Buffer Tank Design for Acceptable Control

Performance


Audun Faanes and Sigurd Skogestad
Department of Chemical Engineering
Norwegian University of Science and Technology
N–7491 Trondheim, Norway
April 4, 2003

Ind. Eng. Chem. Res.

(in press)

Abstract
This paper provides a systematic approach for the design of buffer tanks. We consider
mainly the case where the objective of the buffer tank is to dampen (“average out”) the
fast (i.e., high-frequency) disturbances, which cannot be handled by the feedback con-
trol system. We consider separately design procedures for (I) mixing tanks to dampen
quality disturbances and (II) surge tanks with averaging level control to handle flow-rate
disturbances.

 also affiliated with Norsk Hydro ASA, Corporate Research Centre, N-3907 Porsgrunn, Norway, E-mail:
 Author to whom all correspondence
audun.faanes@hydro.com, Tel.: +47 35 92 40 21, Fax.: +47 35 92 32 63
should be addressed. E-mail: skoge@chemeng.ntnu.no, Tel.: +47 73 59
41 54, Fax.: +47 73 59 40 80

1
1 Introduction
Buffer tanks are common in industry, under many different names, such as intermediate storage
vessels, holdup tanks, surge drums, accumulators, inventories, mixing tanks, continuous stirred
tank reactors (CSTRs), and neutralization vessels. We start with a definition:

A buffer tank is a unit where the holdup (volume) is exploited to provide smoother
operation.

We here focus on buffer tanks for liquids, although most of the results may be easily
extended to gas- or solid-phase systems. Buffer tanks may be divided into two categories,
namely, for (A) disturbance attenuation and (B) independent operation:

A. Buffer tanks are installed between units to avoid propagation of disturbances for contin-
uous processes.

B. Buffer tanks are installed between units to allow independent operation, for example
during a temporary shutdown and between continuous and batch process units.
In this category there is a continuous delivery or outdraw on one side and a discontinuous
delivery or outdraw on the other side. The design of the tank size for these types of
buffer tanks is often fairly straightforward (typically equal to the batch volume) and is
not covered further in this paper.

LC
Quality
Flow rate

(I) Averaging by mixing (mixing tank) (II) Averaging level control (surge tank)

Figure 1: Two types of buffer tanks

In this paper we focus on category A. There are two fundamentally different disturbances,
namely, in quality and flow rate, and two approaches to dampen them (see Figure 1):

I. Quality disturbances, e.g., in concentration or temperature, where we dampen by mix-

ing. Such buffer tanks are often called mixing tanks or neutralization vessels for pH
processes.

II. Flow-rate disturbances, e.g., in the feed rate, where we dampen by temporarily changing
the volume (level variation). Such buffer tanks are often called surge tanks, intermediate
storage vessels, holdup tanks, surge drums, accumulators, or inventories.

2
 In both cases the tank volume is exploited, and a larger volume gives better dampening: In
the first case, mixing of a larger volume means that the in-flow entering during a longer period
is mixed together, and in the second case, larger level variations are allowed.
Often, in the design of buffer tanks, the residence or hold-up time is used as a measure
instead of the volume. The residence time is defined as 

, where is the volume 

and the nominal flow rate   
.
Even if the buffer tanks are designed and implemented for control purposes, control theory
is rarely used when sizing and designing the tanks. Instead, rules of thumb are used. For
example, textbooks on chemical process design seem to agree that a half-full residence time
of 5-10 minutes is appropriate for distillation reflux drums and that this also applies for many
other buffer (surge) tanks. For tanks between distillation columns, a half-full residence time
of 10-20 minutes is recommended (Lieberman, 1983; Sandler and Luckiewicz, 1987; Ulrich,
1984; Walas, 1987; Wells, 1986).
Sigales (1975) sets the total residence time as the sum of the surge time and a possible
settling time. The following surge times are recommended: distillation reflux, 5 minutes;
product to storage, 2 minutes; product to heat exchanger or other process streams, 5 minutes;
product to heater, 10 minutes. The settling time applies when there is an extra liquid phase.
For water in hydrocarbons, a settling time of 5 minutes is proposed.
None of the above references provide any justifications for their rules.
The most complete design procedure for reflux drum volumes is presented by Watkins
(1967), who proposes a half-full volume given by
!"$#&%'()" +* %,.- (1)
Here  (typical range 0.5-2) and $# (typical range 1-2) are instrumentation and labor factors,
respectively, related to buffer tanks of category B mentioned above. For example, the value of
$# may be based on how much time it takes for the operator to replace a disabled pump. ( and
reflux and product rates, and the factor  (typical range 1.25-4) is dependent on how
*wellareexternal 
units are operated (e.g., 1.25 for product to storage). .- (typical range 1-2) is a
level indicator factor. The method gives half-full hold-up times from /021 to 345687 .
In addition to the volumes proposed above, one normally adds about /:9<; of the volume
to prevent overfilling (Wells, 1986). For reflux drums, 25-50% extra volume for the vapor is
recommended (Sandler and Luckiewicz, 1987).
A basic guide to the design of mixing tanks is given by (Ludwig, 1977).
The process control literature refers to the level control of buffer tanks for flow-rate damp-
ening (surge tanks) as averaging level control. Harriott (1964), Hiester et al. (1987), and
Marlin (1995) propose controller and tank size designs that are based on specifying the maxi-
mum allowed change in the flow rate out of the buffer (surge) tank because this flow acts as a
disturbance for the downstream process. However, no guidelines are given for the critical step
of specifying the outlet flow-rate change. Otherwise, these methods have similarities with the
one proposed in the present paper.
To reduce the effect of the material balance control on the quality control loop, Buckley
(1964) recommends designing the buffer tank such that the material balance control can be
made 10 times slower than the quality loop. In practice, this means that the effect of the
disturbance on the quality at the worst-case frequency is reduced by a factor of 10. This
applies to both surge and mixing tanks.
There have also been proposals for optimal averaging level control, e.g., (McDonald et
al., 1986), where the objective is to find the controller that essentially gives the best disturbance
dampening for a given surge tank. To reduce the required surge tank volume, provided one

3
is willing to accept rare and short large changes in the outlet flow, one may use a nonlinear
controller that works as an averaging controller when the flow changes are small but where
the nonlinear part prevents the tank from being completely empty or full, e.g., (McDonald et
al., 1986; Shunta and Fehervari, 1976; Shinskey, 1996).
Another related class of process equipment is neutralization tanks. Neutralization is a
mixing process of two or more liquids of different pH. Normally this takes place in one or more
buffer (mixing) tanks in order to dampen variations in the final product. The process design for
neutralization is discussed by Shinskey (1973) and McMillan (1984). Another design method
and a critical review on the design and control of neutralization processes with emphasis on
chemical wastewater treatment is found in Walsh (1993). In (Faanes and Skogestad, 2002)
tank size selection for neutralization processes is discussed.
Zheng and Mahajanam (1999) propose the use of the necessary buffer tank volume as a
controllability measure.
The objective of this paper is to answer the following questions: When should a buffer
tank be installed to avoid propagation of disturbances, and how large should the tank be?
The preferred way of dealing with disturbances is feedback control. Typically, with integral
feedback control, perfect compensation may be achieved at steady state. However, because
of inherent limitations such as time delays, the control system is generally not effective at
higher frequencies, and the process itself (including any possible buffer tanks) must dampen
high-frequency disturbances. We have the following:
The buffer tank (with transfer function =>@?$% ) should modify the disturbance, A ,
such that the modified disturbance
ACB?$%DE=>@?$%FAG@?$% (2)
can be handled by the control system. The buffer tank design problem can be
solved in two steps:

Step 1. Find the required transfer function . (Typically =>@?$% =>@?$%HI/$JKL?MN/$%PO ,
Q
and the task is to find the order and the time constant .) 
Step 2. Find a physical realization of =>@?$% (tank volume and possibly level control
tuning).

