bolt/stud bearing from
Bearing center depth
considerations involved in measuring machinery for pattern height line
mounting surface (if any)
replacement seals.
S
Temperature differential
team tracer lines are designed to maintain the fluid in a
150 160 180 200 240 300 400 500 700 900 1050
primary pipe at a designated uniform temperature. In most
cases, these tracer lines are used outdoors, which makes 11 11
10 10
ambient weather conditions a critical consideration. 9 9
The primary purpose of steam traps on tracer lines is to retain 8 8
the steam until its latent heat is fully utilized, and then discharge 7 7
the condensate and non-condensable gases. As is true with any
6 6
piece of heat transfer equipment, each tracer line should have its
Positive Positive
Vane Pumps Rotor
150 60 Sliding vane pumps have a rotor with
100 40 radial slots, and it is positioned off-center
Centrifugal in a housing bore. Vanes that fit closely
Centrifugal into rotor slots slide in and out as the ro-
50 20
tor turns. Pumping action is caused by the
0 0 expanding and contracting volumes con-
0 50 100 150 0 250 500 750 1,000 tained by the rotor, vanes and housing.
Capacity, gal/min Viscosity, cSt Vane pumps are effective for low-
viscosity liquids, and when dry-priming is
required. They are not ideal for abrasive
C. Flowrate D. Head liquids. Vane Push rod
110 80
Positive
100 Positive
Efficiency, %
References
Flowrate, %
90 70
1. Petersen, J. and Jacoby R., Selecting a Positive Displacement Pump, Chem.
80 Eng., August 2007, pp. 42–46.
60
70 Centrifugal 2. Viking Pump Inc., When to use a Positive Displacement Pump, The Pump
60 Centrifugal 50 School Website, Viking Pump Inc., www.pumpschool.com, 2007.
50 3. Soares, C., “Process Engineering Equipment Handbook,” McGraw Hill,
40 40 New York, 2002.
0 100 200 300 400 500 55 80 105 4. “Perry’s Chemical Engineer’s Handbook,” 8th ed. McGraw Hill, New
Viscosity, cSt Feet of head York, 2008.
Low-Pressure
Measurement
Department Editor: Scott Jenkins
B
ecause of the behavior of materials under vacuum, including 5IFVTFGVMPQFSBUJOHSBOHFPGBDPOWFOUJPOBM#"UZQFHBHF
lowered boiling points, low-pressure operations are useful tools FYUFOETCFUXFFO–3BOEo torr.
in chemical processing, as well as in research and develop-
ment and manufacturing. Areas where a vacuum system is neces- Crossed-electromagnetic-field IG
sary or useful include the following: "OPUIFSCSPBEUZQFPGJPOJ[BUJPOHBHFTDBOCFDMBTTJmFEBT
t4FQBSBUJOHMJRVJET DSPTTFEFMFDUSPNBHOFUJDmFME &.'
*(T BMTPDBMMFEDPMEDBUIPEF
t1SPDFTTJOHUFNQFSBUVSFTFOTJUJWFNBUFSJBMT *(T
5IJT*(UZQFJTDIBSBDUFSJ[FECZUIFHFOFSBUJPOPGBEJT-
t1SFWFOUJOHVOXBOUFEDIFNJDBMSFBDUJPOTCFUXFFOQSPDFTTNBUFSJ- charge between cathode and anode that is maintained with a
als and the surrounding environment by removing air NBHOFUJDmFME
t*NQSPWJOHIFBUUSBOTGFSUISPVHIMJRVJETCZSFNPWJOHBJS "OFYBNQMFPGUIJT*(UZQFJTB1FOOJOHHBHF"1FOOJOHHBHF
t%SZJOHTPMVUFT SFNPWJOHTPMWFOU
maintains a cathode-anode discharge by applying a magnetic
t*OJUJBUJOHNBUFSJBMTUSBOTGFSPSBJEJOHmMUFSQSPDFTTFTCZBQQMZJOH mFME5IFNBHOFUJDmFMEQSFWFOUTUIFFMFDUSPOTGSPNUSBWFMJOH
differential pressures directly to the anode; instead, they travel in helical paths through
t%FQPTJUJOHUIJOmMNTPOTVSGBDFT
TVDIBTJOUIFNJDSPFMFDUSPOJDT the vacuum space.
industry and other areas 5IFFMFDUSPOTDBOJPOJ[FHBTNPMFDVMFTJOUIFWBDVVNXIFO
t5FTUJOHXJUINBTTTQFDUSPTDPQZ UIFZIBWFTVGmDJFOUFOFSHZ4JODFUIFSFTVMUJOHHBTJPOTIBWFNVDI
greater mass than electrons, they are unaffected by the magnetic
LOW-PRESSURE MEASUREMENT mFMEBOEBSFDBQUVSFE
5IFQSPCBCJMJUZPGDPMMJTJPOTJTQSPQPSUJPOBMUPHBTEFOTJUZ5IF
Because of vacuums’ central role in processing and research, it is current generated by the ion collection process is an indirect indica-
important to be able to measure pressures that may be a tiny frac- tion of vacuum pressure.
tion of typical atmospheric pressures. 5BCMFDPNQBSFTUIFBEWBOUBHFTBOEEJTBEWBOUBHFTPGFNJUUJOH
5BCMFTIPXTBUZQJDBMCSFBLEPXOPGWBDVVNRVBMJUZ5IF DBUIPEF*(TBOEDSPTTFE&.'*(T
SBOHFTEPOPUIBWFVOJWFSTBMMZBHSFFEVQPOEFmOJUJPOT
IPXFWFS
TABLE 2. ADVANTAGES AND DISADVANTAGES OF ECG
TABLE 1. TYPICAL RANGES FOR VACUUM QUALITY VERSUS CROSSED-EMF IONIZATION GAGES
I
nstalled gain graphs can help improve the selection of control valves by using the equations in the ISA/IEC valve sizing standard (ANSI/
for chemical processing. The graphs are plotted to analyze together ISA-75.01.01).
control-valve flow characteristics and process-system flow character-
istics, and better illustrate the relationship between a control valve and Step 4: Express the flowrate in terms of
the system. Predicting installed gain can help to increase controllability percent process variable (%PV)
of the system and help avoid oversized valves. Installed gain graphs Use the range of the process-variable measurement device and its rela-
can reveal ranges of valve travel where the valve gain might impede tionship to flowrate to determine the %PV for the installed characteristic
controllability. They can also show the travels for which the control valve graph points. For example, if the process variable is flowrate, divide
will perform optimally. each flowrate on the curve by the full span of the flowmeter.
2.0
Step 1: Determine the control valve’s inherent flow characteristic
1.8 Maximum
a) Inherent flow characteristic describes how the capacity of a control 1.6
gain
valve changes with valve travel. 1.4
b) The inherent flow characteristic plot has the same shape as valve
Gain
1.2 Minimum
gain
flow coefficient (CV) curve. Common curve shapes are: 1.0
Linear — Slope changes little over the normal working range of the valve 0.8
Quick-opening — Slope changes faster over first 25% of valve travel 0.6
and slower at high travels 0.4
Equal percentage — Slope 0.2
0
changes more slowly at 100 0 10 20 30 40 50 60 70 80 90 100
Percent of rated flow coefficient
high travels
en
op
Find the slope of the installed flow characteristic graph at each valve
ui
ar
Q
ne
Step 2: Determine sys- travel. The plot of Δ%PV / Δ%travel for each percent travel increment is
Li
losses. Plot flowrate vs. The installed gain graph can aid in the analysis of whether the control
ua
Eq
pressure. Assuming the valve inherent characteristic is suitable for the system. An installed gain
control valve is not un- 0
0 100 equal to one for the entire valve travel would indicate that the other
Percent of rated travel components of the control system would not have to compensate for the
dersized, it will have one
position that can fulfill both installed valve gain (that is, the control system tuning parameters used at
the flowrate and pressure conditions required by the system. one value of valve travel would allow equally acceptable controllability
at other travels).
System pressure characteristic It is more than likely the installed gain will not equal one across the
225 full valve travel. Guidelines for desirable installed gain values have
200
been established. In most cases, installed gain values of between 0.5
Maximum
and 2.0 should be the target.
175 Normal If the installed gain falls outside this range for valve travels that are
Pressure, psia
Minimum expected to be used for controlling the process, the controllability will
150
not be optimal. For example, controller tuning setpoints that function
125 P1 for valve well at low valve travel values might cause system instability if used at
travels with a high installed gain.
P2 for valve
100
75 DEFINITIONS
100 500 1,000 1,500 2,000 2,500 3,000 3,500 4,000
Flow, gal/min Valve gain — The change in flow for a given change in travel
Valve travel — The degree of openness of the valve; the valve stroke
Step 3: Determine installed flow characteristic graph Control range — The control range of an installed valve is the range of
a) Pressure conditions across a control valve are not constant. Values of travels for which the installed gain remains within the recommended 0.5
the liquid-pressure recovery factor and the pressure-drop ratio factor for to 2.0 range
control valves vary with valve travel. CV — The valve flow coefficient of a device (such as a valve) represents
b) For several values of valve travel, determine where on the system a relative measure of its efficiency at allowing fluid flow. It involves the
curve (flow versus pressure) the process will be operated and what the relationship between the pressure drop across a valve system and the
flowrate would be. The location on the system curve can be determined corresponding flowrate
Inherent flow characteristic — The relationship between control valve
Installed flow characteristic capacity and valve stem travel
3,000
Installed flow characteristic — Actual system flow plotted against valve
2,500 Maximum opening. Pressure drops vary with valve travel when valves are installed
with pumps, piping, fittings and other process equipment
2,000 Normal
Flow, gal/min
Minimum
1,500
References
1. Niesen, M., Using installed gain to improve valve selection. Chem. Eng.
1,000 October 2008, pp. 34–37.
500 2. Fitzgerald, B. and Linden, C., The control valve’s hidden impact on the
bottom line. Valve Manufacturer’s Association, Washington, D.C., 2003.
0 3. “Perry’s Chemical Engineer’s Handbook,” 8th ed., McGraw Hill, N.Y.,
0 10 20 30 40 50 60 70 80 90 100
Valve travel, % 2008.
Aboveground
and underground
Department Editor: Scott Jenkins storage tanks
T
he ultimate objective of storing liquid,
REGULATIONS AND CODES POTENTIALLY APPLICABLE TO UST AND AST
fluid and gaseous products, which may
be corrosive, flammable or unstable, is Regulation name Applicability Governing
to store material in an environmentally safe Body
and economically viable manner. Storage Resource Conservation and Recovery Act Subtitle C AST and UST EPA
tanks in the chemical process industries — hazardous waste regulations
(CPI) can be most broadly divided into
those buried underground, and those Spill Prevention, Control and Countermeasures within AST and UST EPA
Clear Air Act amendments of 1990
constructed aboveground. The following is
an outline of considerations associated with Hazardous Waste Operations and Emergency Response AST and UST OSHA
each category and positive and negative Regulations — 29 CFR 1910.120
aspects of each. Also included are poten-
Flammable and Combustible Liquids Standard — 29 CFR AST and UST OSHA
tially applicable regulations and codes from 1910.106
the U.S. Environmental Protection Agency
(EPA), the Occupational Health and Safety Hazard Communication Standard — 29 CFR 1910.1200 AST and UST OSHA
Administration (OSHA) and others. Confined Space Safety Standard — 29 CFR 1910.146 AST and UST OSHA
UNDERGROUND STORAGE TANK Oil Pollution Act of of 1990 AST only EPA
(UST) ADVANTAGES National Pollutant Discharge Elimination System AST only EPA
UST DISADVANTAGES International Code Council (ICC) International Fire Code AST and UST ICC code
Space use — more real estate required to Cost — Several aspects of cost should be
Construction — ASTs are simpler and less house ASTs considered, including: storage tank, tank
expensive to construct and install. One system equipment, initial installation cost,
significant aspect eliminated is the need for Temperature variation — ASTs experience ongoing maintenance cost, security cost,
excavation and special backfill materials greater ambient-product-temperature fluctua- land cost, regulatory cost and secondary
Maintenance — Visual inspection is pos- tions than USTs containment cost
sible, which leads to more reliable leak
detection and easier repair
MAKING THE CHOICE References:
1. Geyer, W.B. To Bury or Not to Bury: Steel
Insurance — Lower pollution insurance pre- When making decisions about whether to Tank Technology Decisions. In: “Handbook
miums because of reduced risk of ground- use an AST or a UST, consider the following: of Storage Tank Systems,” Marcel Dekker,
water contamination New York, 2000.
Regulatory — The local authority in the
Regulatory burden — ASTs are subject to 2. Cheremisinoff, P.N., and Vallamar, O. Aboveg-
area where the tank system will be as- round and Underground Storage Tank Com-
less regulatory requirements than USTs
sembled has control of whether a tank parison. In: “Storage Tanks. Advances in
Costs — ASTs have lower monitoring and permit will be issued, based on whether Environmental Control Technology Series,”
record-keeping costs plans for the tank comply with local, state Gulf Publishing, Houston, 1996.
Chemical
Resistance of
Department Editor: Scott Jenkins Thermoplastics
C
orrosion is estimated to account for a plications. However, plastics must be selected References
significant portion — 8–10% — of total based on process specifics. When chemicals 1. “Perry’s Chemical Engineers’ Handbook: 8th
annual-plant-capital expenditures for the affect plastics, it usually happens as: chemical ed.” McGraw Hill, New York, 2008.
chemical process industries. To avoid it, engi- solvation or permeation, where physical prop-
2. Nibco Chemical Resistance Guide. Nibco Inc.
neers can consider corrosion-resistant plastics erties may change, but the polymer molecular Elkhart, Ind. 2003.
for process piping and storage vessels. Ther- structure is not chemically altered; and direct
3. Plastics Pipe Institute. Chemical Resistance of
moplastics are generally resistant to chemical chemical attack, where a chemical reaction Thermoplastics. Techincal Report 19. 2007.
attack and thus suitable for many process ap- with the polymer occurs.
EXPANSION TANK satisfactorily by relatively simple “on-off” attention must be paid to achieving turbu-
or “high-low” controllers. However, units lent flow (without stagnation zones) around
A well designed expansion tank: of all sizes will operate more uniformly if the heat transfer surfaces to eliminate hot
1. Maintains a static pump-suction head equipped with modulating temperature con- spots and localized fluid boiling.
2. Compensates for temperature-related trols. Install user controls to regulate the flow
volume and pressure changes of the heat transfer fluid in proportion to VALVES
3. Provides a means of venting moisture the energy consumption of the equipment.
and low boilers Cast- or forged-steel valves with 13-chrome
In a multiple-user system, separate controls
4. Prevents fluid oxidation trim are satisfactory for service in organic
should be installed on each consuming unit
Usually, the expansion tank is installed heat-transfer-fluid systems. Globe valves
to assure the proper energy delivery.
at the highest point in the system and is with an outside screw (as a protection
connected to the suction side of the pumps. FIRE RESISTANCE against high temperatures) should be used
It should serve as the main venting point of throughout the system when tight sealing of
the system, as well as provide for system Materials considered NOT fire-resistant: fluids is desired, and should be installed up-
fluid expansion, which can be 25% or t-PXNFMUJOHQPJOUNFUBMTBMVNJOVN
stream and downstream of each pump and
more depending on fluid choice and on the copper at each user. Gate valves are acceptable
operating temperature range. For heat- t&MBTUPNFST for use, but not to provide reliable positive
ing circuits, the expansion tank should be t1PMZUFUSBnVPSPFUIZMFOF 15'&
HBTLFUTBOE shut-off.
sized so that it is one-quarter full at ambient packing The use of valve stem seals can be effec-
temperature and three-quarters full when t/POBTCFTUPT
mCFSSFJOGPSDFESVCCFS tive in minimizing system leakage. For valve
the system is at operating temperature. For bound gaskets stem seals:
cooling, it should be vice versa. Materials considered fire resistant: t'MFYJCMFHSBQIJUFQBDLJOHXJUIJOOFS
The double drop-leg expansion tank t)JHINFMUJOHQPJOUNFUBMTDBSCPOBOE and outer anti-extrusion rings of braided
provides greater flexibility of operation stainless steel, nickel alloys graphite fiber gives the best results for
than a single leg tank. From a single-leg t'MFYJCMFHSBQIJUFQBDLJOHBOEHBTLFUT elevated temperature systems.
expansion tank, venting of nonconden- t"TCFTUPTQBDLJOHBOEHBTLFUT t15'&QBDLJOHPGUFOXPSLTJOTZTUFNT
sibles is often difficult in heating systems operating up to 400°F (200°C).
as is purging of air and water on startup. HEATER t.FUBMCFMMPXTTFBMFEWBMWFTBSFGSFRVFOUMZ
A double-leg expansion tank provides unin- used with excellent results, but these
Two basic fired-heater designs for indirect valves are relatively expensive, especially
terrupted flow on startup and significantly
heat transfer systems are available: liquid in larger sizes.
improves the venting capability of the
tube and fire tube types. t'JCFSQBDLJOHNBUFSJBMTIBWFHJWFOQPPS
system. All vent lines should be routed to a
t*OMJRVJEUVCFIFBUFST
nVJEJTQVNQFE performance in service with organic
safe location.
through the tubes as it is heated. The fire fluids, and are not recommended.
