Research On 2D Iterative Learning Predictive Control For The Batch Processes With Actuator Failures

As one of the modern production methods,the batch processes occupies a dominant position in the industry.Research on the optimization control of the batch processes has also achieved gratifying results.With the increasing high requirements and complex processes,uncertainty,actuator failures,Its own non-linear,multi-phase characteristics have become an obstacle to the stable and high-precision production of batch production processes.Therefore,efficient control methods need to be solved urgently.The research work of this paper is as follows:Aiming at a multi-phase batch processes with actuator faults,an infinite time-domain LQ tracking fault-tolerant control method for an batch processes based on a 2D switching system model is proposed.The main work is to introduce new variables based on the established system model to transform a time-delayed system into a system without time-delay,design an iterative learning control law,establish a 2D-Roesser model,and convert it into an equivalent 2D switching system.The quadratic performance function is introduced to design a controller that can resist partial faults of the actuator,while satisfying the optimal control performance.The fault is regarded as interference,and the range of interference allowed by the controller is solved by using Lyapunov stability theory,and the minimum running time of each phase is obtained from it.Taking the injection molding process as an example,the comparison with the one-dimensional model method proves that the proposed two-dimensional method is more effective.For the batch processes with actuator failures and nonlinearities,the fuzzy model and iterative learning predictive control strategy were studied.Intermittent process of 2D T-S fuzzy model,and the prediction model of the fuzzy fault tolerant predictive controller design 2D constraints.According to the design of an optimal performance index and 2D system Lyapunov stability theory,the guarantee system of real-time online design robust asymptotic stability of update law through the form of linear matrix inequality(LMI)constraints are given.Finally,in view of the three water simulation with nonlinear,proves that the proposed method is effective and practical.Aiming at the nonlinear batch processes with partial actuator fault and unknown disturbances,a switching strategy-based a synthetic minmax optimization design of 2D H? model predictive fault tolerant control for piecewise affine batch processes is proposed.Main work: considering the nonlinear system model of the batch processes as multiple linear subsystem models.Based on this,introduce state and output tracking errors,and design an iterative learning control law,which will be converted into an equivalent 2D multiple subsystem model to obtain Corresponding 2D switching system model,and use the recursive method to get the system's prediction form.A maximum and minimum optimized model predictive controller is designed.This controller can not only improve the performance,but also the robustness.The internal disturbances are regarded as disturbances,which solves a certain range of disturbances allowed by the controller through Lyapunov stability theory.From this,the minimum running time of each subsystem is obtained.Finally,the example of a nonlinear stirred tank reactor is compared with the one-dimensional model method,which shows that the two-dimensional method proposed is more practical and effective.