제약조건을 고려한 불안정 시스템의 RCGA 기반 PID 제어

Abstract

The Proportional Integral Derivative(PID) controller is still widely used in the process industries even though control theory has developed significantly since it was first used several decades ago.

As our industry has developed and increased the level of high-technology, many suggestions have appeared to upgrade the PID controller. Most of them are mainly based on experiences and experimentations. As a result, tuning of PID controller depends on designer's experience and intuition. Closed loop tuning method of Ziegler and Nichols(Z-N), open loop tuning method, Cohen-Coon(C-C) tuning method and IMC tuning method are well known to us.

However, it is not easy to apply for unstable process with time delay. The reason is because of unstability due to the poles existing on right-hand side in s-plane and the effect of time delay. Also, we can ascertain through many of the earlier researches that unstable processes with time delay have large overshoot the cause of system characteristics. The well known tuning methods for unstable process are De Paor and O'Malley method, Venkatashankar and Chikambaram method, Poulin and Pomerleau method, Ho and Xu method, Wen and Yingqin method. But the above methods did not consider the control environment.

This paper considers design technique for PID controller in case of predefining overshoot or settling time by designer according to control environment. To deal with constraint problem like this, the RCGA(Real-Coded Genetic Algorithm) incorporating the penalty strategy is used this is the method that if the RCGA violates given constraints, the defined penalty function is summed to the evaluating function depending on the severity and to convert non-constraints optimization problem. The proposed method is applied to the unstable FOPTD(First Order Plus Time Delay) system and simulation is given to illustrate the set-point tracking performance.