In this paper we present design methods for buffer tanks based on this fundamental insight.

2 Introductory example
The following example illustrates how we may use (1) the control system and (2) a buffer
(mixing) tank to keep the output within its specified limits despite disturbances.
R
Example 1 Consider the mixing of two process streams, and , with different components S
R S
(also denoted and ), as illustrated in Figure 2.
R S
The objective is to mix equal amounts of and such that the excess concentration of
the outlet flow T+UVWT+XZY[T&\
is close to zero. More specifically, we require to stay within T+U
9^]_/`bacd!
. The combined component and total material balance gives the following
model:
A<T+U  / 8fTgXih j!YkTgUl%m nX^NT&\'h j!YkT+Uo%m o\'
Ae (3)

4
 r rpqt
pcqs s u
A
A,f
cB,f
B

Mixer

pcq 0
CIC

Buffer tank

pc
Figure 2: Mixing process. The concentration is controlled by manipulating the flow rate of
S
stream . Variations are further dampened by an extra buffer tank.

For the case with no control and no buffer tank, the time response in the outlet concentra-
tion,v)wT+U , to a step disturbance in the feed concentration, AxWTgXih j
, is shown by the solid
line (“Original”) in Figure 3. The value of approaches vbyTgU /:9z^acC 
, which is 10 times
larger than the accepted value.

{ 
o\
1. We first design a feedback control system, based on measuring , and manipulating v^ET U
to counteract the disturbance. We choose a proportional-integral (PI) composi-
tion controller, |~},€}z@?$%
9029m/D€?JN/%‚d?
. Note that the speed of the control system is
limited by an effective delay ƒ„_/…~6†7
, mainly due to the concentration measurement.
The resulting response with control is shown by the dashed line. Because the controller
has integral action, the outlet concentration returns to its desired value of . 95baCcC! 
However, because of the delay, the initial deviation is still unacceptable.

2. To deal with this, we install, in addition, a buffer tank with volume (residence /:‡5
time /.‡5687
) (drawn with dashed lines in Figure 2). We are now able to keep the outlet
T
concentration within its limit of ]z/…^acd!
at all times as shown by the dash-dotted
line in Figure 3.

Instead of the buffer tank, we could have installed a feedforward controller, but this re-
quires a fast (and accurate) measurement of the disturbance, Aˆ‰T+Xih Š
, and a good process
model. In practice, it would be very difficult to make this work for this example.

Comment on notation: Throughout the paper, the main feedback controller for the process
is denoted |W@?$%
, whereas the buffer tank level controller is denoted . ‹Œ@?$%
In the following sections we will show how to design buffer tanks for quality disturbances,
like in the above example, as well as for flow-rate disturbances.

5
 10

Original
8

Concentration
6

4 With control

2 With control and buffer tank

0 10 20 30 40 50 60 70
}
80
Allowed
range

Time [min]

Figure 3: Response in the excess outlet concentration to a step in inlet quality (from
/:9d9~bacLR`!b
to /:4d9^acR`!b :/ 95687
at ) for the system in Figure 2. A composition con-
troller handles the long term (“slow”) disturbance, but a buffer tank is required to handle the
short term deviations. Nominal data: Ž’‘
“:”–•F—™˜mšF—›–œ Ž"‘
“$”•m—ž˜Ÿšm—J›œ  ’‘¢¡“&“,—™£g¤V¥¦šm—ž˜
, , ,
 ­  ‘§¡“&“‚—ž£+¤x¨Jšm— ˜ ,  @©x‘_“‚—ž£+¤Z¥«ª¬¨šF— ˜ . Residence time mixer: ¡—J›œ . Delay in control loop
‘¡—J›œ . The levels in the mixer and the buffer tank are controlled by adjusting the outflow with PI con-
trollers, ®D¯±°²,‘³¯±•g“l°'´µ¡²$š'¯±•g“l°² .

3 Step 2: Physical realization of ¶w·$¸º¹ with a buffer tank

A
Consider the effect of a disturbance, , on the controlled variable . Without any buffer tank, v
the linearized model in terms of deviation variables may be written as

v»?$%M
¼¾½¿@?$%mA™?$% (4)

where ¼ ½¿
is the original disturbance transfer function (without a buffer tank). To illustrate
the effect of the buffer tank, we let =>@?$%
denote the transfer function for the buffer tank.
The disturbance passes through the buffer tank. With a buffer tank, the model becomes (see
Figure 4)

v»@?$%DÀÁ ¼ ½¿  Âo?$à %m=>@?$Ä % A™@?$% (5)

ÅiÆÇÉÈÊ
where ¼¾½@?$%
is the resulting modified disturbance transfer function. A typical buffer tank
transfer function is

=>Ë?$%º ÌL?N// $% O (6)

Note that =>9%D/

so that the buffer tank has no steady-state effect.
We will now consider separately how transfer functions of the form (6) arise for (I) =`?$%
quality and (II) flow-rate disturbances. In both cases, we consider a buffer tank with liquid
volume ¢
, inlet flow rate +Í O L!
, and outlet flow rate . Œ~…!

