0% found this document useful (0 votes)
12 views

Chap 01

The document discusses anti-surge systems for controlling level in tanks. It describes the dynamics and design of anti-surge control loops, which use only proportional action with low gain to provide overdamped response to surges in flow. The document also includes a worked example of sizing a buffer tank to limit rate of change of downstream flow during changes to upstream flow.

Uploaded by

megahedm
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
12 views

Chap 01

The document discusses anti-surge systems for controlling level in tanks. It describes the dynamics and design of anti-surge control loops, which use only proportional action with low gain to provide overdamped response to surges in flow. The document also includes a worked example of sizing a buffer tank to limit rate of change of downstream flow during changes to upstream flow.

Uploaded by

megahedm
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 5

Anti-Surge Systems

Chapter 93
93.1 Dynamics of Anti-Surge
93.2 Anti-Surge Design
93.3 Worked Example
93.4 Anti-Surge Control: Pressure Context
93.5 Comments
93.6 Nomenclature

There are two main categories of level control sys- can be designed to provide any degree of damp-
tem. First, and conventionally, those in which the ing. It is essentially a question of estimating the
level is controlled for its own sake. The dynamics size of surge that is likely to occur, specifying the
of such a level control system were considered in maximum rate of change that is acceptable, deter-
Chapter 84. The second category is where the exact mining the necessary tank capacity, and matching
level is unimportant,provided that the tank doesn’t the controller gain with the height of the chosen
overflow or run dry, and surges in flow are damped tank.
by allowing the level to rise and fall. The level con-
trol system on the reflux drum of a distillation col-
umn is a good example, as depicted in Figure 93.1. 93.1 Dynamics of Anti-Surge
The reflux rate is controlled by the temperature at
Consider the simple level control system depicted
the top of the column and a level controller used
in Figure 93.2. It is assumed that the dynamics of
to average out fluctuations in product rate, as de-
all the control loop elements other than the tank
scribed in more detail in Chapter 35.
are negligible.

F1
LC
LC
TC

h
X F0

Fig. 93.1 Control of level in a reflux drum Fig. 93.2 Tank with level control loop

The basic strategy is for the control loop to have The corresponding block diagram is depicted in
P action only with a low gain, such that the closed Figure 93.3. The process transfer functions are as
loop response is overdamped, and to exploit its in- developed in Chapter 84. Note that all the signals
herent offset. A tank with a level control system shown are in deviation form.
758 93 Anti-Surge Systems

F1(s) KL
TP s + 1
+
hR(s) e(s) u(s) X(s) − KP + h(s)
-KC KX
+ TP s + 1
-

KM

Fig. 93.3 Block diagram of level control system

Using block diagram algebra yields the closed loop h


response: H

KL
h(s) = .F1 (s)
(TP s + 1)
KC .KX .KP t
+ . (hR (s) − KM .h(s))
(TP s + 1) 0
Fig. 93.4 Response to a step change in inlet flow
Assuming the set point is fixed at the normal level
in the tank, which it would be for anti-surge con-
If the steady state closed loop offset is H, as de-
trol, then hR (s) = 0.
picted in Figure 93.4, then:
Whence:
KL .F1
KL H= (93.3)
(1 + K)
(TP s + 1)
h(s) = .F (s)
KM.KC .KX .KP 1 Differentiating Equation 93.2 gives:
1+
(TP s + 1)
KL 1 dh KL .F1 e−t/T
= . .F1 (s) = .
(1 + K) (Ts + 1) dt (1 + K) T
d2 h KL .F1 e−t/T
where: 2
= − .
dt (1 + K) T2
TP
K = KM.KC .KX .KP and T= (93.1)
1+K
93.2 Anti-Surge Design
If a step change in feed rate occurs, which is the Following a step change in feed rate, the important
worst possible surge, i.e. F1 (s) = Fs1 : variable for anti-surge control is the rate of change
of outlet flowrate, dF0 /dt.
KL .F1 1
h(s) = . An unsteady state volume balance gives:
(1 + K) s (Ts + 1)
dh
The response as a function of time is thus: A = F1 − F0
dt
KL .F1 d2 h dF0
h= . 1 − e−t/T (93.2) A = −
(1 + K) dt 2 dt
93.3 Worked Example 759

Thus: 93.3 Worked Example


dF0 A.KL .F1 e−t/T
= . A reaction is carried out by using three reactors in
dt (1 + K) T2
parallel and the product is sent to a recovery col-
Noting that the maximum rate of change of outlet
umn.The normal flow from each reactor is 10 m3 /h.
flowrate, the principal constraint, is at t = 0, and
Each reactor is shut down about every 10 days for
substituting from Equation 93.1 gives:
an 8-h period. To protect the column from changes