Experience indicates that systems with
is outside the tubes.
expansion tanks open to the atmosphere
t*OmSFUVCFIFBUFST
nVJEnPXTUISPVHIUIF References and further reading
have fluid contamination problems related
heater “shell” around the outside of the 1. Wagner W. “Heat Transfer Technique with
to oxidation and excessive moisture. There-
fire tubes. Organic Media.” Grafelfing-München: Technis-
fore, open expansion tanks should not be cher Verlag Resch KG, 1977.
Liquid tube heaters are preferred at all
employed in systems using organic heat 2. Gamble CE. Cost Management in Heat Trans-
temperatures. At temperatures below
transfer fluids. fer Systems. Chemical Engineering Progress,
500°F (260°C), fire tube heaters with a
An effective way to minimize fluid oxida- July 2006 pp. 22–26.
special baffle design to eliminate hot spots
tion is to blanket the expansion tank vapor 3. “Systems Design Data.” Pub. #7239193ver. C,
can be used.
space with an inert gas (for example, nitro- Solutia Inc., 1999.
Two basic configurations for electrical
gen, CO2, or natural gas). When using a 4. “System Design and Maintenance.” Pub. #TBS
heaters are available: container design and 10-25 (E), Solutia Inc., 1998.
nitrogen blanket, moisture should be driven
tubular design. 5. “Liquid Phase Design Guide.” Pub.
off from the fluid before the gas pressure is
t*OUIFDPOUBJOFSEFTJHO
POFPSNPSFFMFD- #7239128C, Solutia Inc., 1999.
set. If this is not practical, air contact can
trical heating elements are inserted into a
be minimized by a cold seal trap arrange-
container through which fluid flows. Acknowledgment
ment. Low boilers and moisture can collect
t*OUIFUVCVMBSEFTJHO
UIFIFBUJOHFMFNFOUT
in the cold seal trap, so the fluid in the trap Material for this “Facts at Your Fingertips” was
are inserted longitudinally into tubes supplied by Solutia Inc.
should be discarded periodically.
through which the fluid flows.
CONTROLS The tubular design is preferred for the heat-
NOTICE: Although these recommendations are
ing of organic fluids. If the container design believed to be correct, Solutia Inc. makes no
Install heater controls to regulate the firing is to be used, due to the unpredictable representations or warranties as to the complete-
mechanism in direct proportion to the flow conditions around the elements, heat ness or accuracy thereof. Nothing contained
required output. These controls should flux should be limited to 1–2 W/cm2. For herein is to be construed as a recommendation
to use any product, process, equipment or
increase or decrease the heat input to all heater types, the maximum heat flux at
formulation in conflict with any patent, and So-
maintain the heat transfer fluid at the oper- the surface of the heat source and the fluid lutia Inc. makes no representation or warranty,
ating temperature required by the energy velocity over it should be in proper balance express or implied, that the use thereof will not
demand. Small units may be operated to avoid excessive film temperature. Careful infringe any patent.
S
NOMENCLATURE
uccessful adsorber design requires
information about certain adsorbent 1
E A Henry’s law coefficient
1 Aj
properties, including: adsorbent densi- E B empirical coefficient
ties and void fractions; isotherms or other Bij
1
E C Concentration (subscripts: 0, initial;
equilibrium data; and data on mass-transfer 1 Ai (7) f, final; i, instantaneous)
kinetics and fixed-bed dynamics. Here, E
dp particle diameter
these adsorbent properties are explained,
DAB adsorbate diffusivity in the liquid
along with corresponding equations. KINETICS DATA Deff effective diffusivity
DENSITIES AND VOID FRACTIONS Kinetics data describe the intraparticle F fractional conversion
mass-transfer resistance and control the i assigned to strongly adsorbed com-
Three densities are relevant: bulk, particle cycle time of a fixed-bed adsorption ponent
and solid. There are four pertinent void frac- process. Fast kinetics, or a high rate of j assigned to less strongly adsorbed
tions, three corresponding to the three den- diffusion, occur when the effluent concen- component
sities, and one representing the overall void tration remains level until the adsorbent is jD Colburn-Chilton j-factor
fraction in the bed of adsorbent (). These almost saturated, then rises sharply. Slow k fluid-to-particle mass transfer coeffi-
terms are related in the following way: kinetics, on the other hand, occur when the cient
effluent concentration begins fluctuating
R B 1
E B R P 1
E B 1
E P R S soon after adsorption starts. The effect of
l particle radius (for a sphere or cyl-
inder), or half-thickness (for a rect-
1
E R S (1) slow diffusion can be overcome by add- angular solid)
ing adsorbent at the product end, or by
L bed length
Bulk density is the mass of adsorbent in a increasing the cycle time. It is also possible
given volume. Particle density is the mass to use small particles. n* adsorbent loading
of adsorbent per volume occupied by the Intraparticle diffusion is characterized Re Reynolds number
particle. Solid density is the mass of the by an effective diffusivity Deff, which equals Sc Schmidt number
adsorbent per volume occupied by the par- DABερ/τ. Dimensionless time is expressed t elapsed time
ticle, but with the pores’ volume deducted. as Deff t/l 2, which allows the determination xi mole fraction in fluid phase
of the fractional change (during adsorption
yi mole fraction in adsorbed phase
EQUILIBRIUM DATA or regeneration), as shown below.
α selectivity, 1<α<infinity
Adsorption-equilibrium data are usually β selectivity, 0<β<1
gathered at a fixed temperature and plot- ε void fraction (subscripts: B, bulk; P,
ted or tabulated as isotherms: adsorbent particle; S, solid)
capacity (loading) versus fluid-phase μ fluid viscosity
concentration (or partial pressure, for gases
ρ fluid density (subscripts: B, bulk; P,
and vapors). Below are three equilibrium (8) particle; S, solid)
equations for isotherm data.
τ tortuosity
Henry’s Law: This equation shows that, when searching υs superficial velocity
for an effective adsorbent, it is usually safe ψ particle shape factor
n* AC (2) to choose one having a large diffusivity, a
small diameter, or both.
The particle shape factor, ψ, is 1.0 for
Langmuir Isotherm:
beads, 0.91 for pellets and 0.86 for flakes.
AC FIXED-BED DYNAMICS αi is the ratio of particle interfacial area to
n* (3)
1 BC volume, which equals 6(1 – ε)dP.
Interstitial mass transfer in fixed beds,
Another major factor in bed dynamics is
primarily fluid-to-particle transfer, can be
pressure drop. Most adsorbers are designed
Freundlich Isotherm: related to the fluid, adsorbent and system
to operate with relatively low pressure drop,
properties via either of two equations,
n* AC B because large particles are used whenever
(4) depending on the value of the system’s
possible and the velocity is typically low,
Reynolds number.
to allow equilibration of the fluid with the
Henry’s Law is the simplest isotherm,
adsorbent. The pressure drop in a fixed bed
while the Langmuir Isotherm takes surface (9)
is represented by the Ergun Equation, below.
coverage into account, and the Freundlich
Isotherm is the result of fitting isotherm data For 10<Re<2,500, use Equation (10),
to a linear equation on log-log coordinates. where Sc equals μ/ρDAB. ΔP ⎛ 1 − εB ⎞⎛ ρυ 2s 1 − ε B ⎞
= ⎜150 + 175⎟⎜⎜ ⎟
⎠⎝ gc dp ε B3 ⎠⎟
In situations that involve multiple adsor-
L ⎝ Re
bate components in the feed, the concept jD k Us Sc0.667 1.17 Re
0.415
of selectivity comes into play. Selectivity, α (10)
(12)
and β in the three equations below, is the
References
ratio of the capacity of an adsorbent for For lower flowrates, corresponding to
1. Knaebel, K., The Basics of Adsorber Design,
one component to its capacity for another. values of a modified Reynolds number, Re’,
Chem. Eng., April 1999, pp. 92–101.
of below 50, use Equation (11).
yi y j 2. Knaebel, K., For Your Next Separation,
A ij Consider Adsorption, Chem. Eng., November
xi x j (5) jD 0.91Y Re'
0.51 (11) 1995, pp. 99–102.
3. Masel, R.I., “Principles of Adsorption and Reac-
A ij Ai A j (6) (11a) tion on Solid Surfaces,” Wiley-Interscience,
New York, 1996.
Flowmeter
Selection
Department Editor: Kate Torzewski
W
hen a flowmeter is needed, the selec-
Fluids Tem-
tion process should include study- Accuracy
Turn- (liquid, Pipe Sizes,
Maximum
perature
ing the characteristics of respective Flowmeter (full scale, pressure,
down gas, solid, in. range,
measurement technologies and analyzing the F; rate, R) psig
slurry) (º F)
advantages/disadvantages for different plant
environments. This effort will help ensure Square-edged,
orifice differential 0.5–1.5% R 4 to 1 L, G, S 0.5–40 8,800 –4–2,300
that a meter with the right performance and pressure
reliability, for a particular installation, is
selected. Some of the most common industrial Electromagnetic 0.2–2% R 10 to 1 L 0.15–60 5,000 –40–350
flowmeter designs are described here.
Turbine 0.15–1% R 10 to 1 L, G 0.5–30 6,000 –450–600
O
ne of the most important factors to con-
Metal UNS Coefficient of Thermal Ex- Temperature Melting tem-
sider when evaluating a specialty metal Number pansion (10–6 mm/(mm°C) range, °C perature, °C
for use in valves, fittings and instrumen-
tation is its corrosion resistance or corrosion Nickel alloy 200 N02200 13.3 20–90 1,440–1,450
rate of the metal in the target media. Each Titanium R50250 8.6 0–100 1,705
material discussed here has its own niche,
Zirconium R60702 5.2 0–100 1,860
and depending on the processing environ-
ment and the needs of the application, the Tantalum R05200 6.5 0–200 2,996
metal’s strengths and weaknesses may or
owes its corrosion resistance to the natural NICKEL ALLOYS
may not be suitable. In either H2SO4 or
formation of a dense, stable, self-healing
HCl solutions, the corrosion resistance of Nickel alloys are commonly used when typi-
oxide film on its surface. Unalloyed zirconi-
tantalum is second to none, followed by cal steel materials don’t offer the corrosion
um has excellent resistance to H2SO4 up to
zirconium, nickel alloys and titanium. Ther- performance that is needed. To enhance the
60% concentration at the boiling point, and
mal expansion coefficient and melting point performance of nickel in aqueous-solution
has excellent corrosion resistance in HCl.
data for each metal are given in the table. service, the most important alloying ele-
Zirconium is also highly resistant to most
alkali solutions up to their boiling point. ments are Fe, Cu, Si, Cr and Mo. Cr and
TANTALUM Zirconium’s corrosion resistance could Mo play a major role in nickel’s corrosion
be compared with titanium in many ways, resistance. Varying the concentrations of
The physical properties of tantalum are these elements in the nickel alloys changes
but it is much more robust than titanium in
similar to those of mild steel, although the corrosive environments in which nickel
withstanding organic acids, such as acetic,
tantalum has a higher melting point. Tan- alloys can be successfully applied, but they
citric, and formic at various concentra-
talum is the most corrosion-resistant metal are typically used in a range of acid, salt
tions and elevated temperatures. However,
that is in common use, due to its tenacious and alkali applications. The addition of Cr
zirconium can still be corrosively attacked
oxide layer. Its superb corrosion resistance (15–30%) improves the corrosion resistance
by fluoride ions, wet chlorine, aqua regia,
is comparable to glass and is practically to oxidizing solutions, while the addition of
concentrated sulfuric acid (above 80%),
inert to most oxidizing and reducing acids, Mo (up to 28%) improves the resistance to
and ferric or cupric chlorides [3]. Zironium
except fuming sulfuric acid, hot alkalis non-oxidizing acids.
has excellent resistance to reducing environ-
and HF. While tantalum is an ideal choice The nickel alloys C-22, C-276, and B-2
ments, but oxidizing agents frequently
from a corrosion-resistance point of view, all have good corrosion resistance in a
cause accelerated attack. Commercial-
it is typically cost prohibitive, even when variety of media. In the case of HCl, the
grade zirconium, which contains up to
cladded. Only in process conditions where corrosion resistance of these alloys depends
2.5% hafnium, is often used in hydrogen
no other material will perform adequately is greatly on the Mo content. The alloy with
peroxide production, rayon manufacture,
tantalum a material of choice, at least in its the highest concentration of Mo, B-2, exhib-
and the handling of phosphoric acid, sulfu-
traditional forms. This limits tantalum’s use its the best corrosion resistance.
ric acid and ethyl benzene.
to heating coils, bayonet heaters, coolers In solutions such as nitric acid (HNO3),
and condensers operating under severe Cr is an essential alloying element for pro-
conditions. When economically justified, TITANIUM viding corrosion resistance. Nickel alloys’
larger items of equipment, such as reactors weaknesses revolve around their interaction
or tanks, may be fabricated with tantalum Titanium is an established metal when deal- with the media and their environment in
liners. Since tantalum linings are usually ing with corrosive applications. Titanium is the form of impurities. Under ideal testing
very thin, very careful attention to design available in a range of different alloys with conditions, these alloys (for example, B-2),
and fabrication details is required. the most-corrosion-resistant grades being work well in pure de-aerated H2SO4 and
Tantalum can typically be found in ap- titanium 7, 11 (containing 0.15% Pd), and HCl, but deteriorate rapidly when oxidizing
plications that deal with hot concentrated 12 (containing 0.3% Mo and 0.8% Ni). impurities, such as oxygen and ferric ions,
acids. Due to its negligible corrosion rate, Titanium and its alloys offer good corro- are present. Another consideration is the
it is ideal for use in the pharmaceutical and sion resistance that is due to a strong oxide presence of chlorides (Cl–), which generally
food manufacturing industries [2]. film. The oxide film formed on titanium is accelerate the corrosion attack at different
Recently, tantalum has been processed to more protective than on stainless steel, and degrees for various alloys.
create a surface alloy on valves and other it often performs well in media such as Having a wide range of applicability in
fittings, instrumentation and equipment. This seawater, wet chlorine and organic chlo- acids, salt solutions, and caustic environ-
relatively new option exhibits all the chemi- rides. While titanium offers good corrosion ments, nickel alloys have found their way
cal properties of tantalum, allowing excel- resistance to these solutions, it certainly is into a variety of industries, such as chemi-
lent corrosion resistance without the high not immune to them, especially at elevated cal, petrochemical, oil and gas, nuclear,
costs. However, this option is not suitable temperatures (for example, seawater at conventional power generation and paper.
for slurries or solutions that contain abrasive temperatures greater than 110°C) [3]. It has
particles that could lead to mechanical ero- a number of disadvantages as well, as it is References
sion and abrasion of the surface. not easy to form, it has a high springback 1. Gambale, D., Valves & Specialty Metal Materi-
and tends to gall, and welding must be car- als, Chem. Eng., October 2008, pp. 38–41.
