Design And Evaluation Of PID Controller For Batch Process

The design and evaluation of PID controllers has been studied in this thesis from the perspective of within batch and between batches.Firstly, the principle of the PID control has been discussed. Then a brief introduction of the PID controller design for two special process models was presented, which were based on the internal model control and the direct synthesis method. A simulation analysis was given to illustrate the major features of the two controllers.The conventional IMC-PID algorithm could not meet the requirements of fast transient response and small overshoot simultaneously. An improved method was proposed where two parameters ? and ? were introduced to determine the value of the IMC filter parameter ?, by combining the square of output error with the two introduced parameters. When there exist the mismatch between the model and the real process, a system identification based on the open-loop unit step response was employed. The input/output data during the transient process were used to obtain a First Order Plus Dead Time (FOPDT) or a Second Order Plus Dead Time (SOPDT) process model. The parameters of PID controller can be obtained based on the model. The simulation results illustrate that the proposed method can meet the requirements of fast transient response and small overshoot simultaneously.As for the problem of performance evaluation for PID controller within a batch run, the deterministic performance indexes of the PID controllers based on DS-d and IMC methods were studied. The performance indexes include the value of the Integral Absolute Error (IAE) and the Integral Time Absolute Error (ITAE). The applicability of the results was verified by the simulation. When the IMC-PID controller tracking performance was constrained by the actuator saturation, a series of expressions of the relationship between the value of the actuator saturation and the controller parameters were derived. The optimal IMC-PID controller parameters can be calculated by considering the actuator saturation explicitly. An optimal value of IAE was obtained to predict the performance of the controller. The simulation results show that the proposed method is suitable to evaluate the performance of the system controller under saturation constraints and the lower bound value of IAE is verified.Finally, this thesis introduced the iterative learning control theory to take advantage of the repeatable characteristic on multiple batches. The simulation results show that the real time tracking of the desired output signal was achieved.