dF0  A.KL .F1 A.KL .F1 . (1 + K) in feedrate it is proposed to install a buffer vessel
= =
dt max (1 + K) .T2 T2P between the reactors and the column. What size of
(93.4) storage tank is needed to guarantee a maximum
Substituting from Equations 84.8 and 84.9, in change in column feedrate of 10%/h?
which it was shown that TP = A.KL , and also from Assume that, because the reactors are shut
Equation 93.1 gives: down every 10 days for an 8-h period, the flow into
 the buffer vessel will always be either 20 or 30 m3/h.
dF0  F1 . (1 + K) F2 A strategic decision has been made that the buffer
 = = 1
dt max A.KL A.H vessel is not intended to cope with the situation
Whence: when two reactors go off stream simultaneously.
F21 The worst case scenario is when the third re-
A.H =  (93.5)
dF0  actor comes on stream which represents a 50% in-
dt max
crease in flow, rather than when the third reactor
goes off stream which represents a 33% decrease
This is the change in volume that occurs when the in flow.
rate of change of outlet flowrate is restricted to If F1 = F0 = 20 m3 /h and F0 is to be limited to
some specified maximum value following a step 10% per hour, then:
change of magnitude F1 in inlet flowrate. For any 
particular anti-surge duty, AH represents the min- dF0 
= 2 m3 h−2 .
imum tank volume required. Clearly, in practice, a dt max
volume slightly larger than AH is required to pre-
vent flooding, and allowance must be made for the Substituting into Equation 93.5 gives the minimum
head necessary to produce the outlet flowrate un- volume necessary:
der minimum flow conditions.
F21 102
The relative values of A and H may be chosen V=  = = 50 m3 .
dF0  2
dt 
to allow for the shape of tanks available. Having
max
established a value for H, the controller gain KC
to give the required damping may be determined A vessel of 4 m diameter and 4 m height (50.3 m3 )
from Equations 93.1 and 93.3: would be adequate for the job. Remember that al-
KL .F1 KL .F1 lowance must be made for the head required to
H= = produce the flowrate of 20 m3 /h when there are
1 + K 1 + KM.KC .KX .KP
only two reactors operating.
Hence:
 
1 KL .F1
KC = −1
KM.KX .KP H
93.4 Anti-Surge Control:
Note that KP and KL are defined explicitly in Chap-
ter 84. KM and KX represent the calibrations of the
Pressure Context
level sensor/transmitter and converter/actuator Consider the anti-surge control system shown in
respectively. Figure 87.6 whose block diagram is as depicted
760 93 Anti-Surge Systems

in Figure 87.7. Block diagram algebra yields the Again, analogous to Equation 93.4, the initial rate
closed loop response: of change is maximum, i.e. at t = 0:

KL dQ0  C.KL .P1
.P1 (s) =
dt 
P0 (s) =
(TP s + 1) (1 + K) .T2
max
(−KC ).KX .(−KP ) Substituting from Equation 93.1 for T and from
+ (PR (s) − KM .P0 (s))
(TP s + 1) Equation 87.6 for TP and KL :
Assuming the set point is fixed at the normal pres- TP R0 R1 C R0
T= TP = KL =
sure, which it would be for anti-surge control, then 1+K R0 + R1 R0 + R1
PR (s) = 0. Whence: gives:

KL 1 dQ0  (R0 + R1 ) . (1 + K) .P1
P0 (s) = . .P1 (s)  = (93.7)
(1 + K) (Ts + 1) dt max C.R0 .R12
where: which relates the resistances of the valves, the min-
TP imum capacity of the anti-surge vessel, the size of
K = KM.KC .KX .KP and T= . (93.6) the step input and the open loop gain.
1+K
If a step change in supply pressure occurs, which is
the worst case scenario, i.e. P1 (s) = Ps1 , then
KL .P1 1 93.5 Comments
P0 (s) = .
(1 + K) s (Ts + 1) Some judgement is necessary in deciding upon the
set point for the level loop. For example, in a level
The response as a function of time is thus analo-
context, if the tank is to protect against sudden,
gous to Equation 93.2:
but reversible, decreases in flow from the normal
KL .P1 flow,it would normally be operated full with the set
P0 = . 1 − e−t/T
(1 + K) point at the top of the tank. Conversely, for sudden
reversible increases, the tank is normally empty
Differentiating this gives:
with the set point at the bottom. If the tank is re-
d2 P0 KL .P1 e−t/T quired to protect against both sudden increases
2
=− . and decreases about the normal flow, a tank of ca-
dt (1 + K) T2
pacity 2AH would be required with the set point at
An unsteady state volume balance gives: half full.
dP0 Note that these designs are based upon worst
C = Q1 − Q0 case scenarios, i.e. a step change in inlet flow and
dt
the maximum slope at time t = 0. In practice, anti-
d2 P0 dQ0
C 2 ≈− surge systems with smaller capacity will provide
dt dt adequate protection for most purposes.In this con-
The approximation is in recognition of the fact that text it is appropriate to introduce a touch of inte-
the flow Q1 will not necessarily be constant follow- gral action to gently restore the level to the normal
ing the step change in P1 . This depends on whether value in between disturbances.
the flow is sonic or turbulent, and on the relative And finally, remember that anti-surge control
values of the pressures. is a capital intensive solution to the problem of av-
eraging out fluctuations. It is much cheaper, if pos-
Thus
dQ0 C.KL .P1 e−t/T sible, to use feedforward control to manipulate the
= . process down stream to accommodate the surges
dt (1 + K) T2
than to put in an anti-surge vessel.
93.6 Nomenclature 761

93.6 Nomenclature
A cross sectional area m2
e error m
F flow rate m3 s−1
h level m
H steady state level m
K gain
P pressure bar
Q flow rate (measured at 1 bar) m3 s−1
t time s
u controller output
T time constant s
X fractional opening of the valve –

Subscripts
C controller
L load
M sensor and transmitter
P process
R set point
X I/P converter and actuator
0 outlet
1 inlet

You might also like