ZIRCONIUM ALLOYS ried out in an inert atmosphere. 2. Burnstein, G.T., others, “Corrosion Volume 1 &
Titanium metal can be found in a variety 2”, Butterworth Heinemann, Oxford, 1994.
Zirconium alloys exhibit excellent resistance of industries, including chemical processing, 3. Roberge, P.R., “Handbook of Corrosion Engi-
to corrosive attack and work well in many pulp and paper, and marine applications. neering”, McGraw Hill, New York, 2000.
organic and inorganic acids, salt solutions, It is also used extensively in the production 4. “Perry’s Chemical Engineers’ Handbook,” 8th
strong alkalies, and some molten salts. It of chlorine. ed. New York: McGraw Hill, 2008.
I
n today’s typical process plants, prevent-
ing steam loss and improving condensate
return are key opportunities to make a
process more energy efficient.
To be the most effective, steam gener-
ally needs to be dry (such as for process High-
usage), or superheated (for instance, for temperature
use in turbines). These requirements dictate fluid
utility-system operating procedures for
generating the highest quality steam pos-
sible, and then distributing it to the points
of use with minimal deterioration. Since
steam becomes condensate after its heat
energy is expended, strategies must be in Trap Trap Trap
place to remove condensate as quickly as it
FIGURE 1. Preferred method to drain jacketed pipe for
is formed, in the steam-supply portion of the
high-melting-point fluids, such as sulfur
circuit and during steam usage alike.
Furthermore, superheated steam is FIGURE 2. Alternative, practical redesign method for ex-
typically desuperheated by injecting hot isting installations to drain jacketed pipe for high-melting-
condensate into the system. As a result, ex- point fluids, such as sulfur (no “stall”)
cessive wetness can also occur downstream High-
of the desuperheating station. In either temperature
fluid
case, if such condensate is not removed
from the steam supply, the negative impact
on the steam system can be substantial, as
seen in Table 1.
Improving condensate return. At many
plants, the operators admittedly realize that
condensate must be removed as quickly
as it is formed, but a suitable condensate
drainage or transportation system is not Trap Trap Trap
in place. In such cases, the condensate is
often sewered or sent to a field drain. Some TABLE 1. RESULTS OF FAILING TO REMOVE CONDENSATE
possible outcomes of removing condensate FROM THE STEAM SUPPLY
but not handling it effectively are outlined in Loss of yield: Entrained water does not carry as much heat to a process as does steam
Table 2. Damage to nozzles: Entrained water erodes nozzles and can adversely affect vacuum
Condensate is traditionally removed from generation or atomization
steam systems by steam traps or by equip- Loss of power: Entrained water causes turbines to operate less efficiently
ment combinations involving level pots and Increase maintenance loading; Water hammer can damage equipment such as turbine
outlet control valves. Process situations in blades and control-valve packing
which high backpressure from the down- Increased safety risk: Water hammer can injure personnel
stream portion of the condensate-return Poor process control: Flooding exchanges can lead to control swings
system tend to create a “stall.” Then, a
different system incorporating both a pump
TABLE 2. RESULTS OF REMOVING CONDENSATE,
and trap in the design is needed to drive
BUT FAILING TO HANDLE IT EFFECTIVELY
the condensate while also trapping the
Profit loss due to waste of heated and treated condensate
steam; this process may be referred to as
The extremely wasteful effect of opening bypass valves around process equipment or
pump-trapping or power-trapping.
turbines to prevent waterlogging or damage
Because there are at least three
A possible increase in system corrosion because too much makeup water must be
condensate-drainage alternatives, it makes treated
more sense to think in terms of required
“condensate discharge locations” rather where steam traps cannot meet the high cally arises when modulating steam pres-
than referring to condensate removal de- pressure or capacity requirements. sure creates a negative pressure differential
vices indiscriminately as “steam traps.” This Special situations. There can be many situa- across the condensate drain device. So-
broader mind-set helps avoid any predispo- tions in a plant where effective condensate called, Type II secondary pressure drainers
sition to install steam traps in applications removal requires specialized drainage of the pump-trap type are used on equip-
that need a different type of condensate designs. For instance, Figures 1 and 2 ment with a negative pressure differential.
drainage solution. show two options for condensate drainage Because wasted condensate is a valuable
Engineered separator-drains remove from a jacketed pipe that conveys high- resource to be saved, use Type I secondary
condensate that is entrained in a moving melting-point materials, such as liquid sulfur pressure drainers of a “pump only” type to
steam supply (including flash or regener- or high-boiling hydrocarbons. recover collected condensate and power it
ated steam). The result is highest quality Other examples of specialized applica- back to the boiler.
steam delivered for plant use. Compare that tions include options to effectively drain
to steam traps, which remove condensate steam-supplied heat exchangers. A key References
that has already fallen out of the steam. As consideration is to first determine whether 1. Risko, J., Handle Steam More Intelligently,
their name suggests, steam traps remove a stall condition exists or not; when it does, Chem. Eng., November 2006, pp. 38–43.
condensate and “trap steam.” Meanwhile, condensate will not drain effectively through 2. Aggarwal, S., Boost Energy Efficiency In Plant
level pots can be used in certain instances a simple steam trap. Such a situation typi- Utilities, Chem. Eng., April 2002, pp. 70–73.
Membrane
Configurations
Department Editor: Kate Torzewski
M
embrane polymers are pack- CHOOSING A CONFIGURATION Tubular
aged into a configuration,
In selecting a membrane configu- Porous
commonly called a device Membrane substrate
or an element. The most-common ration, it is important to consider
element configurations (figure) are how the packing density and con- Concentrate
tubular, capillary fiber, spiral wound, centration polarization of each Feed
Diameter multiplier, ψ
Colebrook [3] and Blasius [4] are written as (13) 1.5
Equations (4) and (5), respectively. Smooth pipes. Substituting Equation (5) for f 1.4
Re yields a correlation for pipe diameter.
1.3
(4) 1.2
(14) 1.1
(5)
Commercial pipes. Determining the diameter 1.0
Commercial pipe. In this case, f is governed
of a rough pipe requires the use of Gu, the 0.9
by both Re and relative roughness, expressed 1,000 1x104 1x105 1x106
dynamic roughness. 1 10 100
as F/D. The Colebrook-White’s Equation (6) is Gu
used to calculate f [3]. (15) Figure 1 is a plot with Gu as abscissa and : as
Manipulating Equation (7) to reflect Gu and ordinate. It has the expected limit at Gu = 0, at
substituting into the expression for pipe di- : = 1. First, estimate a pipe diameter assuming
(6)
ameter gives Equation (17), commercial pipe smooth pipe. Using this diameter, calculate Gu from
As this equation requires trial-and-error solu- diameter. Several design parameters can be Equation (15). Then, from Figure 1, get the value
tion, Altshul [5] has developed Equation (7), a condensed into a constant, named M.
of : and multiply it by the diameter to get actual
computationally simpler choice. diameter needed for the commercial pipe.
H
ere, we present criteria needed for TABLE 1. Guidelines for Selection of Tray Spacing
the intelligent, effective specification
of trays for a distillation or stripping Description Tray Spacing, mm Comments
column. Column diameters The tray support beams restrict
Among the key parameters and acces- larger than 3,000 > 600 crawling space available; hence
mm the large tray spacing
sories during specification are column
diameter, tray spacing, number of trays, Column diameters This spacing is sufficiently wide to
number of passes, type of downcomers, between 1,200 600 allow a worker to freely crawl be-
and 3,000 mm tween trays
weir heights, provision of downcomer
backups, tray pressure drops, design Crawling between the trays is sel-
Column diameters
dom required, because the worker
of bottom seal pan, column bottom between 750 and 450
can reach the column wall from
arrangement, and nozzle location and 1,200 mm
the tray manways
orientation.
Fouling and cor-
> 600 Frequent maintenance is expected
rosive service
TRAY SPACING
Systems with a At least 450 mm,
Required to avoid premature
The vertical distance between adjacent high foaming but preferably 600
flooding
trays varies from 450 to 900 mm in tendency mm or higher
the columns generally employed in the Columns At least 450 mm,
Required to avoid excessive
chemical process industries (CPI), as operating in spray but preferably 600
entrainment
seen in Table 1. regime mm or higher
Spacing involves a tradeoff between Columns Lower tray spacing restricts
column height and column diameter; with operating in froth < 450 allowable vapor velocity, thereby
regime promoting froth-regime operation
greater tray spacing, the column height
increases, while the required diameter
decreases. cross-sectional area should not be speci- exceed the clearance normally employed
Height/diameter ratios greater than fied. Also, the downcomer width should under the tray downcomers. It should be
25–30 are generally not recommended. not be less than 10% of the column at least 50 mm.
diameter. The distance that the downcomer
TRAY PASSES Specify the downcomer clearance to extends downward within the seal pan
be less than the outlet weir height; other- should be about the same as the clear-
Set the number of passes such that liquid wise, vapor will flow up the downcomer ance between downcomer bottom and
loads do not exceed 70 m3/h per meter rather than through the tray deck above. pan floor.
weir length. Downcomer sealing: To achieve a proper The distance between bottom tray floor
After the number of passes has been downcomer seal, the bottom edge of and seal pan floor should be 150 mm
selected, if necessary, adjust the column the downcomer should be about 10 mm greater than the normal tray spacing.
diameter in order to arrive at a minimum below the top edge of the outlet weir.
path length of 400 mm. NOZZLE ORIENTATION
t'PSBDPMVNOEJBNFUFSPG
UP The downcomer clearance should be
2,100 mm: use two passes at maxi- selected such that the liquid velocity 'PSOP[[MFTUIBUGFFEMJRVJEJOUPUIFUPQ
mum under the downcomer does not exceed tray, the nozzle should be perpendicular
t'PSBEJBNFUFSPG
UP
NN 0.45–0.50 m/s. to the downcomer of the top tray.
use three passes at maximum Nozzles feeding liquid at intermediate
OUTLET WEIRS trays can be placed anywhere except
t'PSBEJBNFUFSBCPWF
NNVTF
four passes at maximum in the downcomer segment. The space
t8FJSIFJHIUTJOUIFGSPUISFHJNFBSF
between the two trays should be at least
restricted to 50–80 mm
DOWNCOMERS 800 mm.
t8FJSIFJHIUTGPSDPMVNOTPQFSBUJOHJO
Downcomers are conduits having the spray regime should be 20–25 mm
8FJSMPBEJOHTTIPVMEGBMMXJUIJOUIF COLUMN BOTTOM
circular, segmental or rectangular cross
sections that convey liquid from a tray to range of 15 to 70 m3/h per meter weir Inlets for the bottom feed and reboiler
the one immediately below. length. return lines should be at least 300 mm
Liquid velocities in downcomers: above the high liquid level.
PRESSURE DROP
t.JOJNBMMZGPBNJOHMJRVJET The bottom feed and reboiler return
0.12–0.21 m/s 'PSUSBZTUPGVODUJPOSFBTPOBCMZDMPTFUP should not impinge on the bottom seal
t.FEJVNGPBNJOHMJRVJET their best efficiency point, the dry-tray pan, seal pan overflow, or the bottom
0.09–0.18 m/s pressure drop must be roughly equal to downcomer.
t)JHIMZGPBNJOHMJRVJET the hydraulic-tray pressure drop. The tops of both pipes should be at
0.06–0.09 m/s least 400–450 mm below the bottom tray.
SEAL PANS
Downcomer areas and clearances'PS
References
the cross-sectional area of the downcom- The clearance between the seal pan .VLIFSKFF
4
5SBZ$PMVNO%FTJHO,FFQ$POUSPMPGUIF
ers, values less than 5–8% of the column floor and the bottom downcomer should Details, Chem. Eng., September 2005, pp. 52–58.
Fluid Flow
DEFINITIONS ¨ N N4 · NOMENCLATURE
Newtonian fluid. A fluid is known to be f 16 ©1 1 He
1 3 He7 ¸
NRe © 6 N Re 3 f NRe ¸ (11) a Dimensionless parameter
Newtonian when shear stresses associated with d ª d d¹
flow are directly proportional to the shear rate A Dimensionless parameter
DVR B Dimensionless parameter
of the fluid NRed (12)
Power law fluid. A structural fluid has a structure
Md D Diameter, m
that forms in the undeformed state, but then D2 RT 0 f Fanning friction factor, dimensionless
breaks down as shear rate increases. Such a N He (13) fL Laminar friction factor, dimensionless
fluid exhibits “power law” behavior at intermedi-
M d2
fT Fully turbulent friction factor, dimensionless
ate shear rates
TURBULENT PIPE FLOW fTr Transition friction factor, dimensionless
Bingham plastic fluid. A plastic is a material that
g Gravitational acceleration, m/s2
exhibits a yield stress, meaning that it behaves
as a solid below the stress level and as a fluid Since most turbulent flows cannot be analyzed L Length of cylinder or pipe, m
above the stress level from a purely theoretical perspective, data and m Consistency coefficient, (N)(s)/m2
generalized dimensionless correlations are used. n Power law fluid flow index, dimensionless
Newtonian fluid. The friction factor for a NHe Hedstrom number, dimensionless
LAMINAR PIPE FLOW Newtonian fluid in turbulent flow is a function of NRe Reynolds Number, dimensionless
both NRe and the pipe relative roughness, ε/D,
For steady flow in a pipe (whether laminar or NRe,pl Power law Reynolds Number, dimensionless
which can be read off the Moody diagram [5].
turbulent), a momentum balance on the fluid gives NRe,plc Power law Reynolds Number at transition
The turbulent part of the Moody diagram (for
the shear stress at any distance from the pipe from laminar to turbulent flow, dimensionless
NRe > 4,000) is accurately represented by the
centerline.
Colebrook equation (14). NRe∞ Bingham-plastic Reynolds Number,
T rx $&r
T w r (1) dimensionless
2L R 1
4 log ¨ E D 1.255 · (14) P Pressure, Pa
© ¸
In Equation (1), Φ = P + ρgz. The volumetric f ª¾ 3.7 NRe f ¹¾ Q Volumetric flowrate, m3/s
flowrate Q can be related to the local shear rate
When NRe is very large, the friction factor depends r Radial position in a pipe or a cylinder, m
by doing an integration by parts of Equation (2).
R only on ε/D. This condition is noted with fT as the R Pipe or cylinder radius, m
Q
P ∫¾ r G rx dr
2˙
(2) “fully turbulent” friction factor in Equation (15). V Velocity, m/s
0 z Vertical elevation above a horizontal refer-
1 = – 4 log ⎡ ε D ⎤ (15)
Newtonian fluid. For a Newtonian fluid, ⎢⎣ 3.7 ⎥⎦ ence plane, m
fT α Dimensionless parameter
τrx = μγtrx, which gives the following volumetric
flowrate, known as the Hagen-Poiseuille equation. The Churchill Equation [2] represents the entire γtrx Shear rate in tube flow, s–1
4 Moody diagram, from laminar, through transition
Q
P$&D (3) flow, to fully turbulent flow. It is presented here as
ε Wall roughness, m
128 ML μ Newtonian viscosity, Pa–s
Equations (16), (17), and (18).