6
 Quality/Flow
disturbance

Ó ÏÐ
Buffer
Ô ÒÕ Î h(sÏ Ð)
tank
Process
Gd0(s)
d(s)
GÑ d(sÒ )
Figure 4: Use of buffer tank to dampen the disturbance

I Mixing tank for quality disturbance ( ֒×ÙØÚ8Û )

Let T ÍO T
denote the inlet quality and the outlet quality (for example, concentration or tem-
perature). For quality disturbances, the objective of the buffer tank is to smoothen the quality
response

T!€?$%DE=>@?$%mT Í O ?$% (7)

so that the variations in T are smaller than those in TÍ . A component or simplified energy
balance for a single perfectly mixed tank yields AG@ HTo%‚ACe) Í T Í Y¢ $T . By combining
O
this with the total material balance A< ACeVÜ gÍ Y[ (assuming constant density), we obtain
O O
HAToCAe™E +Í O TÍ O Y«Tl% , which upon linearizationO and taking the Laplace transform yields

à á á&ç á&ç á&ç ánç
è è
T!@?$%M Ýß Þ Þ ?/ N/ Í2âGã@ä$åæ â`ê é ç ê è âGã@ä$å é â`ë é ç ë ãä$åPì (8)
á
áwhere an asterisk denotes the nominal (steady-state) values and the Laplace variables ã@ä$å ,
è âíã@ä$å , ê è âíãËä$å , and ë ã@ä$å now denote deviations from the nominal values. We note that flow-
rate disturbances (in ê ánè ç â ) may á&ç result in quality disturbances if we mix streams of different
compositions (so that âî è ï ). From (8), we find that the transfer function for the tank is
ð ã@ä$å ï ñ
çë +ó çzôõö ò äæ ñ (9)

where ò ï ê is the nominal residence time. We note that the buffer (mixing) tank
works as a first-order filter. Similarly, for ÷ tanks in series, we have
ð ã@ä$å ï ø â ñ
èúùiû ã ò è äJæ ñ å (10)
ç
ò ï
where è ç residence time in tank ü . We find the required volume of each tank from è
ë ï ò ê
è è,
ê è
where is the nominal flow rate through tank ü .

II Surge tank for flow-rate disturbance ( ýÿþ
 )

For flow-rate disturbances, the objective is to use the buffer volume to smoothen the flow-
rate response
ê ãä$å ï ð @ã ä$å ê è íâ ã@ä$å (11)

7
The total mass balance assuming constant density yields
ë
 ï ê èâ é ê (12)

 ë ó  
 ê
ê èâ Œã@ä$å
We want to use an “averaging level control” with a “slow” level controller, because tight level
control yields and . Let denote the transfer function for the level
controller including measurement and actuator dynamics and also the possible dynamics of an

ê ãËä$å ï Œ ãä$å'ã ë ã@ä$å é ë  ã ä$åå

inner flow control loop. Then

ë
(13)

where is the set-point for the volume. Combining this with (12) and taking Laplace trans-
forms yields
ô ö
ë ãä:å ï ñ ê è íâ ã@ä$å…æ¾ã@ä$å ë  ã ä$å
äæ`ãä$å (14)

or from (13):
ô ö
ê ãä$å ï »ã@ä$å ê è âíã@ä$å é ä ë  @ã ä$å
äæŒãä$å (15)

The buffer (surge) tank transfer function is thus given by

ð ã@ä$å ï Œã@ä$å ï  ñ
ä æ¾ã@ä$å   æ ñ (16)
ó
With a proportional controller, ¾ã@ä$å 
  ð
ð
ï , we get that ã@ä$å is a first-order filter with ò ï ñ  .
Alternatively, for a given ã@ä$å , the resulting controller is
ä ð ã@ä$å
`ã@ä$å ñ é ð ã@ä$å
ï (17)

Compared to the quality disturbance case, we have more freedom in selecting ã@ä$å , be-
ð
cause we can quite freely select the controller 'ã@ä$å . However, the liquid level will vary, so
the size of the tank must be chosen so that the level remains between its limits. The volume
variation is given by (14), which upon combination with (17) yields

ë ã@ä$å ï ñ é ð @ã ä$å ê è â ã@ä$å

ä (18)
ë ã@ä$å
ä ï 
Note that represents the deviation from the nominal volume. The maximum value of this

ð @ã ä$å ï ó ã ò Mä æ å â
transfer function occurs for all of our cases at low frequencies (
In Table 1 we have found the level controller
ó
).
ŒãËä$å
and computed the required total volume
ð @ã ä$å ï ó ã ò Mä æ å
for ñ ñ
ñ ò
. For example, for a first-order filter, , the required ñ ë ñ  ï ò ~ ê"!$#%
controller is a P controller with gain
Note that the resulting level controllers, Œã@ä$å
and the required volume of the tank is .
, do not have integral action. A level controller
without integral action was also recommended and further discussed by Buckley (1964, page
167) and Shinskey (1996, page 25).
ð ãä$å
tanks with a P level controller, Œã@ä$å
For flow-rate disturbances, a high-order can alternatively be realized using multiple
, in each tank. However, the required total volume is the

8
choices of .
ð ã@ä$å
Table 1: Averaging level control: Design procedure II for flow-rate disturbances for alternative

ð ã@ä$å & ó û' ò û  û& ' û û û ')(  & '' û û  **.-
Step 1st order 2nd order nth order

'ã@ä$å +/:&ò (, û  &ò û û

2.1. Desired (from Step 1)

ë ã  å ó ê è Gâ ã  å
2.2. Level controller, from (17) ñò
2.3.
2.4.
ë01
from (18)
ò ~ê"!$#% /:ò2 ê!3#% ÷ ò2~ ê"!$#4%
÷

ŒãËä$å , so the latter is most

same as that found above with a single tank and a more complex
likely preferable from an economic point of view.
4 Step 1: Desired buffer transfer function
57698;:
ð ã@ä$å
What is a desirable transfer function, ? We here present a frequency-domain approach for
ð ãä$å ï ó ãã 9ò < ó ÷ åiäJæ å â
÷ ï ò9< ñ = ò<
answering this question. Figure 5 shows the frequency plot of for ñ ñ
÷ ï ñ ï@? ï ó BA ?C?
to , where in most cases is the total residence time in the tanks. With a given value
/
ï ñ ñ == > ÷ ïE? ñ ÷ ï
ó BA ó BAGCH
of , we see that is “best” if we want to reduce the effect of the disturbance at a given

about ( D
frequency by a factor
), and
( ) or less; is “best” if the factor is between 3 and
is “best” if the factor is between about 7 and (
÷ ð
Thus, we find that a larger order is desired when we want a large disturbance reduction. We
). ñ.F ï ñ =
now derive more exactly the desired . ã@ä$å
0
10

Gain 0.33
(min.vol. n=1)

Gain 0.144 (min.vol. n=2)

−1
10
Gain 0.064 (min.vol. n=3)
Magnitude

n=1

−2
10
n=2

n=3
n=4
−3
10 −1 0 1 2

óJI &LK
10 10 10 10
Frequency × τh

ð â
Figure 5: Frequency responses for ã@ä$å ï ñ â Jä æ Nñ M .

 onLettheusoutput O is then
start with an uncontrolled plant without a buffer tank. The effect of the disturbance

¾O ãä$å ïQP ãä$åBHR ã@ä$å…æ PJSUT ãä$å  ã€ä$å (19)

9
To counteract the effect of the disturbances, we apply feedback control (
Figure 6). The resulting closed-loop response becomes
R ï éWV O ) (see
Oíã@ä$å ïQX ã@ä$å P S T ã@ä$å  ã@ä$åZY X ï
ñæ
ñP
V (20)

\d
^h
Gd [
Gd0 [
]y r
+
K
u
G +
]y
- +

Figure 6: Feedback control system

X X PWS T
With integral action in the controller, the sensitivity function approaches zero at low
frequencies. However, at higher frequencies, the disturbance response, , may
still be too large, and this is the reason for installing a buffer tank. The closed-loop response
_ ã`ažå ã`aå._
O¾ãä$å ïQX ã@ä$å b JP SUT giã@ec ä$h d å  ð ã@ä$f å  ã@ä$å
with a buffer tank is
(21)

which is acceptable if _ X;PWS T ð _

is sufficiently small at all frequencies. We need to quantify the

PWS T
term “sufficiently small”, and we define it as “smaller than 1”. More precisely, we assume that

j _  _Bk ñ , il a )
the variables and thus the model ( ) has been scaled such that

j The allowed output variation is less than 1 ( _ Om_0k ñ n l a )

The expected disturbance is less than 1 (


From (21) we see that to keep _ Om_ok ñ when _ _ ï ñ (worst-case disturbance), we must
require

_ X ã`aå PJS"p ã`aå ð ã`aå._qk ñ Yrlia (22)

ð
á  átsiu v
from which we can obtain the required ãä$å . We illustrate the idea with an example.

