EG55P7 - Tutorial 9

Control systems design

Consider the reactor shown in the figure below, where the exothermic reaction A + B → C takes
place. The diagram shows the control of the temperature in the reactor by manipulating the cooling
water valve.

a) Design a control scheme to control the flow of reactants to the reactor. The flows of reactants A
and B should enter the reactor at a certain ratio, R, that is, R = FB/FA. Both flows can be measured
and controlled.

b) Operating experience has shown that the inlet cooling water temperature varies somewhat.
Because of the lags in the system this disturbance usually results in cycling of the temperature in the
reactor. Design a control scheme that can help in improving the temperature control.

c) Operating experience has also shown that under some infrequent conditions the cooling system
does not provide enough cooling. In this case the only way to control the temperature is by reducing
the flow of reactants. Design a control scheme to do this automatically. The scheme must be such
that when the cooling capacity returns to normal the scheme of part (b) is re-established.

d) Develop the DSC computer program of the developed control scheme.

(b) A common procedure to design control schemes is to first think how the process would be
controlled manually. In the case at hand, it would be useful to be able to find out as soon as possible
a change in cooling water temperature is occurring so that action can be taken to negate its effect.
For example, if the cooling water temperature increases then the valve can be opened to feed in
more fresh water. The figure below shows this idea.

However, the above scheme does not consider the temperature controller TC17 at all. Instead the
output of TC17 can be used as the set point in a cascade control scheme as shown in the figure
below.
Next step it to automate the configuration, replacing the manual operator with controller, as shown
below.

The figure above shows that the master controller, TC17, looks at the temperature in the reactor,
compares it to its set point and decides on the set point to the slave controller. That is, the master
𝑠𝑠𝑠𝑠𝑠𝑠
controller decides on the required inlet water temperature, 𝑇𝑇𝐶𝐶𝐶𝐶 . Suppose that the inlet water
𝑠𝑠𝑠𝑠𝑠𝑠
temperature is not equal to the set point, for example, 𝑇𝑇𝐶𝐶𝐶𝐶 > 𝑇𝑇𝐶𝐶𝐶𝐶 . In this case the controller would
𝑠𝑠𝑠𝑠𝑠𝑠
open the valve to add more water. This action, though would not make 𝑇𝑇𝐶𝐶𝐶𝐶 = 𝑇𝑇𝐶𝐶𝐶𝐶 . Opening or
closing the valve does not have any effect on 𝑇𝑇𝐶𝐶𝐶𝐶 . The controller would keep opening the valve until
it winds up (reaches a 100% open position). This is a perfect example where the action taken by the
controller does not affect its measurement. The three operations -measurement, decision, and
action-must always be in a loop. That is, the action taken by the controller must affect its
measurement. The scheme showed above does not provide a "closed-loop" but rather, we have an
"open-loop."
Going through the same analysis as previously it can be concluded that the scheme shown below is
also not valid. That is, this scheme still provides an "open-loop." Opening or closing the valve does
not affect the temperature where it is measured.

The earliest that a change in cooling water can be detected and have a closed-loop is in any place in
the recycle line or in the cooling jacket. The below figures show the transmitter installed in the
recycle line and in the jacket, respectively.
(c) For this part you again the manual approach is considered first. As soon as the cooling system
does not provide enough cooling the flow of reactants to the reactor must be reduced. If the
temperature in the reactor, or in the jacket, reaches a high value then this is an indication that the
cooling system may not be providing the required cooling. Further analysis suggests that the best
indication of the cooling capacity is the opening of the cooling valve. When this valve is fully open no
more cooling is possible. At that time the temperature controller cannot do any more, and the
process is out of control. The figure below shows a manual controller configuration.

In order to automate this manual control configuration, an override control scheme is used, as
shown below. The output signal from TC18 that goes to the cooling valve is also used as the
measurement to a controller which is called VPC20.
This controller compares the measurement, which indicates the valve position, with the set point
and sends a signal to the flow controller FC15. VPC20 replaces the manual action of the previous
figure. Note that before the signal from VPC20 gets to FC15 it goes through a selector. The selector is
used to select which signal, the operator set or the one from VPC20, really goes to FC15.

Under normal condition the cooling valve is less than 90 % open, e.g. 65 % open. As VPC20 “sees”
that this valve is 65 % open it decides that the only way to make it open up to 90 % is by asking for a
lot of reactants and for it to increase the output signal, e.g. up to 100 %, to increase the set point to
FC15. Obviously, under this condition there is plenty of cooling capacity left and there is no need to
change the reactants flow required by the operator.

The selector must be such that it selects the signal from the operator and not from VPC20. Because
the signal from VPC20 is probably 100 %, the selector must be a low selector. This is shown in the
figure below. The selector is essentially used by VPC20 to tell the operator set point "move away"
and let me set the set point. So, when the valve opening of the cooling valve (as regulated from
TC18) becomes larger than 90% the controller VPC20 will attempt to reduce this value to the set
point of 90%. It will do so by requesting a lower amount of reactants, thus overriding the set point to
FA at some point.
Some additional things that need to be discussed are the following. The FC15 controller must be set
to remote set point, so that it “listens” for a signal from LS19. However, if the operator sets FC15 to
local set point, or manual, that is if VPC20 asks for a lower set point but FC15 is not in remote set
point to receive it, then nothing will happen. This further means that the safety provided by VPC20
would not work, being essentially not active. This is a highly unsafe operation, so the control system
needs to be designed so that it allows the operator to set FC15 in local set point or manual, and at
the same time allowing VPC20 to be able to reduce the flow of reactants. The figure below shows
the new control scheme that accounts for this.
In this scheme the output of VPC20 goes directly to the valve, and not to the set point of FC15. So no
matter what the mode of FC15 is, the decision of closing the valve is after the controller. In this case
the selector is also a low selector. If for safety considerations it is necessary to manipulate a flow, it
is always good practice to go directly to the valve and not to the set point of a controller
manipulating the valve. Note that the reset feedback as usual goes to the controllers feeding the
selector.

Next, the set point of VPC20 needs to be addressed. The controller receives the output signal from
TC18 to decide whether the valve is 90 % open or not. However, the cooling water valve needs to be
fail-open (FO), therefore, the valve is 90 % open when the signal is 10 %. Thus, the real set point in
VPC20 must be 10 %, as shown in the figure below.

(d) Below the software program describing the control scheme above is presented.