Safety Promotion of Process Operations
in the Chemical Process Industries
Chun-Yu Chen
Donp-Liang Hwang
Department of Applied Chemistry
Chung-Cheng Institute of Technology
Tashi, Taiwan 33509, Republic of China
ABSTRACT
This paper discusses how to use the simulation
data obtained from the adiabatic reaction calorimeter
(ARC) to design the emergency relief systems for chemi-
cal reactors and storage vessels to overcome the problem
of reactor explosions resulted from excessive pressure
and thus the disastrous incidents can be avoided. Here
the styrene polymerization reaction is choosen as_ the
reaction system to describe the design of emergency
relief system using the ARC data. The simulation re-
sults show that the change of the initiator and the
monomer concentration will influence the maximun tempe-
rature, pressure, and the rate of temperature rise in
the reactor. The simulation results also illustrate
that in the designing of the emergency relief system,
the larger the overpressure is, the smaller the vent
diameter will be needed.
INTRODUCTION
Although there are many safety precautions within
chemical plants, the problem of reactor explosions due
to excessive pressure buildup, which may be caused by
reactor failures or operators’ errors, will cause major
releases of toxic or flammable chemicals and will result
in disastrous consequences. In many chemical processes,
however, materials need not be reactive or flammable in
order to blow up reactors. If one vaporizes liquid ina
closed reactor, one must balance the added volume with
outflow or with a like amount of condensation elsewhere
in the process. Otherwise, the pressure will rise until
the reactor failures, and the event will occur in an
instant ("explosion in the reactor"). Therefore, provi-
ding an adequate pressure relief is necessary for the
processes.
A number of investigators (Boyle, 1967; Huff, 1982;
Fauske, 1984, 1989; Leung, 1986; Leung and Fisher, 1989)
have considered the most realistic situations which are
based on the release of a vapor-liquid mixture, with
two-phase discharge in the relief system. This paper
discusses how to use the theoretical model to calculate
required relief area for batch polymerization reactors.
The method can be tailored to fit almost any batch
polymerization reaction or any other temperature-depen-
dent exothermic chemical reaction.
PROCESS DESCRIPTION
The process model for the free radical solution
polymeriza-tion of styrene in the ARC system (Fig. 1),
i.e. a batch reactor, involves reaction kinetics, a
material balance and an energy balance. The ARC model
(Jaisinghani and Ray, 1977) for polystyrene polymeri-
zation is
a(t]
= -kd[Ij
at
d(M]
= -kp[M](P] (1)
at
2fkd[I]
QI 12
kt
atr
PCpV ——— = -4HrVkp[M][P] - hA’ (Tr - Tw) (2)
at
with initial conditions
t=0, (I]=[I]in, [M]=[MJin, Tr=Trin
The physical constants and operating conditions for
polystyrene reaction are given in Table 1 (Brooks, 1981;
Biesenberger and Sebastian, 1983).
For the polystyrene polymerization, the pressure in
the ARC is dependent on the changes of the air partial
pressure and the styrene vapor pressure (Huff,1982).
- 108 -
The change of the pressure in the ARC is given in equa-
tion 3:
Pr = Pair + PO (3)
where
Pair = 0.00282Tr
log(PO) = 4.12108 - 1496.44/(Tr - 58.47)
At the adiabatic conditions, the temperature (Tra), the
pressure (Pra), and the rate of temperature rise
(aTra/dt) can be calculated by using the ARC data. The
relationship among themselves can be described as follow
(Huff,1982):
Tra = Tr + (1/aARC) (Tr - Trl) . (4)
log (Pra/Pr)=2140[1/(Tr-273.2+C) - 1/(Tra-273.2+C)] (5)
daTra/dt =
(ATr/dt) (1/aARC) exp[{-(74100/8.319) (1/Tra-1/Tr)] (6)
where
aARC = 1/{1+[ (wC)w(aTw/dt) }/[ (WC) x (aTr/dt) ]}
By using the data obtained from equations (4) - (6), we
can predict the vent sizing of the polystyrene polymeri-
zation reactors under the adiabatic conditions.
HOMOGENEOUS-VESSEL VENTING
The governing equations for relief vent rate requi-
rement can be obtained by considering the macroscopic
energy and mass balance on the vessel as shown in Fig.
