Nyquist Array Design Of Multivariable Control System Based On Genetic Algorithm Structural Optimization

Chemical processes are usually high-dimensional multi-variable control systems,and the internal coupling cannot be ignored when designing the decentralized PID control system.For a high coupling multi-variable system,a decoupler is required to weaken the coupling degree,and then decentralized multi-loop PID controllers are applied to the decoupled system.The Nyquist array design method realizes the decoupling through diagonal dominance in frequency domain,but the designed decoupler is often complicated.In order to simplify the structure but maintain the performance of the decoupler,the Nyquist array design method based on genetic algorithm structure optimization is proposed to design the multi-variable control system.Based on genetic algorithm,the two-dimensional genetic operator in the form of 0-1 structure matrix is designed to describe the decoupling structure.Based on the structure matrix,the PID parameters of decouplers are calculated using the sparse Nyquist array design method.The design difficulty and decoupling performance of the control system are measured by the simplicity and decoupling degree,which are combined into a comprehensive performance index as the fitness function of the genetic algorithm.The structure and parameters of the decouplers are optimized by the genetic algorithm to weaken the coupling of the multi-variable system.Multi-variable systems often have different degrees of internal coupling severity.Usually,some parts are seriously correlated with each other,while the other parts are less correlated,which can be divided into multiple subsystems.The internal coupling of subsystems is serious,and the subsystems are almost independent of each other.Therefore,the decoupler is divided into several independent block decouplers,and each block decoupler decouples each subsystem,so that the structure of the decoupler is simplified and easy to implement.Genetic algorithm based on combined chromosome is presented.The chromosomes which describe the row sequence,the column sequence and the block form are introduced,whose corresponding crossover and mutation rules are also designed.After the block decoupling structure matrix is produced from the combined chromosome,the block decoupling Nyquist array design method is used to calculate the PID parameters of decouplers.The optimal block decoupler is obtained by genetic algorithm.After the decoupler design,suitable PID control loops are designed according to the column dominance degree of the decoupled system,so that the multi-variable control system can achieve good control effect.