Example 1 (continued) (Mixing process). Let O ï p , ï , and R ï êew . Linearizing and

P S T ãä$å ï äñ æ  Y X ã@ä$å ï æ pyñ Jx z { -  Y ð ã@ä$å ï äñ æ
scaling the model (3) then yields

ñ.| ñ
(23)
ñ ñ
10
 
 
~ 
€ ‚ ás
ó ~„ƒ
We here usedó for the scaling the following:
w 0
 A ~ ƒ †
~ 
… ‡ á expected

~ 
€ ‚ ó ~ variations
ƒ in , } /
for , }
ê F } ; range

In Figure 7 we plot the disturbance effects _ PJS‰ˆ _ , _ XŠPJSUT _ , and _ XŠPJSUT _ as functions of
; allowed range for : ñ .
ð
frequency. Originally (without any buffer tank or control), we have _ PJS T _ ï ñ at lower

frequencies. The introduction of feedback makes _ X;PWSUT _J‹ ñ at low frequencies, whereas
adding the buffer tank brings _ XŠPJS T _B‹
ð
ñ
also at intermediate frequencies.

1
10
Gd0

SGd0
Magnitude

0
10

SGd0 h

−1
10 −2 −1 0 1 2
10 10 10 10 10

WP SUT XŠPJSUT Œ~ …‡

Frequency [rad/min]

Figure 7: Original disturbance effect (

Š
X P S T ð
), with feedback control ( ) and with feedback
ñ9|
to bring _ X ã`aå PJSUT ã`aå ð ã`a å_C‹ ñ a A
control and a buffer tank ( ). A buffer tank with a residence time of
for all
is required

In the following we will present methods for finding

ð ã@ä$å based on the controllability
requirement (22). There are two main cases:

ready designed, so ãËä$å X

is known. The “ideal”
ð
S. Existing plant with an existing controller: The “counteracting” controller,
@ã ä$å
V
is then simply the inverse of
ãä$å XŠPJSŽT
, is al-
.

N. New plant: The “counteracting” controller, V ãä$å , is not known so X ãËä$å is not known.

ð ð ó
This is the typical situation during the design stage when most buffer tanks are designed.
â
In most cases we will choose ãä$å to be of the form ã@ä$å ï ñ ã ò äMæ ñ å .
4.1  given (existing plant)
ð ã@ä$å such
ð_ ã@ä$å._q‹ ó _ XŠP S T _yY lia
We consider an existing plant where controller V ãä$å
is known. The task is to find
that ñ ó
. Several approaches may be suggested.

ó ð
approach with ã@ä$å ï ã ò äMæ ñ å â : This
ó ð
is done by selecting ã@ä$å ï
ñ ã ò äMæ ñ å and adjusting ò until _ ã@ä$å9_ touches ñ _ XŠPJS T _ at one frequency. As a
S1. Graphical
â ð ñ
starting point we choose the following:

÷ _ XŠP S T _ in a log-log plot in the frequency area where _ XŠP S T _C ñ.

is the inverse of the frequency where _ XŠP S T _ crosses one from below.
(a) is the slope of
(b) ò

11
 ð ó
ó â
with ã@ä$å ï ñ ã ò äJæ ñ å : With a given ÷
just touches ñ _ XŠPJS T _ by solving the following problem:
S2 Numerical approach ò
we find such that _ð _
ò ï ~’” ‘9“ òN•—–L˜ ã1aå (24)

_ X ã `ažå P S T ã`aå_ +U â é ñ Yž_ X ã`aå P S T ã`aå._0 ñ (25)

where

òe•—–L˜ ã™aå  ”;û œ

Y ~’‘9“ ” with
Because it is not practical to calculate òe•—–L˜ ã1
otherwise

~’‘9“ ”Ÿ , where a è¡ £¢ , which is a finite seta å offorfrequencies

all frequencies, we replace

ð ã@ä$å ï ó ã äJæ å
from the range of interest.

ñ .ñ | ñ
The calculation is explicit and fast, so a large number of frequencies can be used. (This
approach was used to obtain in Figure 7.)
As illustrated in Example 2 (below), for ¤ ÷  ñ one may save some volume with the
following approach, which is more involved since it includes nonconvex optimization.
ð
S3. Numerical approach with “free” ã@ä$å : We formulate a constrained optimization prob-
lem that minimizes the (total) volume of the buffer tank(s) subject to (22). As in the
previous method, we formulate the optimization for a finite set of frequencies, , from ¢
the frequency range of interest.

(I) Quality disturbances: For ÷ mixing tanks

ð ãä$å ï ñ
ã ò û Jä æ ñ å ¥.¥.¥.ã ò â äJæ ñ å (26)

ó
of the volumes ( ò ï ë ê
when the tanks are not necessarily equal. Because the flow rate is independent
), we may minimize the total residence time (instead of
minimizing the total volume) subject to (22):
~&L¦ u x …x ‡u & * ò û 槥.¥.¥dæ
Œ òâ

_2ã ò û `a è æ ñ åZ¥.¥.¥:ã ò âN`a è æ ñ å._C¨7_ X ã)`a è å PJS T ã`a è åe_yYŠa è ©¢
subject to (27)

¢
where is a set of frequencies. This is a single-input, single-output variant of a
method proposed by Zheng and Mahajanam (1999).
(II) Flow-rate disturbances:
ð ã@ä$å ï Œãä n4ª å
äæ«ŒãËä nLª å (28)

where we have parametrized the level controller with the parameter vector . We
minimize subject to (22) the required tank volume (14):
ª
~Œ¬ …‡ ë ï †~ ¬ …‡ ~’”Ÿ­.‘.®„“ ¯ è ñ è ¯
¯ ¯ `a æ«Œã`a n4ª å ¯¯
¯ èto ¯
ã)`a è å PJS T ã`a è å `a è »æã`Œa ã)`n4aª åè å ¯ k ñ Y°a è ±¢
subject (29)

¯¯¯ X n4ª ¯¯¯

¯
12
 Many controller formulations are possible, for example, the familiar PI(D) (D=derivative)

state gain, ,   m÷ ²
real zeros, and real poles: ÷ ³
controller or a state-space formulation. We here express the controller by a steady-

 û +
Œãä nLª å ï   ãã ´ ò û äJäæ æ ñ å'åiã1ã ´ ò + ääæ æ ñ å å ¥.¥.¥.¥.¥.¥.ã ã™ò ´ ââ¶iµ ääJæ æ ñ å å (30)
ô  AeAeA ñ ö ñ ñ
and thus ª ï  n ´ û n n ´
' â µ n ò û n AeANA n ò ⍶ .
With ÷ ² ï
 and ÷ ³ ï ñ in (30) we get
ð ã@ä$å ï + + ñ
ò ä æ /$ò · äºæ ñ (31)

· ‹ ñ does not give real time  constants asó the previous approaches. For a first-
order filter (with ŒãËä:å ï  and ð ã@ä$å ï ñ ã ò äæ ñ å ), there is no extra degree of
freedom in the optimization, and we get the same result as that with (24).

