空间电子技术2016，Vol.13Issue(3)：41-49,9.DOI:10.3969/j.issn.1674-7135.2016.03.008
BP神经网络改进的内模控制方法对积分时滞系统的控制实现
A Back-propagation Neural Network Modified IMC Method for the Control of IPDT Process①
陈昊 1白春江 2王瑶 1段维俊1
作者信息
- 1. 西南民族大学，电气信息工程学院，成都 610041
- 2. 中国空间技术研究院西安分院，空间微波技术国家级重点实验室，西安 710000
- 折叠
摘要
Abstract
In this paper, an innovative control approach is developed to deal with steady state error and process model-ling errors for the control of time delayed process with integral section is developed by using back-propagation neural network approach. Instead of the internal model control ( IMC) , a neural network modified IMC ( NNIMC) is proposed as a novel control scheme. Lyapunov direct method is employed to prove that the convergence of the developed NNIMC is guaranteed by selecting a proper learning rate. Simulation results show that this new development can successfully nullify the caused steady state error for the control of IPDT process in the presence of disturbance load and also present strong robustness to process modelling errors.关键词
神经网络/时滞/内模控制/稳定性Key words
Neural network/Time delay/Internal model control/Stability analysis引用本文复制引用
陈昊,白春江,王瑶,段维俊..BP神经网络改进的内模控制方法对积分时滞系统的控制实现[J].空间电子技术,2016,13(3):41-49,9.