2 For the bulk of the fluid in the vessel, the energy
balance and the mass balance can be described individua-
lly as follows:
Energy balance:
da P
—(Pvu) = Q - W(ul + >- ) (7)
dat 4,
Mass balance:
a
—(Pv) = -w (8)
at
For an ideal gas and an incompressible liquid, the
combined material and energy balance is (Leung, 1986;
Huff, 1982):
atr vt
mcpf = Q - Whfg(xl + —— ) (9)
dt vig
In this paper, we only discuss the particular case
of zero disengagement of liquid and vapor within the
vessel, the so-called uniform-froth or homogeneous-
vessel venting case. For this case xl = x, and vl = v =
vV/m; Eq. 9 then takes the form
atr Vv ohfg
mCp——— = mg - GA —— —— (10)
dat m vig
By using Eq. 8, Eq. 9, and the turnaround criterion, the
result equation for the homogeneous venting rate re-
quired to stop the temperature rise with zero disengage-
ment can be obtained:
m0q
W=GA= (11)
V hfg 2
((—————) 1/2 + (Cp T)1/2]
movfg
where T is simply the "overtemperature (Tm - Ts)". The
average heat generation rate q is evaluated from the
relationship:
a arr atr
Qe ==" Cpl( )s + (—)n) (12)
2 at at
where (dTr/dt)s and (dTr/dt)m are the self-heat rates at
the set temperature and turnaround temperature, respec-
tively. Eq. (11) is a simplified methodology for esti-
mating emergency vent sizes by computations that do not
require detailed kinetic data. The key data required
for the vent size estimation are the adiabatic self-heat
rate, the vapor pressure-temperature relationship and
the specific heat of the liquid.
DISCHARGE FLOW MODELS
Leung (1986) has proposed a generalized correlation
for HEM (Homogeneous Equilibrium Model) model with the
scaling parameter w given entirely in terms of known
stagnation properties
xvig CpTrP vfg
v=o + (2 (13)
v v hfg
In equation form, the generalized_correlation gives the
following normalized mass flux G/JP/v:
For w>4.0 (low-quality region)
G/(P/ve [0.6055 + 0.1356(1nw) - 0.0131(1nw)* j/w7® (24)
and for w<4.0 (high-quality region)
G/(P/vi= 0.66/w°?? (15)
The generalized HEM correlation greatly simplifies the
discharge critical flow calculation and is applicable
over the entire two-phase region. Furthermore, in the
all-liquid inlet condition, to get a good approximation,
the correlation may be replaced by
hfg 1
(
vfg CpTr
Mm
G 0.9
)0.5 (16)
In this paper, the generalized HEM correlation (Leung,
1986), Eq. (16), is used to calculate the relief discha-
rge of the styrene polymerization reactor. The process
parameters of styrene polymerization reactor is listed
in Table 2 (Huff, 1982).
RESULTS
Figures 3 - 5 illustrate the temperature profiles,
the pressure profiles, and the rate of temperature rise
- lll -
data for various initiator concentrations. They indi-
cate that the temperature and the pressure in the ARC
system will increase when a higher initiator concentra-
tion is used. The increase of the temperature and the
pressure will affect the sizing of the required vent
area for the styrene polymerization reactor. Figure 6
describes that a large vent diameter is needed when the
concentration of initiator increases. Because the ARC
system is assumed to be a good adiabatic reactor so that
the temperature in the ARC system wili not decrease
immediately when the styrene polymerization raction
comes to the end. Therefore, in Figures 3 and 4, the
tail end of the curves stays at a constant value for a
while. The same results can be obtained for various
styrene monomer concentrations.
Figure 7 is a plot of the rate of temperature rise
vs. time. The plot shows the difference of the rate of
temperature rise between the observed and the a=1 con-
dition. It is importent to design the required vent area
for a reactor to protect the opera-tors and the reactor
in safety. The safety of the predicted vent area can be
trusted by using the data of a=1 condition (i.e. adia-
batic condition). The impact of overpressure on vent
size is important in designing an emergency relief sys-
tem. Figure 8 depicts how the vent diameter varies
with overpressure for the styrene polymerization. Here
the percent overpressure is defined in terms of gage set
pressure as
Pm(bar g)
%overpressure = ( - 1) x 100 (17)
Ps(bar g)
Figure 12 illustrates an important aspect of homogeneou-
s-vessel venting: a drastic reduction in vent diameter
can be obtained with a small overpressure; and at higher
overpressure the relative reduction in vent diameter
becomes increasingly smaller.