P ã@ä$å ï C / ä æ { - 
Example 2 (Temperature control with flow-rate disturbance).

WP S T ã@ä$å ï ñ C Y (32)
ñ ñ
V†¸q¹)º ã@ä$å ï BA / F2» äæ ä ñ (33)
»
This may represent the process in Figure 8, where two streams ¼ and ½ are mixed, and we want
to control the temperature (O ) after the mixing point. Stream ¼ is heated in a heat exchanger,
¿ B ¾
A u d = Flow in

À Buffer
tank with
LIC
Á "slow" level

 y=ÃTemperature
control

TIC

Figure 8: Temperature control with flow-rate disturbance

R 
½  R O PJS T ï 
and the manipulated input, , is the secondary flow rate in this exchanger. The disturbance, ,
is variation from the nominal flow rate of . , , and are scaled as outlined above.
First consider the case without the buffer tank. Because ñ
, the disturbance has a

is not sufficient because, as seen in Figure 9,

approaches 100 at high frequencies.
_ XŠPJS T _
large impact on the output, and a temperature controller is certainly required. However, this
exceeds 1 at higher frequencies and it

disturbance at higher frequencies. The slope of

expect that a second order is the best.
_ XŠPJSŽT _
We thus need to install a buffer tank with averaging level control to dampen the flow-rate
ð$ėŠå is 2 after it has crossed 1, so one would

13
 1
10
SGd0(V=0)

SGdfree(V= 56)

Magnitude
2
0
10
SGd2(V=72)

SGd1(V=242)
−1
10 −3 −2 −1 0

_ X;PWS T _ 
10 10 10 10
Frequency [rad/s]

BAG / =Æ ‘ Ç ó!õ

Figure 9: A buffer tank is needed for the temperature control problem:
quencies above
Table 2.
. Comparison of
for fre-
for designs 1, 2 and 3 in _ XŠP S _ ï _ ŠX P S T ð _
Table 2: Buffer (surge) tank design procedure II (flow-rate disturbance) applied to the temper-
ature control example

Step
ð3Ä—ÅÈ Design 1
ð3ėōÈ
Design 2 Design 3
3ð Ä—ÅÈ ð3Ä—ÅÈ 2nd order
ðÉÄ—ÅÈ û '  p‰ƒÌx Ë p''û Ê û  (
û û pyz x ÊUp ÍLû Ï ( ' z ƒ x ƒ ' û
û
0+ʎAÎC+  = û ñ
1. Numerical approach to obtain S2: 1st order S2: 2nd order S3(II):

ėōÈ
ó ÐtÑÎÒ + ÏH ' û
 / û Í ¥ ? ûH Ð ï D /
2.1. Desired (from Step 1)

ë ÄÌqÈ Ä BÈ // Ð
/ == /C !3#% F H  Ð !$#4%
2.2. Level controller,
ë0 /D C 3! #% F
2.3.
2.4.

É —
Ä 
Å È ó Ä ÅŠÓ È +

ð ó 0AÎ ñ ò ñ . _ XŠP SUp _ crosses 1 at about
ï
frequency 0.024 rad/s, corresponding to ò
For the graphical approach S1, we use
0
ë 
 
 ñ  / / ò =  ï ÐyÔo= ÕUÖ/ ,
and because  y
Ð o
Ô U
Õ Ö
controller is 
ó
Ä—ÅÈ ï BAG ñ / Ä / ñ Å×Ó ñ È .
disturbance (II), we have from Table 1 that ï $
» = . The required level this is a flow-rate

É
ð —
Ä 
Å È ó Ä weÅØÓ consider È
For the more exact numerical approaches (S2 and S3),
ï ò three designs, and

controller, but as expected, +the required volume is large because

Ù
ð U
Ä 
Å È ó
the results are given in Table 2. Design 1 (with
Ä Š
Å Ó È
ñ $
ð ñ Ä—ÅÈ requires a P level
) only

ï ò is first-order. Design
2 (with
ðÉÄ—ÅÈ in (31)), ñ ñ ) gives a considerably smaller required volume. From design 3
(with
ðÙÄUÅÈ above 2.
the required volume is even smaller than with design 2, as expected. Little

In Figure 9 we plot the resulting _ XŠPJS _ for the three designs, which confirms that they
is gained by increasing the order of

stay below 1 in magnitude at all frequencies. These results are further confirmed by the time
responses to a unit step disturbance shown in Figure 10.
Buckley’s method (Buckley, 1964) gives a residence time of ñ

H Œ
~ 
… ‡ 
 Œ
~ 
… ‡
the minimum required residence time of about F
, which is much less than
C

disturbance needs to be reduced by a factor of ñ , and not ñ as Buckley implicitly assumes.
 
(see Table 2). The reason is that the

4.2  not given

ÚÉÄ—ÅÈ must be designed such that ÛÝÜ;ÞWßUà Ú ÛBá£â at all frequencies. However, at the design stage
The requirement is that (22) must be fulfilled; that is, the buffer tank with transfer function

the controller and thus Ü is not known. Three approaches are suggested:

14
 1

0.8

0.6
Scaled output

Design 2 (V=72)

0.4
Design 3 (V=56)

0.2

Design 1 (V=242)

−0.2
0 20 40 60 80 100 120 140 160 180 200
Time[s]

Figure 10: Temperature control with flow-rate disturbance: Response in the scaled output to a
unit step in the disturbance (flow rate) with different tank sizes and level controllers (Table 2).