It can be written in dimensionless form in Equation 1
μ∞ Bingham Plastic limiting viscosity, Pa–s
(4) with the two terms defined in Equations (5) ¨¥ ´
12 · 12 ρ Density, kg/m3
and (6). f 2 ©¦ 8 µ 1
1.5
¸ (16) τ0 Yield stress, N/m2
f = 16 / NRe (4)
©§ NRe ¶
ª A B ¸
¹ τrx Stress due to force in x direction acting on r
16 surface, N/m2
P 2 D5
$& ¨ ¥ ´· τw Stress exerted by fluid on tube wall, N/m2
f (5) ©
32RLQ2 ¦ µ¸ Φ Flow potential, P + ρgz, Pa
A ©2.457 ln¦ 1 µ¸
© (17) ΔΦ Ιncrease in flow potential, Pa
4QR ¦ ¥ 7 ´ 0.9 0.27E µ ¸
NRe (6) © ¦§ ¶ µ¸
PDM ª¾ § Re D ¶ ¹¾
Bingham plastic. For the Bingham plastic, fT is solely
Power law. A fluid that follows the power law 16
¥ 37, 530 ´ a function of NRe∞ and NHe, as represented by
model obeys the relationship τrx = –μ(–γtrx)n. This B¦ (18)
µ Equations (26)–(29).
gives the following equation. § NRe ¶
1 1
3n+ 1
τ
Q=π ⎞ w⎞
n
⎞ n ⎞R n (7)
Power law. For a power-law fluid, the friction fac-
tor depends only upon Equation (9) and the flow
f f Lm fTm m (26)
27
3 n
RQ 2
n
fL 16
;
a
14.7 1 0.146 e
2.9 s 10
5
N He = (28)
NRe, pl (9) (20)
n NRe, pl
¨3n 1 · m 1.7
40, 000 (29)
mP 2
n D 4
3n © n¸
ª ¹ 0.0682n
0.5 NRe
fT 1 (21)
Bingham plastic. In this case, there is a solid-like [ NRe, pl ] (1.87 2.39 n)
“plug flow” region from the pipe centerline (where References
τrx = 0) to the point where –τrx = τ0 (that is, at r 0.414 0.757 n [
5.24 n]
1.Darby, R., Take the Mystery Out of Non-Newtonian
= r0 = R x τ0/ τw). The result is a flow integral fTr 1.79 s 10
4 NRe, pl e Fluids, Chem. Eng., March 2001, pp. 66–73.
(22)
modified from Equation (2). For a Bingham plas- 2.Churchil, S. W., Friction Factor Equation Spans all Fluid-
Flow Regimes, Chem. Eng., November 1997, p. 91.
tic, –τrx = τ0 + μ∞(–γtrx). Using this expression and A 1
(23) 3.Darby, R., and Chang, H. D., A Generalized Correla-
the modified flow integral, the Buckingham-Reiner 1 4
$ tion for Friction Loss in Drag-reducing Polymer Solutions,
Equation (10) is found. AIChE J., 30, p. 274, 1984.
$ NRe, pl
NRe, plc (24)
4
PR3T w ¨© 4 ¥ T 0 ´ 1 ¥ T 0 ´ ·¸ 4.Darby, R., and Chang, H. D., A Friction Factor Equation
Q 1
¦ µ ¦ µ (10) The value of NRe where transition from laminar to for Bingham Plastics, Slurries and Suspensions for all
4 M d © 3 § T w ¶ 3 §T w ¶ ¸ turbulent flow occurs (NRe,plc) is given by Equa- Fluid Flow Regimes, Chem. Eng., December 28, 1981,
ª ¹ pp. 59–61.
tion (25).
The equivalent dimensionless form is given by 5.Darby, R., “Fluid Mechanics for Chemical Engineers,”
Equations (11), (12) and (13). NRe, plc 2, 100 875(1
n) (25) Vol. 2, Marcel Dekker, New York, N.Y., 2001.
Biodiesel
Production
Department Editor: Kate Torzewski
B
iodiesel can be produced from formation and the separation of
vegetable oils by three types glycerin downstream.
of reactions: base catalyzed Methanol Catalyst
transesterification of the oil; direct Separation. Glycerin and biodiesel
acid-catalyzed transesterification of are the two main products of reac-
the oil; and conversion of the oil to its tion, with each containing an amount
fatty acids, and then to biodiesel. of unreacted alcohol. Since the Mixer
Soybean oil
Biodiesel is typically produced by glycerin phase is much more dense
a base-catalyzed reaction (Figure 2). than biodiesel phase, the two phases
This method of production has several can be separated by gravity in a
advantages, including the following: settling vessel, with glycerin simply
low temperature (150˚F) and pressure drawn off the bottom of the settling Reactor
Excess methanol
(20 psi) reaction that requires only vessel. Alternatively, a centrifuge can
standard materials of construction; be used to separate the two materials
direct conversion to biodiesel with no more quickly.
intermediate compounds; and high
conversion (98%) with minimal side Glycerin neutralization. The Settler
reactions and a low reaction time. separated glycerin contains unused
In the chemical reaction for base- catalyst and soaps, which are neutral- Biodiesel Glycerin
catalyzed biodiesel production, ized with an acid. Water and alcohol
vegetable oil is reacted with a short are removed to produce glycerin
chain alcohol (signified by ROH) at 80–88% purity to sell as crude Neutral-
in the presence of a catalyst to glycerin. Alternatively, glycerin can Wash
ization
produce glycerin and biodiesel. The distilled to 99% purity or higher for
fatty acid chains associated with selling to the cosmetic and pharma-
the oil, which are mostly palmitic, ceutical industries.
stearic, oleic, and linoleic acids for
naturally occurring oils, are repre- Methyl ester wash. After the
sented by R', R'' and R''' (Figure 1). biodiesel is separated from glyc-
Distillation
Distillation
erin, residual catalyst or soaps can
be removed with a gentle warm
PRODUCTION STEPS water wash.
Mixing of alcohol and cata-
Alcohol removal. Unreacted alco-
lyst. The catalyst is typically sodium
hol in both the glycerin and biodiesel
hydroxide (caustic soda) or potassium
phases is removed by flash evapora- Biodiesel Glycerin
hydroxide (potash). It is dissolved in
tion or distillation. The recovered alco-
the alcohol using a standard agitator
hol is then reused for mixing with the ASTM specifications. Additionally, all
or mixer. Methanol or ethanol is com-
catalyst. Alcohol removal can occur biodiesel produced must be regis-
monly used as the alcohol.
after the wash and neutralization, as tered with the U.S. Environmental
shown in Figure 2 to the right, but it Protection Agency (Washington, D.C)
Reaction. The mixture of alcohol
can occur before these steps as well. under 40 CFR Part 79.
and catalyst is charged into a closed
reaction vessel, and the oil is added.
Product quality and registra- References
The reaction mix is kept just above the
tion. Prior to use as a commercial 1. Biodiesel Production & Quality Standards,
boiling point of the alcohol, 160°F,
fuel in the U.S., the finished biodiesel July, 2008. National Biodiesel Board,
to speed up the reaction, although it www.biodiesel.org/resources/fuelfactsheets
must be analyzed to ensure it meets
is sometimes recommend to run the
reaction at room temperature. The
reaction time can vary from 1–8 h. CH2OCOR’’’ CH2OH R’’’ COOR
Excess alcohol is used to ensure total Catalyst
conversion of the oil to its esters. CH2OCOR’’ + 3 ROH CH2OH + R’’ COOR
The amount of water and free
CH2OCOR’ CH2OH R’ COOR
fatty acids in the incoming oil must
be monitored, because if either Vegetable Alcohol Glycerin Biodiesel
level is too high, it can inhibit soap oil
Heat Transfer
System Design
Department Editor: Kate Torzewski
/05*$&"MUIPVHIUIFJOGPSNBUJPOBOESFDPNNFOEBUJPOTTFUGPSUIIFSFJO IFSFJOBGUFSi*OGPSNBUJPOw
BSFQSFTFOUFEJOHPPEGBJUIBOECFMJFWFEUPCFDPSSFDUBTPGUIFEBUFIFSFPG
4PMVUJB*ODNBLFTOP
SFQSFTFOUBUJPOTPSXBSSBOUJFTBTUPUIFDPNQMFUFOFTTPSBDDVSBDZUIFSFPG*OGPSNBUJPOJTTVQQMJFEVQPOUIFDPOEJUJPOUIBUUIFQFSTPOTSFDFJWJOHTBNFXJMMNBLFUIFJSPXOEFUFSNJOBUJPOBTUPJUTTVJUBCJMJUZ
GPSUIFJSQVSQPTFTQSJPSUPVTF*OOPFWFOUXJMM4PMVUJBCFSFTQPOTJCMFGPSEBNBHFTPGBOZOBUVSFXIBUTPFWFSSFTVMUJOHGSPNUIFVTFPGPSSFMJBODFVQPO*OGPSNBUJPOPSUIFQSPEVDUUPXIJDI*OGPSNBUJPO
SFGFST/PUIJOHDPOUBJOFEIFSFJOJTUPCFDPOTUSVFEBTBSFDPNNFOEBUJPOUPVTFBOZQSPEVDU
QSPDFTT
FRVJQNFOUPSGPSNVMBUJPOJODPOnJDUXJUIBOZQBUFOU
BOE4PMVUJBNBLFTOPSFQSFTFOUBUJPOPSXBS-
SBOUZ
FYQSFTTPSJNQMJFE
UIBUUIFVTFUIFSFPGXJMMOPUJOGSJOHFBOZQBUFOU/03&13&4&/5"5*0/4038"33"/5*&4
&*5)&3&913&4403*.1-*&%
0'.&3$)"/5"#*-*5:
'*5/&44'03"1"35*$6-"3
163104&030'"/:05)&3/"563&"3&."%&)&3&6/%&38*5)3&41&$550*/'03."5*0/035)&130%6$5508)*$)*/'03."5*0/3&'&34
C
rystallization is a method of solid- supersaturated solution that is free of
fluid separation in which pure foreign matter. First, molecules in the
Nomenclature
chemical crystals are formed. solution will associate into a micro- A Frequency factor
Crystallization kinetics consists of scopic cluster, which will either dis- As Surface area of crystal
three major phenomena: nucleation sociate or continue to grow. When the B0 Rate of homogeneous pri-
(the birth of a crystal), transfer of the cluster develops until it forms a lattice mary nucleation
solute from the supersaturated solution structure, it is then called an embryo. A
c Mass solute concentration
to the crystal surface, and a reaction stable crystalline nucleus is established
in the bulk supersaturated
during which the solute becomes incor- when the crystal size exceeds Dp given
solution
porated into the crystal. by the Kelvin equation for the solution’s
specific supersaturation ratio. cs Mass solute concentration
in the solution at satura-
Combining the Kelvin equation with
SUPERSATURATION tion
laws of chemical kinetics gives the rate
of homogeneous nucleation, which is c/cs Supersaturation ratio
In order to drive the process of crystal- described as follows: dm/dt Rate of mass deposited on
lization, the solution must be supersatu- the crystal surface
rated with solute. Dp Crystal diameter
The solubility of the solution, related kc Mass transfer coefficient
to crystal size, is defined by the Kelvin
ki Kinetic coefficient
equation:
kN, b, j, r Constants determined ex-
perimentally
MT Mass of crystals per vol-
ume of magma
Secondary nucleation N Agitation rate
Industrial crystallizers typically rely on Na Avogadro’s number
Supersaturation can be quantified by secondary nucleation, which is caused R Gas Constant
the ratio of the mass solute concen- by the presence of existing crystals in s Relative supersaturation
tration in the bulk solution to the the supersaturated solution. This can
concentration in the solution at the t Time
occur by one of three mechanisms in
point of saturation. Alternatively, it which nuclei are removed from a crys- T Temperature
can be described by relative super- tal surface, including the following: s,L Interfacial tension
saturation, which is calculated by the —fluid shear Molar volume of crystals
s
following equation: —collision of crystals with each other Number of ions/molecule
—collision of crystals with metal of solute
surfaces
Solute diffusivity
The phenomena of secondary nucle-
ation is too complex to derive a simple Solution density
kinetic theory, so an empirical power-
NUCLEATION law function has been developed to
describe this process, which is based incorporated into the crystal-lattice
Nucleation occurs at the point that a structure. By combining these steps, the
on experimentally derived constants for
crystal begins to form. The relative rate of mass transfer in crystal growth
a particular system:
rates of nucleation and growth are is expressed as follows:
critical to crystallization kinetics, as
they determine both crystal size and
size distribution. Nucleation is catego-
rized as primary if the supersaturated
solution is free of crystals, or second- Through a series of calculations, this
CRYSTAL GROWTH
ary if the solution already contains expression can be used to define the
crystals. Primary nucleation requires a particle size of a crystal at any time,
Based on data proving that a solution
higher level of supersaturation and is assuming that Dp0 << Dp:
in contact with a crystal is supersatu-
the principal mechanism occurring in rated, there is an accepted two-step
precipitation. Secondary nucleation oc- theory of crystal growth, referred to as
curs in commercial crystallizers, where the diffusion-reaction theory. In the first
crystalline surfaces are present in order step, mass transfer of solute from the
to produce large crystals. solution to the crystal-solution interface
occurs. In the second step, the kinetic
References
Primary nucleation step, a first-order reaction occurs at the
1.Seader, J. D. and Henley, E. J., “Separation
Primary nucleation, when occurring surface of the crystal, during which sol- Process Principles,” 2nd ed., John Wiley and
homogeneously, takes place in a ute molecules from the solution become Sons, Inc., New Jersey, 2006.
Valves
GLOBE
The basic principle of globe valve
operation is the perpendicular
movement of the disk toward, or
away from, the seat. This causes
the annular space between the
disk and seat ring to gradually
Source for figures: Valve Manufacturers
close as the valve is closed. It is Association of America, Washington, D.C.
this characteristic that gives the
Butterfly Check
globe valve good throttling ability. valve valve
When the valve is closed, there BUTTERFLY
is no blocked-in volume, as occurs
in a gate valve, so a globe valve Butterfly valves get their name from
has much less leakage around the winglike action of the flow-con-
the seat. Also, the disk-to-seat-ring trolling disk that opens and closes
contact is much closer to forming at right angles to the flow path.
right angles, so the force of clos- Butterfly valves were introduced to
ing tightly seats the disk. counteract the problems associated
with linear-valve designs (espe-
PINCH cially gate valves), such as the
relatively large size and weight,
This valve consists of a flexible to be the valve least likely to Ball valves are available in the high operating force required,
tube that is mechanically pinched cause contamination. For this Venturi, reduced and full-port and the tendency to leak. Instead
from the outside of the valve reason, it is popular in high- patterns. The full-port pattern has of a long stroke, the butterfly valve
body. The principal advantages of purity applications. a ball with a bore equal to the requires a quarter turn to cycle
this type of valve are that the flow It is available in two general inside diameter of pipe. Most ball from a fully open to fully closed
passage is straight without any designs, weir and straightway. valves instead have a reduced position. Butterfly valves can be
crevices, and there are no internal The weir-style diaphragm valve is bore with a Venturi shaped flow used for both on/off and throttling
moving parts. The soft valve body utilized for higher-pressure applica- passage of about three quarters applications.
has the ability to seal around tions. The straightway diaphragm the nominal valve size.
trapped solids, so pinch valves valve, having no flow path obstruc- CHECK
are suitable for handling slurries tions, is well suited for higher-flow PLUG
and solids, which would clog in and slurry applications. The purpose of a check (or non-
the obstructed flow passages of Plug valves have a cylindrical or return) valve is to allow fluid flow
other valve types. They are also BALL tapered plug with a hole bored in one preferred direction and
used for the sanitary handling of through. As with ball valves, fluid to prevent back flow, or flow in
foodstuffs and pharmaceuticals This rotational-motion valve uses flows when the hole in the plug is the opposite direction. Ideally, a
because the media are isolated a ball-shaped disk with a hole aligned with the pipe, and a quar- check valve will begin to close as
from the working parts. bored through to stop or start fluid ter turn of the plug stops the flow. the pressure drops in a pipeline
flow. When the valve handle is Plug valves often have fluoro- and the fluid momentum slows.
