U.S. Banu∗ and G. Uma∗∗


Internal model control, hybrid control, continuous stirred tank reactor, genetic algorithm-based PID control Specifications • Fi/V = 1 l/min • Ti = 298.5 K • UA = 150 kcal/m3 ◦ Kh • Ci = 10 mol/l • Tci = 298.5 K • k0 = 34930800 h−1 • ΔE/R = 11843 kcal/kgmol • −ΔH = 2(105 ) kcal/mol • ρCp = 500 kcal/m3 ◦ K • ρcCpc = 500 kcal/m3 ◦ K ∗ Instrumentation and Control, BSA Crescent Engineering Col- lege, Vandalur, Chennai – 600 048, Tamil Nadu, India; e-mail: sabura_banu@yahoo.co.in ∗


This paper describes the design of a hybrid control for continuous stirred tank reactor (CSTR), with parameter uncertainty and system disturbance by intelligent techniques. CSTR is characterized by nonlinear and time-varying behaviour. Control of CSTR is a challenging problem. The proposed hybrid controller consists of a genetic algorithm (GA) based proportional integral derivative (PID) control and an internal model control (IMC). The IMC is constructed with a backpropagation (BPN) algorithm based neural emulator for the model and an Takagi Sugeno Fuzzy based inverse adaptive neuro-fuzzy inference system (ANFIS) as the controller. The IMC provides a very good tracking control and leads to control performance degradation in the case of regulatory problem which is solved using the GA-based PID control. GA-based PID implements the characteristics of GA’s global optimization to optimize the PID’s control parameters: Kp, Ki , Kd to provide best control effect. The proposed model provides an effective controllable range and gives robust control performance.

Important Links:

Go Back