ãä åÛ ÜæÛçèâ
N1. Shortcut approach: The requirement (22) must, in particular, be satisfied at the band-

î ÛåÞ ßUàêïeé)ò ðë ãoäíì.ñ Û Û Ú éë ãoäêì.ÛCá£âôóõ Û Ú é ë ãoäíì9ÛCá£â ö÷

width frequency where , and this gives the (minimum) requirement

(34)

In Skogestad and Postlethwaite (1996, p. 173-4) it is suggested that , where

is the effective delay around the feedback loop. However, to get acceptable robustness,
ã ä á 1ù úüø û ý.þ ÿ
we here suggest to use a somewhat lower value

ãoä
 â
.ý þ ÿ (35)

Skogestad (2003) proposes the following simple rule for estimating

 ý þÿ :
   

 
ý.þ ÿ’ç§ý 
 for PI-control
 for PID-control (36)
ëë çç


ý
where is the delay,  ¡ç¤âö
 is the inverse of a right half-plane zero  , and  is the
We now assume 
Ú é ì°ç£âö é â ì
time lag (time constant) number  ordered by size so that  is the largest time constant.
    , use ã ä ç â ö é ý.þ ÿ ì
 , and solve (34) to get ø
   !#" %$ ý þ1ÿ ÷ â (37)

÷ ç þ Þ ßŽà të ù ø úüû
)(
Û Ú éë ì.Ûmç¤âöC÷
where '& +* -, /. ** . Alternatively, Figure 5 may be used for a given 0 to read off
ç ö é Zãoäì rç ã
**
the normalized frequency * 1 23 where 1 , and the required  for each
tank is then  41 0 .
15
 ãä
N2. Numerical approach based on preliminary controller design: The above shortcut
method only considers the frequency . To get a more exact design, we must consider
all frequencies, and a preliminary controller design is needed. This approach consists of
two steps:

N2a. Find a preliminary controller for the process, and from this, obtain Ü é ì .
N2b. Use one of the approaches S1, S2, or S3 from section 4.1.

value on the peak of ): Ü <>=?

ç yö
For step N2a, we have used the method of Schei (1994), where we maximize the low-
frequency controller gain 576 98;: 6 , subject to a robustness restriction (maximum


 6 @8 :

HGJtoI
ö
ÛÝÜ é)ë ã ìNÛ ã Ü stable
subject (38)
BADCFE and

é ì ç é
where for a PI controller 5  K8 :  6   § 
â 
ì ö é
6 ì
problem that Schei uses, we have added the constraint that Ü is stable. This is imple-
. Compared to the optimization

mented by requiring the eigenvalues of Ü L to be in the left half-plane, where Ü L is obtained

from Ü by replacing the delay with a Padé approximation. To obtain a robust design,
CE should be chosen low, typically @â MON $  . With this controller design, we then use
one of the methods S1-S3 to design the buffer tank.

N3. Numerical approach with a simultaneous controller and buffer tank design. A more
exact approach is to combine the controller tuning and the buffer tank design optimiza-
tion into one problem. For (I) quality disturbances, the optimization problem may be
formulated as an extension of (27):<>=?

PRQTS U U URS PRVS WYX  4ZZZ[ 
subject to _^`!a
ø  2GbI
Û é  ø ë ã §âì é ë ã §âì.Û 7ÛÝÜ é)ë ã ì ÞWßUà é)ë ã ìeÛ ã
\ZZZ    c^`!a ]  HGeI
ÛåÜ éë ã ì9Û ã
(39)
` a dADCFE
`/a Ü é ì stable
where is the controller parameter vector for 5 é  ì . Likewise for (II) flow-rate dis-
turbances, we get from (29):
<7=? <>=? <ihj
â ^R`
WS W Xgf ç WS W X k;lmn ** ë ã o8 é)ë ã ì ***
*
* to ^` *
_^`!a subject 8 éë ã ì ^R`  GeI
** Ü é)ë ã ìÞWßUà éë ã c^`!ì a ë ã o8 é)ë  ã HGeì I *** á£â ã (40)
**
ÛåÜ éë ã `/ì9a dÛ ADCFE ã *
` Ü é ì stable
where is the controller parameter vector for the level controller 8 é  ì , which enters in
`a
Ú é  ì , and is the controller parameter vector for the feedback controller 5 _é  ì , which
16
 Üéì
enters in  . To ensure effective integral action in 5 , these optimization problems
éì
must be extended by a constraint; for example, if 5  is a PI controller, a maximum
value must be put on the integral time.


Example 2 (continued) (Temperature control with flow-rate disturbance (II))
 dp p
ÞJß à3é ìŠçèâp;p Þ é ì;ç â §â
p;p   rqsut (41)

not known. The delay is ý çèâ

The available information of the process is given by (41), and we assume that the controller is
wv . We get the following
h| results:
N1. The shortcut approach yields (
(37) (or Figure 5) the following:
ã ä çxpBMOy{z ö{v and ÷ ç
ÛåÞJß à Ûêç âpdp for all ã ) from
} First-order filter (0©çèâ ): ç  pdpd>‚„ƒ…c† .
fu~€~
} Second-order filter (0©ç  ): 4
f‡~€~ ç ˆ]pd>‚„ƒ‰…† .
Ú é ì (0wŠçp and 0\‹±çèâ ) the following:
N2. The Schei tuning in (38) followed by the optimal design (29) yields for a second order

} CFE ç£@â MON :  ç y  7‚Œƒ‰…† .

u
f €
~ 
 ~
} CFE ç  : 
f‡~€~ ç ;d>‚„ƒ‰…† .
N3. Simultaneous controller tuning and optimal design (40) yields with second-order é ì
Ú
(0 Š  ç p and 0 ‹ çèâ ) the following:
} CFE ç£@â MON :  ç y  7‚Œƒ‰…† (as for method N2)
u
f €
~ 
 ~
} CFE ç  : 
f‡~€~ ç ;d>‚„ƒ‰…† (as for method N2)
Note that CE ç¤@â MON gives more robust (and “slow”) controller tunings than CFE ç


a second-order filter is fu~€~ ç

BŽ 7‚Œƒ‰…† (found with method N3 with CE free). Methods
and therefore requires a larger tank volume. The smallest achievable tank volume with

gives a tank volume very similar to that found with CE¡ç .

N2 and N3 yield almost identical results for this example. The shortcut method N1 also


5 Before or after?
If the buffer tank is placed upstream of the process, the disturbance itself is dampened before
entering the process. If it is placed downstream of the process, the resulting variations in the
product are dampened. The control properties are mainly determined by the effect of input  on
Þ Þ
output  (as given by the transfer function ). An upstream buffer tank has no effect on , and Þ
also a downstream buffer tank has no effect on provided we keep the original measurement.
On the other hand, placement “inside” the process normally affects . In the following we
list some points that may be considered when choosing the placement. We assume that we
Þ
prefer to have as few and small buffer tanks as possible (sometimes other issues come into
consideration, like differences in cost due to different pressure or risk of corrosion, but this is
not covered).

17
 1. In a “splitting process”, the feed flow is split into two or more flows (Figure 11(a)). One
common example is a distillation column. To reduce the number of tanks, it will then be
best to place the buffer tank at the feed (upstream placement). An exception is if only
one of the product streams needs to be dampened, in which case a smaller product tank
can be used because each of the product streams are smaller than the feed stream.

2. In a “mixing” process, two or more streams are mixed into one stream (Figure 11(b)). To
reduce the number of tanks, it is here best with a downstream placement. An exception
is if we only have disturbances in one of the feed streams because the feed streams are
smaller than the product stream, leading to a smaller required size.