DIAPHRAGM turned to the open position, the carbon seating materials and in When the flow direction reverses,
ball is rotated so that the hole some cases are fully lined with the check valve should close
The flow passage in diaphragm lines up with the valve body’s fluorocarbons, which provides completely. Check valves can be
valves is free of crevices and inlet and outlet. When the ball is excellent protection for corrosive of the following types: swing, lift
is unobstructed by moving rotated so the hole is perpendicu- applications that require bubble- and tilting disk.
parts, making them suitable for lar to the flow, the valve is closed. tight shutoff. There are several
applications where cleanliness, Because the ball moves across different types of plug valves References
bubble-tight shutoff and chemical the seats with a wiping motion, commonly used in the CPI, includ- 1.Sahoo, T., Pick the Right Valve,
compatibility are important. The ball valves can handle fluids with ing lubricated, nonlubricated and Chem. Eng., August 2004, pp.
diaphragm valve is considered suspended solids. eccentric types. 34–39.
Vacuum
Pumps
Department Editor: Kate Torzewski
V
acuum is any system of reduced FIGURE 1.-JRVJE3JOH1VNQ FIGURE 2.3PUBSZ$MBX1VNQ FIGURE 3.
pressure, relative to local 3PUBSZ
(typically atmospheric) pressure. -PCF
Achieved with a pump, vacuum 1VNQ
systems are commonly used to:
Source: Kurt
t3FNPWFFYDFTTBJSBOE J. Lesker Co.
its constituents
t3FNPWFFYDFTTSFBDUBOUTPS FIGURE 4. 3PUBSZ4DSFX1VNQ
unwanted byproducts
t3FEVDFUIFCPJMJOHQPJOU
t%SZTPMVUFNBUFSJBM
t$SFBUFBQSFTTVSFEJGGFSFOUJBMGPS
JOJUJBUJOHUSBOTQPSUPGNBUFSJBM
-JRVJESJOHBOEESZQVNQTPGGFSUIF
NPTUBEWBOUBHFTGPSUIFDIFNJDBM
QSPDFTTJOEVTUSJFT $1*
#PUIPG
Source: Medical Gas Info Source: Gardner Denver Source: Kurt J. Lesker Co.
UIFTFQVNQUZQFTIBWFCFBSJOHT
Hanover, Inc.
TFBMFEPGGGSPNUIFQVNQJOHDIBN-
ber and do not require any internal UPXBSEUIFEJTDIBSHFTJEFPGUIF void space between the rotors and Rotary Screw. 5XPMPOHIFMJDBM
lubrication because the rotors do pump, the pockets decrease in QVNQIPVTJOH0OUIFOFYUSPUBUJPO
rotors in parallel rotate in opposite
OPUDPOUBDUUIFIPVTJOH#PUI
XIFO TJ[F
BOEUIFFWBDVBUFEHBTJTDPN- UIBUTBNFUSBQQFETBNQMFPGHBTJT EJSFDUJPOTXJUIPVUUPVDIJOH
TZO-
FNQMPZJOHBDPPMBOUTZTUFN
QSFWFOU QSFTTFE
FOBCMJOHJUTEJTDIBSHF DPNQSFTTFEBOEEJTDIBSHFEBTUIF DISPOJ[FECZIFMJDBMUJNJOHHFBST
UIFDPPMBOUGSPNDPOUBDUJOHUIFQSP- 5IFSJOHPGMJRVJEOPUPOMZBDUTBT EJTDIBSHFQPSUPQFOT 'JHVSF
(BTnPXNPWFTBYJBMMZ
DFTTnVJEBOEDBVTJOHDPOUBNJOB- a seal; it also absorbs the heat of "NJOJNVNPGUISFFTUBHFTJO BMPOHUIFTDSFXXJUIPVUBOZJOUFSOBM
tion, and both use mechanical shaft compression, friction and condensa- series is required to achieve pres- compression from suction to dis-
seals for containment. UJPO1PQVMBSMJRVJEDIPJDFTJODMVEF sures comparable to those of an DIBSHF1PDLFUTPGHBTBSFUSBQQFE
XBUFS
FUIZMFOFHMZDPM
NJOFSBMPJM oil-sealed mechanical pump. Some within the convolutions of the rotors
BOEPSHBOJDTPMWFOUT ESZEFTJHOTVTFUXPUFDIOPMPHJFTJO BOEUIFDBTJOH
BOEUSBOTQPSUFEUP
LIQUID-RING PUMPS DPNCJOBUJPOGPSFYBNQMF
BSPUBSMZ UIFEJTDIBSHF$PNQSFTTJPOPDDVST
DRY PUMPS lobe as a booster for a claw pump. BUUIFEJTDIBSHFQPSU
XIFSFUIF
*OUIFDZMJOESJDBMCPEZPGUIFQVNQ
USBQQFEHBTNVTUCFEJTDIBSHFE
BTFBMBOUnVJEVOEFSDFOUSJGVHBM 3PUBSZDMBX
SPUBSZMPCFBOE Rotary Lobe. The rotary-lobe pump
BHBJOTUBUNPTQIFSJDQSFTTVSF&BDI
GPSDFGPSNTBSJOHBHBJOTUUIFJOTJEF rotary-screw pumps dominate as 'JHVSF
JTUZQJDBMMZVTFEBTBNF-
convolution of the rotor acts similarly
PGUIFDBTJOH 'JHVSF
ESZQVNQTJOUIF$1*
QBSUJDVMBSMZJO DIBOJDBMCPPTUFSPQFSBUJOHJOTFSJFT
UPBTUBHFJOTFSJFTXJUIUIFPOFCF-
The source of that force is a MBSHFSTJ[FQVNQBQQMJDBUJPOT with an oil-sealed piston or vane
IJOEJUBUMFBTUUISFFDPOWPMVUFEHBT
multi-bladed impeller whose shaft Rotary Claw. 5IFHFPNFUSJDTIBQF QVNQUPCPPTUQVNQJOHDBQBDJUZBU
pockets in the rotor are required to
is mounted so as to be eccentric to PGUIJTQVNQBMMPXTGPSBHSFBUFS low pressures. achieve acceptable vacuum levels.
UIFSJOHPGMJRVJE#FDBVTFPGUIJT compression ratio to be taken This pump consists of two sym-
eccentricity, the pockets bounded BDSPTTUIFSPUPSTBUIJHIFSQSFTTVSFT metrical two-lobe rotors mounted on
by adjacent impeller blades (also 'JHVSF
5XPDMBXSPUPSTSPUBUF separate shafts in parallel, which References
DBMMFECVDLFUT
BOEUIFSJOHJODSFBTF in opposite directions of rotation rotate in opposite directions to each 7
JMCFSU
1
.FDIBOJDBM1VNQTGPS
7BDVVN1SPDFTTJOH
Chem. Eng.
in size on the inlet side of the pump, XJUIPVUUPVDIJOH
VTJOHUJNJOHHFBST PUIFSBUIJHITQFFET5JNJOHHFBST
0DUPCFS
QQo
BOEUIFSFTVMUJOHTVDUJPODPOUJOVBMMZ UPTZODISPOJ[FUIFSPUBUJPO5IFHBT are used to synchronize the rotation "
MJBTTP
+
$IPPTFUIF3JHIU
ESBXTHBTPVUPGUIFWFTTFMCFJOH FOUFSTUISPVHIBOJOMFUQPSUBGUFSJU of the lobes to provide constant 7BDVVN1VNQ
Chem. Eng..BSDI
evacuated. As the blades rotate has been uncovered and fills the clearance between the two.
QQo
Advantages Disadvantages
t4JNQMFSEFTJHOFNQMPZTPOMZPOFSPUBUJOHBTTFNCMZ t.JYJOHPGUIFFWBDVBUFEHBTXJUI
Liquid-Ring Vacuum Pumps
t$BOCFGBCSJDBUFEGSPNBOZDBTUBCMFNFUBM UIFTFBMJOHnVJE
t.JOJNBMOPJTFBOEWJCSBUJPO t3JTLPGDBWJUBUJPOSFRVJSFTBQPSUJPO
t-JUUMFJODSFBTFJOUIFUFNQFSBUVSFPGUIFEJTDIBSHFEHBT PGQSPDFTTMPBEUPCFOPODPOEFOT-
t/PEBNBHFGSPNMJRVJEPSTNBMMQBSUJDVMBUFTJOUIFQSPDFTTnVJE BCMFVOEFSPQFSBUJOHDPOEJUJPOT
t.BJOUFOBODFBOESFCVJMEJOHBSFTJNQMF t)JHIQPXFSSFRVJSFNFOUUPGPSNBOE
t4MPXSPUBUJPOBMTQFFE
SQNPSMFTT
NBYJNJ[JOHPQFSBUJOHMJGF NBJOUBJOUIFMJRVJESJOH
SFTVMUJOHJO
t$BOVTFBOZUZQFPGMJRVJEGPSUIFTFBMBOUnVJEJOTJUVBUJPOTXIFSFNJO- MBSHFNPUPST
HMJOHXJUIUIFQSPDFTTWBQPSJTQFSNJTTJCMF t"DIJFWBCMFWBDVVNJTMJNJUFECZUIF
t/PMVCSJDBUJOHMJRVJEJOUIFWBDVVNDIBNCFSUPCFDPOUBNJOBUFE WBQPSQSFTTVSFPGTFBMBOUnVJEBU
t"DDPNNPEBUJPOPGCPUIDPOEFOTBCMFWBQPSTBOEOPODPOEFOTBCMFT
UIFPQFSBUJOHUFNQFSBUVSF
XIJMFPQFSBUJOHBTCPUIBWBDVVNQVNQBOEDPOEFOTFS t1PXFSDPOTVNQUJPO
t3VHHFESPUPSEFTJHO
DPOTUSVDUFEPGTUVSEZDBTUPSEVDUJMFJSPOXJUIPVU t$BOOPUIBOEMFQBSUJDVMBUFNBUUFS
BOZnJNTZSPUBUJOHDPNQPOFOUT OPSMBSHFTMVHTPGMJRVJE
Dry Vacuum Pumps
t/PODPOUBDUEFTJHOGBDJMJUBUFECZUJNJOHHFBST t.BZSFRVJSFBTJMFODFS
t)JHISPUBUJPOBMTQFFESFEVDFTUIFSBUJPPGHBTTMJQUPEJTQMBDFNFOU
t.BZEJTDIBSHFHBTFTBUIJHI
JODSFBTFTOFUQVNQJOHDBQBDJUZBOESFEVDFTVMUJNBUFQSFTTVSF UFNQFSBUVSFT
t.VMUJQMFTUBHJOHQSPWJEFTJOMFUQSFTTVSFTCFMPXNN)HBCTPMVUFXIJMF t.PTUEJGmDVMUUPSFQBJSPSSFCVJME
EJTDIBSHJOHUPBUNPTQIFSF t.BZSFRVJSFBHBTQVSHFGPSDPPMJOH
t/PDPOUBNJOBUJPOPGFWBDVBUFEHBT PSUPQSPUFDUUIFCFBSJOHTBOETFBMT
t%VFUPMBDLPGDPOEFOTBUJPO
QVNQDBOCFGBCSJDBUFEPGTUBOEBSE
GSPNUIFQSPDFTTHBT
JOFYQFOTJWFDBTUJSPO t%VFUPIJHIPQFSBUJOHUFNQFSBUVSFT
TPNFQSPDFTTHBTFTNBZQPMZNFSJ[F
Humidity
Control
Department Editor: Kate Torzewski
APPLICATIONS
Compressor
Heater
Dehumidification by cooling or dessication Condenser
Receiver
has a variety of applications, including:
Desiccant
Preventing moisture regain. Nearly all Fan Cooling wheel
coil
materials have some affinity for moisture
based on surface characteristics and Process air Dry air
the amount of surface exposed to humid Heater
air. Moisture regain occurs when moist
Condensate pan
particles stick together. Drive motor
160 160
Preventing condensation. Air holds water 100% 100%
vapor in proportion to its temperature. Cold
Specific humidity, grains/lb
M
ost common acids can be stored in Standard flanged DImension all nozzle
horizontal or vertical ASME-type and dished heads locations from this line
tanks, as shown in the figures to the Manhole Vent
right, or vertical API-type tanks. Horizontal,
carbon-steel ASME-type tanks of 10,000—
40,000 gal capacity should have a plate
thickness of 3/8 in. with dished heads of
2A* 2A*
the same thickness. The thickness includes
a corrosion allowance of 1/4 in., which Knuckle radius line
provides a tank life of 15—20 years.
Weld lines
4VMGVSJDBDJE )40
prox)
ers for concentrated acid storage tanks at
Polyvinyl chloride pipe is recommended OD
temperatures up to 60˚C with a minimum
D (ap
for ordinary sulfuric acid, but for oleum,
lining thickness of 3/16 in. Semi-hard Submerged
Type 316 stainless steel or carbon-steel
R=O
rubber is used for lining equipment and Name fill pipe
lined with a fluorocarbon is best. plate
piping for acid up to 70˚C with FRP tanks
of vinyl-ester resin. Grounding
15 in.
1IPTQIPSJDBDJE )10
clip
Storage. The tank bottom should be rolled
to a height of 3 in. (upward). This allows /JUSJDBDJE )/0
S
eparation by a membrane is achieved by Micrometers, 0.001 0.01 0.1 1.0 10
log scale
creating a boundary between different
bulk gas or liquid mixtures. As different
Scanning electron microscope Optical microscope
solvents and solutes flow through a membrane
at different rates, separation is achieved. Particle size
Here, we will focus on three filtration range Molecular Macromolecular Microparticle
techniques: microfiltration (MF), ultrafiltration Albumin protein Yeast cells
(UF) and nanofiltration (NF). These processes
are characterized by the size of the particle Aqueous salts Carbon black Paint pigment
that can be separated by the membrane, as Bacteria
Typical
illustrated in the figure. Each membrane type particles Sugar Virus A.C. fine test dust
is best suited for unique applications and is
Synthetic Tobacco smoke Milled flour
designed with the module and material that dye
will allow the best separation. Latex emulsion
Flow through a membrane is characterized Pesticide Colloidal silica Blue indigo dye Red
as either tangential flow filtration (TFF), where blood
Asbestos fibers cells
the feed stream flows at a velocity vector nor-
mal to the membrane surface, or normal flow Gelatin Coal dust
filtration (NFF), where the stream flows tangent
to the membrane surface. The flow pattern is Process for Ultrafiltration
filtration
dependent on the type of module utilized. NFF Nanofiltration Microfiltration
modules include: cartridges, stacked disks
and flat sheets. TFF modules include: plate-
and-frame (cassettes), hollow fibers, tubes, of crystallite orientation, fracture in such a way NANOFILTRATION
monoliths, spirals and vortex flow. that reproducible microchannels are made.