CONCLUSIONS
In this paper, a simple ARC model has been provided
to estimate the required vent diameter for styrene poly-
merization, reaction by using the simplified-vent sizing
equations proposed by Leung in 1986. From the simula-
tion results, we can find that the vent diameter will
- 112 -
increase when a higher concentration of the initiator
and the styrene monomer is used. Although the vent
diameter becomes smaller at higher overpressure, it is
not recommended to design the relief system under such
high overpressure. Because a slight reduction in area
can bring about a drastic increase in peak pressure
reached.
REFERENCES
Boyle, W. J. Jr.,"Sizing Relief Area for Polymerization
Reactor", Chem. Eng. Prog., 63(8), 61(1967).
Fauske, H. K.,"A Quick Approach to Reactor Vent Sizing",
Plant/Operations Progress, 3(3), 145(1984).
Fauske, H. K.,"Emergency Relief System Design for Runa-
way Chemical Reaction: Extension of the Diers Methodolo-
gy", Chem. Eng. Res. Des. 67(3), 199(1989).
Huff, J. E., “Emergency Venting Requirements", Plant/O-
perations Progress, 1(4), 211(1982).
Jaisinghani, R. and W. H. Ray, "On the Dynamic Behavior
of a Class Of Homogeneous Continuous Stirred Tank Poly-
merization Reactors", Chem. Engng Sci., 32, 811(1977).
Leung, J. C., "Simplified Vent Sizing Equations for
Emergency Relief Requirements in Reactors and Storage
vessels", AIChE J., 32(10), 1622(1986).
Leung, J..C., and H. G. Fisher, "Two-Phase Flow Venting
from Reactor Vessels", J. Loss Prev. Process Ind., 2,
78(April, 1989).
able 1, Operating conditions and physical conatanta for poly-
styrene reaction
Inlet concentrations:
11) = 0.008 wo1/i
(4) = 10.881 wot
Inlet temperature:
tp 2 38214 K
feos
kp = 1:08 x 107 exp(-3657/RT,} (1/mol #)
kg = 6.95 x 10!exp(=14807/8T,) (07) (AKBND
key + 1:26 x 10% oxp(-843/nt,) (1/mol #)
Pcp # 360 cal/i K
om ADS:
‘Table 2, Process parameters of styrene polymerization resctor
V = 13.16 2° (3,500 gal)
mo = 8,600 kg
4.5 bar abs.
+ 5:4 bar mbs.(asmuming 10% above HAW?)
4,8 Bar Set 5.4 Bar Peak
vey wre 0.001988 0.001414
vgr #°/ks; Adel ges assumed 0.07278
Gps i/ke K 2.614
hegs ki /ke 302.3
W=GA
VOLUME=V
@ Lew
a
RADIANT
HEATER
Figure 1.
Schenstic diagram of the adiabatic reaction calorimeter
Figure 2. Reference ves
L for model development.
a
i
g, fa
- ?
Es
: :
a zl
re Gan.
igure 4 Tenpanature profites for various initiator emorstne
ona. (12160.01(1), 0.008(2), 0.0025(9, 0.00164).
- 114 -
20
10.0
PRESSURE (MR?
0
40
Figure 4,
to
+ RISE (C/MIN.
0%
10"
RATE OF TEMP
20"
o ny
TIME CHIN.
Figure 5. Rate of temperature rise data for various
initiator concentrations. ({1]=0.01(1),
9.00512), 0.0025(3), 9.091(4)),
@
TIME CHIN.
({71=0-01(2), 0.005(2), 0.0025(3), 0.001(4)).
Pressure profiles for various initiator concentrations.
osc
VENT OIANETER C11)
0.40
o.70
: Bo” ooo. 1.00
CaNC. OF INITIATOR CHOLEALITERD 0
Figure 6. The required vent di.
concentrations.
jeter for various initiator
- 115
wo
8
Si
RATE OF 1:
= rary
Tae GN.
Figure 7. Rate of teaperature rise data at various adiabatiosty.
(me1(1), raw date(2)).
1,00
am
VENT OICHETER. CH}
= 8 @
PERCENT GVERPRESSURE
Figure 8. The effect of overpressure on the required vent dia-
meter.
