Chemical reactor control project


Thread Starter

Fardad Zand

I am a university student at Polytechnic inversity
and I am going to do a research for my B.Sc project about controlling a chemical reactor which is used for a Polyester resin synthesis.
There are some facts about our project and they are :

1)The dynamics of the process is known and we know a lot about the process mathematical model and the form of the equation, which is used to model the process, but we don't know the exact values of the model parameters (kinethic, reaction &rate contants and we could just estimate them prior to beginning of the experiment); We could just approximate these values by performing some experiments prior to the main experiments.

2)The process is to some extent sluggish (not so
sensitive). We are going to design a controoler that should follow a predetermined temperature profile at the range of 150 c to 250 c with a maximum of 5 c deviations from the set point.It is preferable to use a control strategy that help us to online monitoring of two process variables(except of the temperature) which guide us to reach high quality products (resins with known molecular weights and -OH saturation values which have special reaction temperature profiles).

Now I have some question from you who are control specialists and they are as follow :

1)What are the best control methods that we could
apply to our system to get the best results? why?

2)what are the things we should know about our system prior to performing the control strategies mentioned above?

If it is possible please answer me as soon as possible because we are seriously concerned to start our work at the university.

Thanks before hand.

Best Regards
1 Dec 2001

Andrew Burden

Hello Fardad,
I am a chemical engineering graduate student at the University of Louisville. I am specializing in control systems. If you could send me a more info of your process. How many inputs, outputs. What are your objectives? Can you obtain
process dynamics or experimental models for a control system to utilize. If the process is nonlinear...not a problem...need optimization?...we can do that too. Just let me know.
Andrew Burden
Chemical Engineering
University of Louisville
[email protected]

david mertens


the way to control a reactors pressure and temperature are largely dependent of the type of reactor. The most common control is base on
minimum two temperature measurements, one in the reactor, measuring the product temperature and one in the jacket, measuring the heating agent (mostly steam or hot oil). When heating up the jacket temperature will be controlled to a setpoint that is a given delta T above the
product temperature. This ensures a heating with constant heat flow to the product. Once a certain temperature is reached the control switches to cascaded control where the valve is controlled by the jacket temperature controller with a setpoint that is the result of the actual product temperature controller. A good distribution of heat in the product must be guaranteed, therefore the reactor contents has to be recycled or stirred. For pressure control a standard PID controller is often used with split-range (two valves, one supplying nitrogen, the other to atmosphere (via scrubber/flare) or to vacuum unit).
However this is only a description of most common situation.