T
he failure of a device or of a group of range to maximum pressure. As a general "QSFTTVSFSFMJFGWBMWF 137
components can lead to overpressuriza- rule, when sizing a PRV, maximum heat-duty JTBOBVUPNBUJDQSFTTVSF
tion and subsequent adverse events, such assumed for the abnormal case should be no SFMJFWJOHEFWJDFUIBUJTBDUV
BUFECZQSFTTVSFBUUIFJOMFU
as fire, explosion, spill or release. The most more than 125% of normal heat duty.
PGUIFWBMWF5IPVHITBGFUZ
common causes of overpressurization are WBMWF
PSTBGFUZSFMJFG
listed below. Understanding the circumstanc- Abnormal vapor input WBMWF
JTUIFUFSNJOPMPHZ
es surrounding overpressurization will help Abnormal vapor input can be caused by the GPSWBMWFTSFMJFWJOHHBT
an engineer to avoid these failures. failure of the upstream control valve to fully PSWBQPS
XFXJMMVTFi137u
open, or upstream-relieving or inadvertent UPEFTDSJCFBMMUZQFTPG
External fire valve opening. The required relieving QSFTTVSFSFMJFGWBMWFT"
According to API RP 520 and 521 standards, capacity must be equal to or greater than the SFMJFGWBMWF
VTFEGPS
MJRVJETFSWJDF
HFOFSBMMZ
a fire-exposed area is within an area of amount of the vapor accumulation expected
PQFOTJOQSPQPSUJPOUP
2,500 and 5,000 ft2, and below a height of under the relieving conditions. BOZJODSFBTFJOQSFTTVSF
25-ft above the grade. In this scenario, the PWFSPQFOJOHQSFTTVSF 4QSJOH
exposed vessel is blocked in. Potential vapors Loss of absorbent flow "TBGFUZWBMWFIBTDIBS
resulting from the fire must be relieved using When gas removal by absorbent is more BDUFSJTUJDTTJNJMBSUPB
a PRV on the vessel, or via a vent path that than 25% of the total inlet-vapor flow, an SFMJFGWBMWFFYDFQUUIBU
remains in a locked-open position between interruption of absorbent flow could cause JUVTVBMMZPQFOTSBQJEMZ
the vessel and an adjoining vessel. pressure to rise in the absorber. The PRV QPQT
BOEJTQSJNBSJMZ
VTFEGPSHBTPS
should be sized base on the net accumulation
WBQPSTFSWJDF
Blocked outlets of the vapor at the relieving conditions.
The closure of a block valve on the outlet of a
pressure vessel can cause the vessel's internal Entrance of volatile materials
pressure to exceed its maximum allowable The entrance of a volatile liquid, such as
working pressure if the source pressure water or light hydrocarbons into hot oil dur-
exceeds the vessel design pressure. Blocked ing a process upset, can cause instantaneous
outlets can be caused by control valve failure, phase expansion. Instead of relying on PRVs,
inadvertent valve operation, instrument-air processes should be properly designed with
or power failure, and other factors. A PRV the use of double block valves, the avoid- Source:
must have sufficient capacity to pass a fluid ance of water-collecting pockets and use of Farris
flowrate that is equivalent to the difference steam condensate traps and bleeds on water Engineering,
Brecksville,
between those of the incoming fluids and the connections. Ohio
outgoing fluids.
%JTD
Accumulation of non-condensibles
Utility failures Accumulation can result from blocking of the /P[[MF
These failures can include the following: gen- normal non-condensible vent or accumulation the downstream vessel should should be
eral power failure, partial power failure, loss in the pocket of a piping configuration or designed to handle the pressure and volume
of instrument air, loss of cooling water, loss of equipment. Because this can result in a loss of of the incoming stream without overpressur-
steam, and loss of fuel gas or fuel oil. For these cooling duty, PRV analysis should be handled izing. If the upstream vessel does not have
cases, a flare header should be designed and the same way. adequate relief capacity, the downstream
sized based on the maximum relief load that vessel should have a PRV of its own.
could result from a potential utility failure. Valve malfunction When two vessels are connected by an
Check-valve malfunction results in backflow, open path and the first has its own PRV and
Loss of cooling duty which can be from 5 to 25% of the normal discharges to a flare header, the second
Cooling-duty losses can include the following: flowrate. Required relief capacity should be will experience the impact from the relieving
loss of quench stream, air-cooled exchanger based on this. pressure of the first vessel and should be
failure, loss of cold feed and loss of reflux. Inadvertent valve operation results in a analyzed accordingly.
Relieving capacity should be calculated by valve position that is opposite from normal
performing a heat balance on the system, operating conditions, which is largely caused Upstream relieving
based on the loss of the condensing duty. by human error and can be avoided by care- Required relief capacity should be greater
ful operation. than the vapor generated because of heat
Thermal expansion Control valve failure to open or close is buildup in the system.
When liquid is blocked in a vessel or caused by electronic- or mechanical-signal
pipeline, external heat input can cause liquid failure. This typically will affect just one valve Runaway chemical reaction
temperature, and hence volume, to rise. This at a time and should be analyzed on a case- Runaway reactions tend to accelerate with
can be caused by the following: liquid that by-case basis. rising temperature; extremely high volumes
is blocked in a pipeline and is being heated, of non-condensibles with high energy can
the cold side of a heat exchanger being filled Process control failure cause the internal pressure of a vessel or
while the hot side is flowing, or a filled vessel This situation refers to the failure of process pipeline to rise rapidly. PRVs may not provide
at ambient temperature that is being heated controllers, such as programmable logic sufficient relief, so vapor-depressurizing
by direct solar radiation. PRVs used in these controllers and distributed control systems. systems, rupture disks and emergency vents
cases can be easily analyzed and sized. The potential impact of failure of every con- are preferable.
trol loop should be analyzed, as well as the
Abnormal heat input impact if one loop fails but all others remain References
This failure can be caused by: the supply of active. As a general rule, the required relief 1.Wong, W., Protect Plants Against Overpres-
heating medium, such as fuel oil or fuel gas to capacity must be greater than the vapor gen- sure, Chem. Eng. June 2001, pp. 66–73.
a fired heater, being increased; heat transfer erated because of heat buildup in the system. 2.Goodner, H., A New Way of Quantifying
occuring in a new and clean heat exchanger Risks: Part 2, Chem. Eng. November 1993, pp.
after revamp; control valve for the fuel supply Exchange tube rupture 140–146.
failing to fully open; or supply pressure of the When an upstream vessel is relieving by 3.Emerson, G., Selecting Pressure Relief Valves,
heating steam being changed from normal discharge fluid to a downstream vessel, Chem. Eng. March 18, 1985, pp. 195–200.
Sedimentation
Centrifuging
Department Editor: Kate Torzewski
C
entrifugation is the method of choice picked up by a skimmer. Periodically, the Stokes’ Law tells us that settling velocity
in the chemical process industries (CPI) solid must be removed from the centrifuge can be maximized with a high centrifugal
for separating solids from liquids. It manually or with a cutter knife. Alterna- speed, large particle size, large density
relies on the G-forces generated by high- tively, it can be removed continuously with difference between solids and liquid, large
speed rotation to recover solids or liquids a screw conveyor. Sedimentation centrifu- separation radius and low liquid viscosity.
from slurries, as well as clarify liquids or gation allows material to be separated
classify solids. hundreds or thousands of times faster than APPLICATIONS
Centrifuges can be categorized as either simple sedimentation by gravity alone.
sedimentation or filtration units. Sedimenta- Centrifuge selection is heavily dependent
tion centrifugation relies on a difference in STOKES' LAW on characteristics of the incoming slurry, in-
density between the solid and liquid being cluding particle size, solids concentration,
separated. Filtering is performed with a According to Stokes’ Law, the terminal liquid viscosity and density differential. Oth-
rotating basket fitted with a filter medium, velocity of a particle is determined by the er factors that come into play are the need
where the centrifugal force of rotation centrifugal gravity (Ω2r) created by the to remove solids periodically or continuous-
expels the liquid through the filter. centrifuge with particle-balancing buoy- ly and the degree of purity required of the
ancy and viscous drag taken into account. separated products. Table 1 summarizes
BASICS OF SEDIMENTATION This terminal (settling) velocity is deter- the mechanics and suitable applications of
mined by the equation below: common sedimentation centrifuges.
CENTRIFUGING 1
Vs 7 2r ( R s
R L ) d 2 3FGFSFODFT
18 M
The mechanics of sedimentation centrifug- 1.Scroder, T. Selecting The Right Centrifuge, Chem.
ing make it ideal for two-phase systems where Eng. September 1998, pp. 82–88.
with a high-density differential. As an Vs = Settling velocity, m/s 2.Moir, D. N. Sedimentation Centrifuges: Know What
incoming slurry spins in a sedimentation μ = viscosity, kg/m·s You Need, Chem. Eng. March 1988, pp. 42–51.
centrifuge, it forms an annulus adjacent to Ω = angular speed of rotation, rev/min 3.Bershad, B. C., Chaffiotte, R. M., Leung, W. F.
the bowl wall. The centrifugal force causes ρs = density of solid, kg/m3 Making Centrifugation Work For You, Chem.
the denser material to move outwardly to- ρL = density of liquid, kg/m3 Eng. August 1990, pp. 84–89.
ward the wall of the centrifuge bowl, while d = particle diameter, m 4.“Perry’s Chemical Engineers’ Handbook,” 8th ed.
the liquid overflows from the bowl or is r = centrifuge radius of curvature, m New York: McGraw Hill, 2008.
Multi- • Constructed of a series of tubular • Larger solid particles settle in the small • Clarifying fruit juices, wort
chamber sections arranged concentrically tubes, and particles of smaller sizes settle and beer
• The slurry feed enters in the in subsequent tubes
smallest tube and continues • Up to six chambers are typical with a
through the outer tubes as they maximum holding capacity of 0.064 m3
increase in size
Skimmer • Feed enters the hub end and • When a thick solid layer begins to form • Heavy-duty applications,
pipe / is accelerated to speed before on the bowl wall, supernatent liquid is such as coal dewatering
knife entering the separation pool removed with a skimmer, and solids are
discharge • Solids settle on the bowl wall while knifed out with centrifugal filters
liquid overflows the ring weir
Disc • Feed enters through the top axis • Solids settle under the disc and move • Self-cleaning types:
of the bowl and is accelerated by downward to be released at the bottom of purification of beverages,
a radial-vane assembly the bowl wall mineral oils, and edible oils
• The unit is constructed of a stack • Liquids travel up the conical channel, • Disc nozzle: corn wet
of typically 50 to 150 closely and their upward movement in the milling (starch separation,
spaced conical discs arranged at centrifuge is facilitated by holes across gluten thickening), clay
an angle between 40 and 50 deg each disc classification, acid crystal
washing, lube oil dewaxing
Decanter • Constructed of a solid external • Solids are removed from the conical • Applications that require
bowl and an internal screw discharge end (the beach) continuous removal of solids,
conveyor mounted horizontally • Bowl and conveyor rotate in the same where feed solids are high
direction, but at different speeds, creating and volume reduction is
a speed differential that controls the important
speed of solid removal
Screen- • A solid bowl decanter with a cylindrical screen added to the conical end • Improved cake dryness and
bowl highest product purity
Avoiding
Seal Failure
Department Editor: Kate Torzewski
S
eals are assemblies of elements that prevent SEAL FAILURE
the passage of a solid, liquid, gas or vapor
from one system to another. When a seal Seal failure is caused by a wide variety of circumstances, including improper in-
allows leakage of material, failure has occurred. stallation and environmental factors such as temperature, pressure, fluid incompat-
This guide provides an overview of common seal ibilities, time and human factors.
types and a discussion of seal failure to aid in Most causes of failure can be described as mechanical difficulties or system
choosing the most effective seal and avoiding operations problems. Examples of mechanical difficulties include strain on the
future failure. seal face caused by improper installation and vibration caused by improper net
positive suction head. Meanwhile, system operating problems can include condi-
SEAL TYPES tions that are outside of a pump’s best performance envelope, such as upsets, dry
running, and pressure or temperature fluctuations. Changes in the fluid being pro-
Seals types can be classified within two broad
cessed can cause problems as well, especially with fluids that flash or carbonize.
categories: static and dynamic. Static seals have
Common visual indicators of failure include short cuts, V-shaped notches in the
no relative motion between mating surfaces, while
seal, skinned surface in localized areas, or thin, peeled-away area on the seal.
dynamic seals do have relative motion between
Table 2 describes causes of some of the most prevalent types of seal failure with
a moving surface and a stationary surface. Seals
recommended methods of action.
do not have to fit into one category or the other;
In some cases, the cause of failure may be difficult to determine due to the com-
rather, seal types can fall anywhere on a spectrum
plexity of the seal construction. These unique failure modes can result in flaking or
between static and dynamic, and few seals are
peeling of the seal face, corrosion, flaking or pitting of the carbon faces, degrada-
strictly one type or the other. Table 1 describes the
tion of the elastomer energizer seals, and spring or bellows breakage. It is likely
applications and requirements of several common
that these rapid degradations are a result of contamination, which can be avoided
seal types.
with careful installation or using pre-assembled, cartridge-type mechanical seals.
I
n choosing a pump, it is important to match a pump’s capabili- required capacity that your pump must handle. Based upon a
ties with system requirements and the characteristics of the liquid required capacity in gal/min. and a pressure in lbf/in.2, the pump
being processed. These factors include the inlet conditions, coverage chart below can help engineers focus their selection to a
required flowrate, differential pressure and liquid characteristics. just a few pump styles.
Generally, the quality of the liquid should remain unchanged after
passage through a pump. Therefore, material compatibility, viscos-
ity, shear sensitivity and the presence of particulate matter in a
liquid are important considerations in pump selection.
Most engineering applications employ either centrifugal or
positive displacement (PD) pumps for fluid handling. These pumps
function in very different ways, so pump selection should be based
1SFTTVSF
MCGJO
)FBE
GUPGMJRVJE
on the unique conditions of a process.
$FOUSJGVHBMQVNQT
The most widely used pump in the chemical process industries for
liquid transfer is the centrifugal pump. Available in a wide range of
sizes and capacities, these pumps are suitable for a wide range of
applications. Advantages of the centrifugal style include: simplic-
ity, low initial cost, uniform flow, small footprint, low maintenance
expense and quiet operation.
$BQBDJUZ
HBMNJO
1PTJUJWFEJTQMBDFNFOUQVNQT
Though engineers may be first inclined to install centrifugal pumps, .FUFSJOHmQMVOHFS 4DSFX $FOUSJGVHBMNVMUJTUBHF
.FUFSJOHmEJBQISBHN (FBS %JSFDUBDUJOHTUFBN
many applications dictate the need for PD pumps. Because of their $FOUSJGVHBMEPVCMFTVDUJPO "YJBMGMPX .VMUJDZMJOEFS
mechanical design and ability to create flow from a pressure input, $FOUSJGVHBMTJOHMFTUBHF
TJOHMFTVDUJPO 3FHFOFSBUJWF
PD pumps provide a high efficiency under most conditions, thus Adapted from Perry’s Chemical Engineers’ Handbook
reducing energy use and operation costs.
$IPPTJOHDFOUSJGVHBMWFSTVTQPTJUJWFEJTQMBDFNFOU
PUMP SPECIFICATIONS
These two main pump styles respond very differently to various
Based on the application in which a pump will be used, the pump
operating conditions, so it is essential to evaluate the requirements
type, and service and operating conditions, the specifications of a
of a process prior to choosing an appropriate pump. Table 1 il-
pump can be determined.
lustrates the mechanical differences between these pumps, as well
as the effects of pressure, viscosity and inlet conditions on flowrate • Casting connection: Volute casing efficiently converts velocity en-
and pump efficiency. ergy impacted to the liquid from the impeller into pressure energy.