(a) A splitting process (b) A mixing process

Figure 11: Two types of processes

3. An advantage of a downstream placement is that a downstream buffer tank dampens

all disturbances, including disturbances in the control inputs. This is not the case with
upstream tanks, which only dampen disturbances entering upstream of the tank.

4. An advantage of an upstream placement is that the process stays closer to its nominal
operation point and thus simplifies controller tuning and makes the response more linear
and predictable (see Example 3).

5. An advantage of the “inside” placement is that it may be possible to avoid installation

of a new tank by making use of an already planned or existing unit, for example, by
increasing the size of a chemical reactor.

6. A disadvantage with placing the buffer tank inside or downstream of the process is that
the buffer tank then may be within the control loop, and the control performance will
generally be poorer. Also, its size will effect the tuning, and the simultaneous approach
(N3) is recommended. For the downstream placement, these problems may be avoided
if we keep the measurement before the buffer tank, but then we may need an extra
measurement in the buffer tank to get a more representative value for the final product.

Example 3 (Distillation column). We apply the methods from section 4.1 to a distillation
column and compare the use of a single feed tank with the use of two product tanks (Fig-
ure 12). We consider a distillation column with 40 stages (the linearized model has 82 states;

øç ç
see column A from (Skogestad and Postlethwaite, 1996, p.425)). The disturbances to the col-
umn are feed flow rate and composition ( ‘ “’ and ‘ ”–• ), and the outputs are the mole

18
fractions of the component in top and bottom products, respectively ( and  ). The manip-
ulated variables are the reflux and the boilup ( ˜— and 
øç
f ). The variables have ç ø ç â
been scaled so that a variation of ™%;pBš in the feed flow rate corresponds to ‘ K™
â
a variation of ™ p]š in the feed composition corresponds to ‘ K™ . A change in the top
and
ç ¡â ø
â
and bottom product composition of ™%pBMOp mole fraction units corresponds to a change ™ ¡â
éø ø ìWç â
in  and  <7. =?Decentralized PI controllers are used to control the compositions. In the top,
5  ›NBMOœˆ  p  
delay of p
é §â ìö é ì
 p  , and in the bottom, 5 ›yuMˆ]N  p  
<>ž <>=? ç
 p  . There is aé Eâìö é ì
linear model. Nominally, the feed flow rate is
tions are 0.99 and 0.01, respectively.
â ö
in each loop, which we represent with fifth-order Padé approximations in the
, and the top and bottom concentra-

¡ y  =x¤
1 D
u1=L

V=49
LIC

¥ Ÿ  
V=68 d1=F
Ÿ
d2=zF

¦
u2=V V=94
¡ y¢ =x£
2 B

Figure 12: Distillation column with either one feed surge tank or two product mixing tanks to
dampen disturbances.

The holdup in the reflux and the boiler are controlled with § controllers (with gain p ) by
the top and bottom product streams, respectively.
â
and  without any buffer tank,
â ÛÝ܊ÞJß àÛ ÛÝÜ;ÞWß 1à Û ø
We consider the effect of the flow-rate disturbance, ‘ . The closed-loop gains from ‘ to 
Q and #¨ are shown with solid lines in Figure 13.
The gains are both above at intermediate frequencies, so our purity requirements will not be
ø ø
fulfilled, unless<> =? install a buffer tank.
we
Upstream placement (feed surge tank). is known, and with 0 Ü §ç â
ÛÝ܊ÞJçèß Û ââ Û ÜŠâ ÞJö ß Û Ú Û Û ÛÝÜ;ÞWß Û , (24) in method S2
yields  ©ˆ Q ¨ are shown with dashed lines, and we see
that â
. The resulting
#¨ just hits (as expected).
frequency, which is not at the maximum of
and

ÛåÜ;ÞWß 1à Û
is also plotted (dash-dotted) to indicate the limiting
¨ , but at a lower frequency “shoulder”.
Ú é ìŠçèâö é â⩈  Eâ ì
Following design procedure II, we now get the following:

2.1

2.2 The required level controller for the buffer tank is 8 é ìŠçèâ öBâCâYˆ çpuMOp;pdœ;œ
19

2.3 é Bìö é BìŠç çèâCâ
f p @‚  p ª > < Yˆ =? < ž <7=€? < ž
ç
2.4 f ~€~ ª«>‚ ƒ‰…† èç âCâ Yˆ Z  –Z pBMO Ndœ . ö ç
<7=€?
Comment: Since the slope of
order filters will increase the volume demand. For example, with 0
åÛ ÜŠÞ ß à Û
#< ¨ ž is less that 1 around the limiting frequency, higher
, (24) gives  ç 7ç
Ž NBMON , and f ~€~
 ç
\7‚ ƒ‰…c† ¬ MO . ç â
1
10

1/h

Limiting ω
SGd20
↓
Magnitude

0
10

SGd2

SGd1 SGd10

−1
10
−3 −2 −1 0
10 10 10 10
Frequency [rad/s]

Figure 13: Feed flow disturbance for the distillation column: Q and #¨ (for top
and bottom) are both above 1 (solid line). A feed tank with averaging level control, 
Û ÜŠÞ ß à Û ÛåÜŠÞ ß 1à Û Ú é ìÙç
â â öö Ú é âé â ì §â ì
©ˆ   , brings the disturbance gain to both top and bottom below 1 (dashed). Note that
 is just touching T¨ . Ûå܊ÞJß 1àÛ
â Downstream placement (product mixing tank). Because both Q u­ and #¨ ]­ ÛåÜ;Þ ß à Û @â ÝÛ ÜŠÞ ß à Û
at some frequencies, we must apply one mixing tank for each of the two products. When
we designed the ? tank, we had to consider the worst of <7=?Q and
<>=feed
Ûå܊ÞJß àÛ
Q for the top product and
¨ , but now we
ÛÝÜ;ÞWß 1à Û âCâ Ý
Û Š
Ü J
Þ ß à Û Û Š
Ü J
Þ ß 1 à Û ç â
may consider
 <'ž #¨ for the bottom product.
<>ž With 0 ,
(24) yields  as before Yˆ
corresponding volumes are ®Z]puMOy
 ž buffer tank@MOy and (top)
for the < top
ç âCâ
and ©ˆ>ZBpBMOy ¯y
Ž (bottom), which
for the bottom tank. The
ââ ©ç
gives a total volume of Nd , which is the same as that for the feed tank. However, the feed
tank placement is preferred because we then need only one tank.
Nonlinear simulations. The above design is based on a linearized model, and (as ex-
pected) the feed tank placement is further justified if we consider a nonlinear model because
the column is then less perturbed from its nominal state. This is illustrated by the simulations
in Figures 14, 15 and 16. If the buffer tanks are placed downstream, the nonlinear response
deviates considerably<7=? from the linear <7=€?response, and the tanks designed by linear analysis are
< ž < ž we
find that  ~ ;œ ç
too small. By trial and error

gives a total volume of ©ˆ]

and °
with
œ;œ
â oçèâ
disturbance step simulations on the nonlinear model,
are needed for the top and bottom product tanks. This
, considerably larger than the required feed tank of Ndœ .
In conclusion, an upstream feed tank with a P controller (averaging level control) proves
best for this example. The example also illustrates that for nonlinear processes the buffer tank
design methods that we have proposed are most reliable for the design of upstream buffer tanks.

