基于Markov参数整定的自适应迭代学习PID控制

Parameters of semi-batch processes often varies with time. A single neuron adaptive PID iterative learning control strategy based on Markov parameter tuning was investigated in this study. A two-dimensional iterative learning PID controller (2D-ILC-PID) was first established, and the initial values of the parameters were tuned offline by the Markov parameter method. The controller parameters were adaptively adjusted online by single neuron adaptive adjustment mechanism. The algorithm can make full use of the repeating information between batches and improve the iterative learning rate, and achieve effective improvement of the control performance. The control method was verified by a simulated reaction process, and the results show that the single neuron adaptive iterative learning control method based on Markov parameter tuning can effectively achieve accurate tracking of reaction temperature.