A casing with guide vanes reduce loses and improve efficiency
3BOHFPGPQFSBUJPO over a wide range of capacities, and are best for multistage high-
Pump styles range far beyond simply PD and centrifugal pumps. head pumps
PD pumps encompass many specific styles, including a variety of • Impeller details: Closed-type impellers are most efficient. Open-
reciprocating, rotary and blow-cover pumps. Likewise, centrifugal type impellers are best for viscous liquids, liquids containing solid
pumps encompass radial, mixed, and axial flow styles, which all matter, and general purposes
• Sealings: Rotating shafts must have proper sealing methods to
PUMP COMPARISON CHART prevent leakage without affecting process efficiency negatively.
Centrifugal Positive displacement Seals can be grouped into the categories of noncontacting seals
Pump pump
and mechanical face seals. Noncontacting seals are often used
The pump captures
The pump imparts a
confined amounts of
for gas service in high-speed rotating equipment. Mechanical face
velocity to the liquid, seals provide excellent sealing for high leakage protection
liquid and transfers
resulting in a pressure
Mechanics them from the suction • Bearings: Factors to take into consideration while choosing a
at the outlet.
to discharge port.
Pressure is created bearing type include shaft-speed range, maximum tolerable shaft
Flow is created and
and flow results misalignment, critical-speed analysis, loading of compressor
pressure results
Flow varies with Flow is constant with impellers, and more. Bearing styles include: cylindrical bore; cy-
Performance
changing pressure changing pressure lindrical bore with dammed groove; lemon bore; three lobe; offset
Viscosity
Efficiency decreases Efficiency increases with halves; tilting pad; plain washer; and taper land
with increasing viscosity increasing viscosity
• Materials: Pump material is often stainless steel. Material should
Efficiency peaks at the
be chosen to reduce costs and maintain personnel safety while
best-of-efficiency point.
Efficiency increases avoiding materials that will react with the process liquid to create
Efficiency At higher or
with increasing pressure
lower pressures, corrosion, erosion or liquid contamination
efficiency decreases
Liquid must be in the 3FGFSFODFT
Negative pressure is
pump to create a pres- 1. “Perry’s Chemical Engineers’ Handbook,” 7th ed. New York: McGraw
Inlet created at the inlet port.
sure differential. A dry Hill, 1997.
conditions A dry pump will prime
pump will not prime on 2. Petersen, J. and Jacoby, Rodger. Selecting a Positive Displacement Pump,
on its own
its own Chem. Eng. August 2007, pp. 42–46.
Pristine
Processing
Department Editor: Kate Torzewski Equipment
P
rocesses in the pharmaceutical, biotechnology, food and semiconductor industries must 7BMWFT
meet a high set of standards to ensure high product purity. Equipment criteria specific to Valves should not harbor contaminants and
high-purity processes are established to minimize contamination and maintain prod- must be easy to clean. By these criteria,
uct integrity. In designing a pristine process, material and equipment style are of upmost diaphragm and pinch valves are excellent
importance. Bacteria is the main cause of contamination and is prone to growing in the dead choices for ultrapure processes, as they
cavities of equipment created by sharp corners, crevices, seams and rough surfaces. Another have smooth, gently curved surfaces that
source of contamination is leaking, which allows undesirable chemicals to compromise the will not harbor contaminants. Ball check,
quality of the process ingredients, by causing contamination, rusting and particle generation. full-port plug and full-port ball valves are
good choices as well, while butterfly, spring
MATERIALS OF CONSTRUCTION check, gate and swing check valves are all
unacceptable, since contamination can col-
Many factors must be taken into consideration when selecting materials of construction for lect in the corners that are essential to their
use in pristine process applications where high-purity and sanitation are paramount. All sur- design. Though several valves are appropri-
faces should be constructed of a smooth material that will not corrode, generate particles ate for pristine processes, certain valves
or harbor dead cavities. These criteria can be met with three standard materials: 316L are better suited for particular applications.
stainless steel (SS), polyvinylidene fluoride (PVDF) and polytetrafluoroethylene (PTFE). The Diaphragm valves are the most widely used
advantages and disadvantages of these materials are summarized below to facilitate the in high-purity systems for their resistance
material selection process for a given application with consideration of chemical compat- to contamination and ability to be used as
ibility, cost, and temperature stability. a control valve. Ball and plug valves, on
the other hand, are less costly and are not
MATERIALS OF CONSTRUCTION limited by temperature and pressure. Also,
Material Advantages Disadvantages in applications using sterile steam and
• Mechanical strength Vulnerability to corrosion by
freeze-drying, ball valves are preferred over
Stainless diaphragm valves because they eliminate
• Functions at 121°C certain chemicals, which
Steel the risk of catastrophic seat failure.
(steam-sterilization temperature) increases with temperature
• Chemically inert
• Resistant to corrosion and leaching Functions only intermittently 4FBMT
PVDF As with all pristine processing equipment,
• Durable and long-lasting at 121°C
• Retains circumferential strength high-purity seals should not have any
• The most chemically inert plastic cavities where contaminants can breed. By
• Resistant to corrosion and leaching Complex shapes are choosing a seal with gland rings that do not
PTFE
• Avoids leaching difficult to form need to be threaded or ported, the areas
• Suitable for coating equipment where bacteria can breed are minimized. In
choosing a seal material, it is important to
EQUIPMENT STYLE SELECTION find a compound that will not swell, crack,
pit or flake, thus reducing seal failure and
Critical factors in high-purity equipment selection include cleanability, cost, flow capabili- contamination. To ensure the success of
ties and product compatibility. With these considerations in mind, criteria useful for choos- seals, fluroelastomers are a top choice in
ing pumps, valves, seals and piping are described in this section. pristine processing applications for their ex-
cellent thermal stability, chemical resistance
1VNQT and mechanical durability.
A fundamental requirement of pristine processing pumps is the ability to clean a pump in
place without disassembly. Pump seals, gaskets and internal surfaces should eliminate the 1JQJOH
buildup of material and should clean out easily during wash cycles. The most common pump The surface of piping, as well as any wet-
styles for high-purity processes are centrifugal, lobe-style and peristaltic pumps, which are ted equipment parts, should have a very
outlined below. smooth surface. When 316 SS is being
PUMPS used, electropolishing is a good method
Pump Applications best for achieving an ultra-smooth finish. Joining
Advantages Disadvantages methods should minimize crevices and dead
style suited for this style
• Handling low- cavities, and all materials should be free of
Efficiency and flow de-
• Low cost viscosity products biological degradable substances, leachable
Centrifugal crease with increasing
• Easy cleanability • Handling high flowrates substances, and glues and solvents that may
pressure and volume
(40–1,500 gal/min) migrate into the product stream.
• Low cost The need for hoses may
• Easy cleanability cause issues in elastomeric • Small, batch-oriented 3FGFSFODFT
• No mechanical compatibility, temperature applications
Peristaltics 1. Smith, B., What Makes a Pump for High-Purity
seals and pressure limitations, • Laboratory or pilot-
Fluids?, Chem. Eng., pp. 87–89, April 2002.
• Non-damaging to and a need to change scale plants
delicate products hose regularly 2. Schmidt, M., Selecting Clean Valves, Chem.
Eng., pp. 107–111, June 2001.
• Higher pressure
• Large, continuous 3. Wulf, B., Pristine Processing: Designing Sanitary
and flow Systems, Chem. Eng., pp. 76–79, Nov. 1996.
Rotary duty applications
capabilities High cost
Lobes • Steaming and high 4. Weeks, D. T. and Bennett, T., Specifying Equip-
• Unaffected by
pressure applications ment for High-Purity Process Flow, Chem. Eng.,
pressure variations pp. 27–30, Aug. 2006.
Heat Transfer
Fluids and Systems
Department Editor: Rebekkah Marshall
toughness down to temperatures as low as gravity pipe (dia. from 8 to 144 in.) and Zirconium2 20.77
–250°C. For good IS at temperatures down pressure pipe (dia. from 1 to 16 in.). It is Graphite 121.15
to –45°C, C-Mn-Si steels are recommended. not unusual to see FRP pressure pipe han- Hexoloy 125.65
The most preferred grades are fine-grained dling pressures as high as 2,000–5,000 Glass1 1.00
steels of pressure-vessel quality, such as ASTM psi during chemical processing, with the Lead 35.30
A 516 and ASTM A 537 (in all grades). For higher-pressure pipe at the lower end of Inconel2 12.00
temperatures between –45 and –100°C (for the diameter scale. CPVC 0.14
PTFE (Polytetra- 0.25
example, for liquid-ethylene storage), steels fluoroethylene)1
containing 2.5–9% Ni are useful. Between HEAT TRANSFER PFA (Perfluoro- 0.19
–150 and –250°C, the Ni-Cr austenitic steels alkoxy resin)1
(300 series, of 18/8 varieties), are highly
PROPERTIES [4]
ETFE (Ethylene 0.24
recommended. In the nonferrous category, tetrafluoroethylene)1
Al has excellent properties for temperatures Metals, including specialty materials, are PVDF (Polyvinylidene 0.23
as low as –250°C. Also attractive are Cu and the best choice in terms of good heat trans- fluoride)1
some of its alloys, which can withstand tem- fer. In the lined category, glass is used ex- ECTFE (Ethylene chlo- 0.16
peratures down to –195°C. tensively for process equipment where good rotrifluoroethylene)1
heat transfer is required. Lined materials, 1. Common choice for lining material
however, often have the problem of uneven 2. Exotic metals
CHEMICAL RESISTANCE thermal expansion, which may weaken the
bonding of the lining in due course. While
References
1. Nalli, K., Materials of Construction For Low-Tem-
CPVC [2]. Many nonmetals do not have the fluoropolymers have excellent compatibility perature and Cryogenic Processes, Chem. Eng.
tensile strength to meet the pressure require- with various chemicals and special surface July 2006, pp. 44–47.
ments of various process applications, espe- and physical chemistries, they are gener- 2. Newby, R. and Knight, M., Specifying CPVC In
Chemical Process Environments, Chem. Eng., Oc-
cially at elevated temperatures. But years of ally not used for reaction vessels because tober 2006, pp. 34–38.
testing and actual field performance prove of their poor heat-transfer properties. Ther- 3. Beckwith, S., and Greenwood, M., Don’t Over-
that chlorinated polyvinyl chloride (CPVC) mal conductivities for various materials are look Composite FRP Pipe, Chem. Eng., May
2006, pp. 42-48.
systems can be pressure rated for operation listed in the Table, and typical applications 4. Robert, J., Selecting Materials of Construction,
as high as 200°F. CPVC’s high heat-distor- are shown in Figure 2. Chem. Eng., September 2005, pp. 60–62.
tion temperature and resistance to corrosion
make it suitable for applications such as 300 Exotic Exotic Exotic Exotic
metal processing, pulp and paper, and in-
dustrial wastewater treatment, where harsh 250 Exotic Exotic Exotic
Temperature, °C
Aliphatics
Application: Storage Transport Agitation (Agitation +
Excellent Strong bases heat transfer)
Typical Tanks, Pipelines, Mixers
Strong oxidants
Reactors
equipment: vessels valves,
Good Halogens
flowmeters
Aromatic solvents
Fair Esters and ketones FIGURE 2.8IFOMPPLJOHCFZPOETUFFMGPSNBUFSJBMTPGDPOTUSVDUJPO
JUJTJNQPS
Poor UBOUUPDPOTJEFSUIFJOUFOEFEBQQMJDBUJPOBOEUFNQFSBUVSFSBOHF&YPUJD TQFDJBMUZ
NFUBMT TFF5BCMF
BSFTIPXOIFSFUPTFSWFXFMMJOBMMBQQMJDBUJPOT"OPUIFSNBUF
FIGURE 1.$17$PGGFSTSFTJTUBODFUPB SJBM
FRVBMMZTVJUFEUPBTQFDJGJDSFRVJSFNFOU
IPXFWFS
NBZCFDIPTFOBTUIFNPSF
WBSJFUZPGIBSTIDIFNJDBMT DPTUFGGFDUJWFPQUJPO
Fuel Selection
Considerations
Department Editor: Rita L. D'Aquino
T
he selection and application of fuels to various 6OJUIFBUPVUQVU
#UVI
combustors are complex. Most existing units have "DIJFWBCMFDPNCVTUJCMFTJOBTI
limited flexibility in their ability to fire alternative A
fuels. New units must be carefully planned to assure
the lowest first costs without jeopardizing the future
capability to switch to a different fuel.
NATURAL GAS
Natural gas has traditionally been the most attrac-
tive fuel type for combustors because of the limited
need for fuel-handling equipment (e.g., pipelines, B
metering, a liquid-knockout drum, and appropriate
controls) and the freedom from pollution-control
$PNCVTUJCMFTJOBTI
"WFSBHFGVFMDPTU
#UV
equipment. Drawbacks include rising fuel costs, in-
A STEPWISE PROCEDURE TABLE 1. Some well-known databases and solvent selection tools
Databases Address and comments
Organic solvents have been used in many industries for centu-
ChemFinder Searchable data and hyperlink index: http://chemfinder.cambridgesoft.com
ries, but the methods and tools to select optimal solvents while
minimizing their adverse environmental, health, safety and op- Solvents Databases Solvent substitution data systems at http://es.epa.gov/ssds/ssds.html;
erational concerns are still evolving. The appropriate selection “Handbook of Solvents” from www.chemtec.org/cd/ct_23.html; and
of solvents depends to a large extent on the application — more SOLVDB at http://solvdb.ncms.org/index.html
specifically on what needs to be dissolved, and under what NIST Webbook Source of physical and chemical data at http://webbook.nist.gov
conditions. This article presents a four-step approach to solvent DIPPR and www.aiche.org/TechnicalSocieties/DIPPR/About/Mission.aspx; and
selection based upon Ref. 1*, where the reader will find a list of TAPP www.chempute.com/tapp.htm
additional resources on this topic. CAPEC Database Pure as well as mixture properties data, including solvent-solute database:
www.capec.kt.dtu.dk/Software/ICAS-and-its-Tools
Identify the challenge and solvent characteristics. Selection Tools Address and comments
The first two steps are: 1) identifying the actual problem and SMSwin A specialized software for property estimation and solvent classification:
technology or unit operation required to solve it; and 2) defining www.capec.kt.dtu.dk/documents/software/SMSWIN.htm
the requirements that must be met by the solvent, using criteria
NRTL-SAC and Activity coefficient method based on segment contributions. Predictive based
related to its physical and chemical properties (e.g., pure-solvent
eNRTL-SAC on a small set of solubility data. Useful for crystallization solvent selection
properties, such as normal boiling point, the Hildebrand solubility and extends to LLE and VLE: www.aspentech.com
parameter at 300 K, the Hansen solubility parameters; solvent-
solute properties, such as the solubility of the solute as a function
of the composition of the mixture; and functional constraints, such TABLE 2. Well-known solvents together with their related properties
as solute loss in solute).
Solvent Name Molecule type Group type Charge NBP NMP Sol. Par.