20
 1 8

0 6

−1 4

−2 2

−3 0

−4 −2
0 50 100 150 200 250 300 350 400 450 500 0 50 100 150 200 250 300 350 400 450 500
Time [min.] Time [min.]

±@²
(a) Output . Nonlinear simulation (solid) ±©³
(b) Output . Nonlinear simulation (solid)
and linear simulation (dashed). and linear simulation (dashed).

Figure 14: Distillation example with no buffer tanks installed. The control system is not able
to handle the disturbance. There is a large deviation between nonlinear and linear simulation.

0 4

−1 3

−2 2

−3 1

−4 0
0 50 100 150 200 250 300 350 400 450 500 0 50 100 150 200 250 300 350 400 450 500
Time [min.] Time [min.]

±@²
(a) Output . Nonlinear simulation (solid) ±©³
(b) Output . Nonlinear simulation (solid)
and linear simulation (dashed).
ž
and linear simulation (dashed).

Figure 15: Distillation example with a feed tank of

the nonlinear simulation is close to the linear one.
N;œ´ . Both outputs stay within ™ â , and

0 4

−1 3

−2 2

−3 1

−4 0
0 50 100 150 200 250 300 350 400 450 500 0 50 100 150 200 250 300 350 400 450 500
Time [min.] Time [min.]

±@²
(a) Output . Nonlinear simulation (solid) ±©³
(b) Output . Nonlinear simulation (solid)
and linear simulation (dashed). and linear simulation (dashed).
ž
ž
Figure 16: Distillation example with product tanks at the top ( MOy@´
Ž
( y ´ ). The outputs deviate from ™ in the nonlinear simulations. â ââ ) and at the bottom

21
 For (highly) nonlinear processes, the results should, if possible, be checked with simulations
on a nonlinear model.

6 Conclusions
The controlled variables ( ) must be kept within certain limits despite disturbances ( ‘ ) enter-
ing the process. High-frequency components of disturbances are dampened by the process
itself, while low-frequency components, e.g., the long-term effect of a step, are handled by the
control system. There are, however, always limitations in how quickly a control system can
react, for example, as a result of delays. Thus, for some processes there is a frequency range
where the original process and the controller do not dampen the disturbance sufficiently. In
this paper we introduce methods for designing buffer tanks based on this insight. The methods
¶
Ú éë ãØì ÛåÜ éë ãØìÞ ßUà$éë ãØì Ú é)ë ãØìÛ
â ã
consist of two steps:

Step 1. Find the required transfer function such that µA

(with scaled variables). The methods for this have been divided into two groups de-
pending on whether the control system for the process is already designed (methods
S1-S3) or not (methods N1-N3). The shortcut methods (S1/S2 or N1), supplemented

Ú é ì . For a first-order transfer

with nonlinear simulations, are recommended for most practical designs.

function, c
Ú é ìŠçèâ ö é Eâ ì
Step 2. Design a buffer tank that realizes this transfer function
   , we have the following:
I. Quality disturbances Install a mixing tank with volume ç
f ·‚Y , where ‚ is the

é ì°ç
nominal flow rate.
II. Flow-rate disturbances Install a tank with averaging level control with gain 8 
â ö ç
© and volume f 4«>‚ ƒ…c† where 7‚ ƒ‰…† is the expected range (from minimum
Úéì
to maximum) in the flow-rate variation.

Sometimes a higher-order  is preferable, in which case we need (I) for quality dis-

éì
turbances more than one mixing tanks and (II) for flow-rate disturbances a more com-
plicated level controller 8  (with lags) (see Table 1).

References
Buckley, P. S. (1964). Techniques of Process Control. John Wiley & Sons, Inc.. New York.

Faanes, A. and S. Skogestad (2002). pH-neutralization: Integrated process and control design.
Submitted to Comput. Chem. Engng.

Harriott, P. (1964). Process Control. McGraw-Hill. New York.

Hiester, A. C., S. S. Melsheimer and E. F. Vogel (1987). Optimum size and location of surge
capacity in continuous chemical processes. AIChe Annual Meet, Nov. 15-20, 1987, pa-
per86c.

Lieberman, N. P. (1983). Process Design for Reliable Operations. Gulf Publishing Company.
Houston.

22
Ludwig, E. E. (1977). Applied Process Design for Chemical and Petrochemical Plants. Vol. 1.
Gulf Publishing Company. Houston.

Marlin, T. E. (1995). Process Control. Designing Processes and Control Systems for Dynamic
Performance. McGraw-Hill, Inc.. New York.

McDonald, K. A., T. J. McAvoy and A. Tits (1986). Optimal averaging level control. AIChE
J. 32(1), 75–86.

McMillan, G. K. (1984). pH Control. Instrument Society of America. Research Triangle Park,

NC, USA.

Sandler, H. J. and E. T. Luckiewicz (1987). Practical Process Engineering. McGraw-Hill Book

Company. New York.

Schei, T. S. (1994). Automatic tuning of PID controllers based on transfer function estimation.
Automatica 30(12), 1983–1989.

Shinskey, F. G. (1973). pH and pIon Control in Process and Waste Streams. John Wiley &
Sons. New York.

Shinskey, F. G. (1996). Process Control Systems - Application, Design, and Tuning, 4th Ed..
McGraw-Hill Inc., New York.

Shunta, J. P. and W. Fehervari (1976). Nonlinear control of liquid level. Instrum. Technol.
pp. 43–48.

Sigales, B. (1975). How to design reflux drums. Chem. Eng. 82(5), 157–160.

Skogestad, S. (2003). Simple analytic rules for model reduction and PID controller tuning. J.
Proc. Contr. 13(4), 291–309.

Skogestad, S. and I. Postlethwaite (1996). Multivariable Feedback Control. John Wiley &
Sons. Chichester, New York.

Ulrich, G. D. (1984). A Guide to Chemical Engingeering Process Design and Economics. John
Wiley & Sons. New York.

Walas, S. M. (1987). Rules of thumb, selecting and designing equipment. Chem. Eng.
94(4), 75–81.

Walsh, S. (1993). Integrated Design of Chemical Waste Water Treatment Systems. PhD thesis.
Imperial College, UK.

Watkins, R. N. (1967). Sizing separators and accumulators. Hydrocarbon Processing

46(11), 253–256.

Wells, G. L. (1986). The Art of Chemical Process Design. Elsevier. Amsterdam.

Zheng, A. and R. V. Mahajanam (1999). A quantitative controllability index. Ind. Eng. Chem.
Res. 38, 999–1006.

23