SELECTING THE CONTROL CONFIGURATION FOR A NONLINEAR SYSTEM BY ON-LINE CONTROLLERS’ PARAMETERS OPTIMIZATION

Abstract

AbstractThe multiobjective nature of the most real engineering control systems requirements and how to determine the best control strategy for a specific problem motivate the development of optimization-based design frameworks and computer aided control system design methods to search for acceptable designs with low computational load. It is proposed an approach to find in real time the best control strategy by optimizing the controllers’ parameters of three control strategies for the various operating points of a non linear system. The proportional-integral-derivative (PID) controller, the coprime factor uncertainty using H-infinity optimization and the mixed sensitivity H-infinity control strategies are used for this purpose. The genetic algorithm is used to tune and optimize the controllers’ parameters, penalizing the individuals of the control strategies that do not satisfy the problem's given performance, and then defining the best control strategy for each operating point of the system. Designing controllers as presented offers the advantage of achieving the plant best working conditions in real time.