Optimal iterative learning control for endpoint product qualities in semi-batch process based on neural network model
A neural network model based optimal iterative learning control (ILC) strategy for improving endpoint product qualities in semibatch processes is proposed in this paper.Control affine feed-forward neural network (CAFNN) is constructed by a special structure and can be used to build nonlinear models of semi-batch processes.In terms of the repetitive nature of semi-batch processes,ILC is used to improve endpoint product qualities from batch to batch.Due to the structure of CAFNN,its gradient of endpoint product qualities with respect to input profile can be computed analytically and a tracking error transition model is built.Therefore,an optimal ILC law with direct error feedback is explicitly obtained by minimizing a quadratic objective function.Sufficient conditions of tracking error convergence are derived for the optimal ILC.It has been proved that the tracking error converges to a small constant but depends on CAFNN model accuracy.The proposed ILC method is illustrated on a simulated isothermal semi-batch reactor.
iterative learning control neural networks semi-batch processes.
Zhihua Xiong Jie Zhang
Department of Automation,Tsinghua University,Beijing 100084,P.R.China School of Chemical Engineering and Advanced Materials,Newcastle University,Newcastle upon Tyne,NE1 7
国际会议
International Conference on Modelling,Identification and Control(模拟、鉴定、控制国际会议)
上海
英文
2008-06-29（万方平台首次上网日期，不代表论文的发表时间）