(K) (K)
Obtain reliable values of solvent properties and
narrow down selection. There are several alternatives for this 1-Methyl-2-pyrrolidinone Amide 1 NE/EPD 475.15 249.15 23.16
third step. For example, one can measure the required properties, Acetonitrile Nitrile 1 E/NPG 354.75 229.35 24.05
use a database of properties of chemicals (or solvents), or, use Dimethyl sulphoxide S-oxide 1 E/NPG 462.15 291.65 26.75
property models to estimate them. For solvent-selection problems Dimethyl formamide Amide 1 NE/NPG 426.15 212.75 23.95
not involving chemical reactions, the pattern of the desired Dimethylacetamide Amide 1 NE/NPG 438.15 253.15 22.35
solvent is established through analysis of the solute, application Diisopropyl ether Ether 2 NE/EPD 341.65 181.35 14.45
type, and other constraints. Dimethyl ether Ether 2 NE/EPD 248.35 131.65 15.12
Once this is established, a database of known solvents can Methyl tertbutyl ether Ether 2 NE/EPD 328.35 164.55 15.07
be used to identify the solvents that match the necessary pattern Tetrahydrofuran Ether 2 NE/EPD 338.15 164.85 18.97
(Table 1). On the other hand, when chemical reactions are Chlorobenzene Chloride 3 NE/P 632.35 404.9 19.35
involved, the approach is based on transition-state theory and m-xylene (also o-; p-) Aromatic HC 3 NE/P 412.27 225.3 18.05
requires consideration of the solvation energies of the reactants,
Toluene Aromatic HC 3 NE/P 383.95 178.25 18.32
products and transition states, and thus, knowledge of the reac-
Acetic acid Acid 4 PG 391.05 289.81 19.01
tion mechanism.
When the crucial values have been found, the solvent search Propionic acid Acid 4 E/PG 414.25 252.45 19.41
could be such that first, solvent-pure properties are used, followed Sulfuric acid Acid 4 E/PG 610 283.46 28.41
ordered
by solvent-EHS, then solvent-solute, and finally solvent-function.
Narrow down the list by removing the compounds that do not Propanol Alcohol 5 E/N 370.35 147.05 24.45
match desired properties. Ethanol Alcohol 5 E/N 351.35 159.05 26.13
A protocol derived by Britest Ltd. (www.britest.co.uk) seeks to Butanol Alcohol 5 E/N 390.81 183.85 23.35
use mechanistic principles to guide solvent selection (Figure). The Ethylene glycol Alcohol 5 E/N 470.45 260.15 33.7
objective is to follow the arrows according to the problem defini- Dichloromethane Chloride 6 NE/EPD 313.15 178.05 20.37
tion and a search criterion until an end-point is reached, thereby Heptane Alkane 7 NE/I 371.65 182.55 15.2
obtaining the characteristics of the candidate solvents. These Hexane Alkane 7 NE/I 341.85 177.85 14.9
characteristics are used to identify the group to which the solvents Pentane Alkane 7 NE/I 309.22 143.42 14.4
belong using solvents database (see Table 2). The corresponding Methanol Alcohol 4, 5 E/N 337.85 175.47 29.59
group-types are evaluated and a final selection is made. Water Aqueous 4, 5 E/N 373.15 273.15 47.81
Verify selection. The fourth step is to verify that the solvent NE = non-electrolytic solvent; E = electrolytic solvent; P = polarizable; EPD = electron-pair
donor; I = inert; PG = protogenic (proton donor); N= neutral (donor & acceptor); NPG =
works as expected by performing a computational validation non-protogenic (proton acceptor); NBP = normal boiling point; NMP = normal melting point;
by simulation. Experimental validation of a solvent candidate is Sol. Par. = Hildebrand solubility parameter at 300 K (MPa1/2)
required at all stages of process development.
*Reference: 1. Gani, R., et al., A Modern Approach to Solvent Selection, Chem. Eng., Vol. 113, No. 3, pp. 30–43, Mar. 2006. Author E-mail: rag@kt.dtu.dk
Controlling
Crystal Growth
Department Editor: Rita L. D'Aquino
T
he formation of crystals requires the birth of 1) represents the start of this new set of batch
new particles, also called nucleation, and dynamics. It is safe to assume that significant NOMENCLATURE
the growth of these particles to the final nucleation is now occurring and that the rate A Crystal surface area, ft2
product size. The driving force for both rates is of energy transfer is too high.
the degree of supersaturation, or the numerical 5. By taking the slope of the curve represent- B° Nucleation rate, (number of nuclei)/
ft3/s
difference between the concentration of solute in ing the estimated number of nuclei present
the supersaturated solution in which nucleation at the measured point in time (Nti) vs. time G Crystal growth rate, μm/s
and growth occurs vs. concentration of solute in (ti), one can determine the nucleation rate. kv Crystal-volume shape factor,
a solution that is theoretically in equilibrium with Using your representation of Figure 3, create dimensionless
the crystals. a dashed, horizontal line across the lower L Crystal size, μm
In a batch crystallizer, the crystal size portion of the graph depicting the selected,
L’ Smallest-measurable size, μm
distribution (CSD) is controlled by first seeding cumulative number of crystals (Ni’), and their
the initially supersaturated batch with a known sizes (L1–L4) over time (t1–t4). Lav Size of crystal fraction, μm
number and size distribution of crystals, and 6. For a selected cumulative number of crystals Lf Final size of crystal, μm
then controlling the rate of evaporation or (Ni’), plot the crystal size (L) vs. time (t), as Ls Seed size, μm
cooling (i.e., rate of energy transfer) so as to demonstrated in Figure 3. The slopes repre-
achieve a level of supersaturation that supports sent the crystal growth rate (G). If the level of N Number of seeds
N i’ Constant, cumulative number
adequate crystal growth and an acceptable supersaturation changes during the run, the of crystals in crystallizer
rate of nucleation. The relationship between growth rate also changes. Non-parallel lines
NLi Number of crystals of a given size,
supersaturation and growth is linear, but that would indicate that the larger crystals are Lav
between nucleation and growth is raised to a growing at a faster rate, due to a reduced
power that is usually greater than one, making diffusional resistance [layer] at the crystal Nti Number of crystal nuclei at any time
it difficult to grow large crystals when nucleation surface. With larger particles, the resistance S Rate of supersaturation
is occurring. The following procedure describes layer may be smaller, allowing the solute to S* Maximum allowable supersatura-
how to achieve the optimal growth rate: more readily reach the crystal surface and tion, lb/ft3 solvent
1. Screen the seeds at the beginning of the ex- incorporate itself into the lattice. These fac- t, ti, tf Time, h
periment to determine the cumulative number tors collectively contribute to the accelerated
of crystals that are greater than a given size growth rate of the larger particles. Parallel ΔWi Weight of crystals on screen
N’. Estimate NLi, the number of crystals of a lines indicate that the growth rate is not Rc Crystal density, lb/ft3
given size (Lav) obtained from the screening: dependent on crystal size.
7. Increase the rate of cooling or evaporation
$Wi (1) until additional nucleation occurs, upon compete for supersaturation and growth.
N Li which you can safely assume that the growth • Studies show that milled seeds may not grow
Lav 3 kv R c rate is too high. as well as unmilled seeds. Furthermore, not
The parameters are defined in the table 8. Develop a seeding and evaporation profile all crystals of a given size grow at the same
of nomenclature. To convert from μm to ft, that will yield a growth rate that is lower than constant rate. This is sometimes attributed to
multiply by 3.28 x 10–6. the value found in Step 6. the differences in the surface characteristics
2. Continue to measure the number and size When determining the growth rate, keep in of particles that have equal dimensions.
of crystals as the cooling or evaporation mind the difference in mixing characteristics • Fines destruction in a batch system can
program is in progress. Prepare an inverse between a laboratory-scale vessel and a com- greatly reduce the effects of secondary nucle-
cumulative plot of the number of crystals mercial configuration. A small tank generally ation on the CSD, and significantly increase
greater than a given size vs. size of the offers a higher relative pumping capacity, crystal size while narrowing the CSD.
crystal (Figure 1). The crystal growth rate de- shorter blend time, and higher average shear • In practice, not all additional nucleation can
pends on the energy transfer rate, so modify rates within a narrower range. be suppressed. Crystallizations carried out at
the rate of energy transfer until a desirable low levels of supersaturation near the meta-
product is obtained. USEFUL OBSERVATIONS stable zone (i.e., the conditions under which
3. Repeat the first two steps at intervals crystals grow, but do not typically nucleate)
throughout the batch cycle and plot the • Most processors will agree that when it will display some secondary nucleation, due
results as shown in Figure 1. The family of comes to crystals, the larger, the better. Large to crystal-crystal interactions and contact
curves resulting from data plotted under the crystals are easier to handle in downstream between the crystals and the impeller. Nev-
selected conditions indicates that the number operations, such as washing, centrifugation ertheless, the mean crystal size, shape and
of crystals is not increasing with time. Thus, and drying. distribution are dramatically improved when
no additional nucleation is occurring yet. • As previously mentioned, it is desirable for seeding is followed by a programmed rate of
4. Proceed to collect crystal samples, an- the seeds’ size distribution to reflect a nar- energy transfer.
ticipating the onset of nucleation. Figure row cut of particles. In this cut, the weight
of crystals with sizes finer than Ls should be Reference: Genck, W., Better Growth in Batch Crystal-
2 indicates that the number of crystals is
lizers, Chem. Eng., Vol. 106, No. 8, pp. 90–95, Aug.
significantly increasing with time. In this minimal because these tiny particles add 2000. E-mail: genckintl@aol.com
figure, t1 (not to be confused with t1 in Figure enormously to the number of crystals that
/h OVNCFSPGDSZTUBMT
/U4 /h3
/U3
/h2
/U2
MBSHFSUIBO -
/h1
/h number of crystals
/U1
4J[F
-
U4
U2 U3
larger than -
U1 U2 U3 U4 U1
/h1 U1 U2 U3 U4
-h -1 -2 -3 -h E-
( = EU
Size, - (Mm) -h = smallest measurable size
-h = smallest measurable size Time, U (min)
FIGURE 1. FIGURE 2. FIGURE 3.
Hazardous Area
Classification
Department Editor: Rebekkah Marshall
Acknowledgement and references Classification of Flammable Liquids, Gases, or 4. IEC 60079-10-2002 Electrical apparatus for explo-
Vapors and of Hazardous (Classified) Locations for sive gas atmospheres — Part 10: Classification of
We would like to thank Vladimir Stetsovsky of Chilworth Electrical Installations in Chemical Process Areas. hazardous areas.
Technology, Inc. for his contributions to this page
3. NFPA 499-2004, Recommended Practice for the 5. IEC 61241-3-2005 Electrical apparatus for use in
1. National Electrical Code-2005-NFPA 70, National Classification of Combustible Dusts and of Hazard- the presence of combustible dust — Part 3: Classifi-
Fire Protection Association. ous (Classified) Locations for Electrical Installations cation of areas where combustible dusts are or may
2. NFPA 497-2004, Recommended Practice for the in Chemical Process Areas. be present.
Preventing
Runaway
Department Editor: Rebekkah Marshall Reactions
GENERAL CONSIDERATIONS [1] DATA COLLECTION THERMAL STABILITY
A process is considered to be thermally The following data are especially rel- CRITERIA [1, 4]
safe only if the reactions can easily be evant in avoiding runaway reactions: As a guideline, three levels are sufficient
controlled, and if the raw material, the • Physical and chemical properties, ig- to characterize the severity and prob-
products, the intermediates and the re- nition and burning behavior, electro- ability of a runaway reaction, as shown
action masses are thermally stable under static properties, explosion behavior in the Table.
the considered process conditions. Check and properties, and drying, milling,
into the process equipment, its design, its DEFINING HIGH, MEDIUM
and toxicological properties
sequence of operation and the control AND LOW RISK [1]
strategies. In addition to the engineer- • Interactions among the chemicals Severity Probability
ing aspects, get detailed information on • Interactions between the chemicals High %Tad > 200K TMRad < 8 h
thermodynamic and kinetic properties and the materials of construction Medium 50K < %Tad < 200K 8 h < TMRad <
of the substances involved, such as the • Thermal data for reactions and de- 24 h
reaction rates or heat-release rates as composition reactions Low %Tad < 50K and TMRad > 24 h
a function of process conditions. Deter- the boiling point
• Cooling-failure scenarios cannot be sur-
mine the physical and chemical proper-
passed
ties, as well.
Understanding of thermal-hazard po- DESIGN OPTIONS [2]
tential requires knowledge of various
Adiabatic temperature rise
skills and disciplines [3]. These include: If a reaction is has the potential for The adiabatic temperature rise is calculated
runaway, the following design changes by dividing the energy of reaction by
Operating mode: The mode of opera-
should be considered: the specific heat capacity as shown in
tion is an important factor. For instance,
Equation (1).
a batch reaction, where all the reactants • Batch to continuous. Batch reactors
are charged initially, is more difficult to require a larger inventory of reac- %Tad = 1,000Qr/Cp (1)
control than a semi-batch operation in tants than continuous reactors do, so where:
which one of the reactants is charged the potential for runaway in continu- %Tad = adiabatic temperature rise, K
progressively as the reaction proceeds ous systems is less by comparison
(for more, see Design Options). Qr = energy of reaction, kJ/kg
• Batch to semi-batch. In a semi-batch
Engineering: Design and layout of the reaction, one or more of the reactants Cp = heat capacity, J/(kg)(K)
plant and equipment and its built-in con- is added over a period of time. There-
trols impact the entire process. The ca- fore, in the event of a temperature or Time to maximum rate (TMR)
pacity of the heating or cooling system pressure excursion, the feed can be TMRad (the time to maximum rate, adiabatic) is
is important in this context. Process en- switched off, thereby minimizing the a semiquantitative indicator of the probability
gineering is used to understand the con- chemical energy stored up for a sub- of a runaway reaction. Equation (2), defining
trol of the chemical processes on a plant sequent exothermic release TMRad in hours, is derived for zero-order
scale. It determines which equipment reaction kinetics:
• Continuous, well-mixed reactors to
should be used and how the chemical TMRad = CpRTo2/3,600qoEa (2)
plug flow designs. Plug-flow reactors
processes should be performed. In ad-
require comparatively smaller volumes where:
dition, take into account technical failure
and therefore smaller (less dangerous) R = gas constant, 8.314 J/molK
of equipment, human errors (deviations
inventories for the same conversion
from operating instructions), unclear To = absolute initial temperature, K
operating instructions, interruption of • Reduction of reaction inventory via
qo = specific heat output at To, W/kg
energy supply, and external influences, increased temperature or pressure,
such as frost or rain (for more, see De- changing catalyst or better mix- Ea = activation energy, J/mol
sign Options). ing. A very small reactor operating The TMR value provides operating personnel
Chemistry: The nature of the process and at a high temperature and pressure with a measure of response time. Knowledge
the behavior of products must be known, may be inherently safer than one of the TMR allows decisions to be based on
not only under reaction conditions, but operating as less extreme conditions an understanding of the time-frame available
also in case of unexpected deviations because it contains a much lower in- for corrective measures in case heat transfer
(for example, side reactions, instability ventory [3]. Note that while extreme is lost during processing.
of intermediates). Chemistry is used to conditions often result in improved
gain information regarding the reaction reaction rates, they also present their References
pathways that the materials in question own safety challenges. Meanwhile, a 1. Venugopal, Bob, Avoiding Runaway Reac-
compromise solution employing mod- tions, Chem. Eng., June 2002, pp. 54–58.
follow. 2. Smith, Robin, ”Chemical Process Design,”
erate pressure and temperature and
Physical chemistry and reaction kinetics: McGraw-Hill, New York, 1995.
medium inventory may combine the
The thermophysical properties of the reac- 3. Kletz, T. A., “Cheaper, Safer Plants,”
worst features of the extremes [3]. IChemE Hazard Workshop, 2d., IChemE,
tion masses and the kinetics of the chemi- Rugby, U.K., 1984.
cal reaction are of primary importance. • Less-hazardous solvent
4. Gygax, R., Reaction Engineering Safety,
Physical chemistry is used to describe the • Externally heated or cooled to inter- Chem. Eng. Sci., 43, 8, pp. 1759–71, Au-
reaction pathways quantitatively. nally heated or cooled gust 1998.