Improved PSO Optimized PID Controller For Paper Manufacturing Quantitative System

Quantitative control system in paper manufacturing process is a high order non-linear system with time delay and complicated interference.In such scenario,conventional PID parameter tuning methods relying heavily on man experiences fail to provide satisfactory control performance.Therefore,an updated PID tuning strategy based on optimization method is needed.PSO algorithm possesses advantages in terms of easy implementation and less parameter tuning requirement,which is ideal for PID design tasks.However,while applied to PID parameter tuning,standard PSO algorithm tends to reside in local optimal because of the convergence rate is too fast at the begining.To conquer this problem,an improved PSO algorithm is proposed and applied to the quantitative control system of a paper machine.The main contributions of this thesis are as follow:1.Convergence rate of the PSO algorithm is expected to be equally distributed during the optimization process thus to avoid early convergence.Therefore,an improved PSO algorithm based on nonlinear weight distributed method is proposed.During the optimization process,an adaptive tuning strategy is adopted by the PSO algorithm.Simulation results are used to verify the effectiveness of the proposed algorithm.2.Tuning the PID parameters via PSO approach requires considerably large computational work,and can be easily affected by the initial value chosen for the optimization algorithm.A modified PSO-PID optimization based on H? theory is proposed to solve such problem.The improved method narrows the feasible solution region for the PSO algorithm and eases computational complexity.The simulation results shows that the improved PSO algorithm manage to enhance the convergence performance,and boost the PID control efficiency.3.The aforementioned method is applied to a factory paper mill in Jiaxing District:a H? PID controller is designed and the parameters were optimized by the improved PSO.Experimental results show that the proposed method can achieve expected performance in practical applications.