Abstract:
Exothermic reaction process releases energy during mixing of different substances within the
stirred tank reactor. The nature of reacting substances may exhibit different chemical properties
that lead for an exothermic reaction to be occurring. The generated heat on the stirred tank
reactor due to exothermic process should be regulated properly as the system shows non-linear
and uncertain process. In this thesis the design of cascade control of exothermic stirred tank
reactor is only using adaptive self-tuning regulator. The purpose of self-tuning regulator is to
make online parameter estimation for the uncertain plant. It has both ordinary feedback control
loop and parameter estimation block. For parameter estimation recursive least square algorithm
is used to estimate the plant parameters. The control system has two controlled outputs the
concentration of feed stream and temperature of the reactor. The flow rate of water through the
cooling jacket around the reactor is a common manipulated input variable. The controlled
outputs have two disturbance variables like inlet feed temperature and the inlet feed
concentration to the reactor. By considering the stirred tank reactor as ideal and assuming perfect
mixing of the reactant in reacting medium, the volume of reactor will be constant. On the other
hand, the concentration of reactant inside the reactor is also the same with that of the output
concentration. The cascade stirred tank reactor processes presented involves two internal control
loop stages for each controlled outputs. Each control loop has its own slave controller and the
corresponding plant dynamics. For the two outputs of the cascade exothermic stirred tank
reactor, a PI controller was used for controlling the secondary internal loop plant of the system
under the desired condition. Such loops will be considered as secondary control loops for the
each controlled outputs. On the other hand the self-tuning controller is used as a master
controller for each controlled variables. So far a PID controller based designs of cascade CSTR
were made to the linear plant models so as to analyze the control action and the stability of the
system. But the PID controller does not have the ability to estimate the parameters of the plant
exhibiting non-linear characteristics. .Thus self-tuning regulator control technique used in this
thesis allows a more stable output of for the plant having dynamic behavior.
