Intelligent control of chemical processes using neural networks and fyzzy systems

Abstract

In Chapter 2, neural network architectures with supervised and unsupervised learning paradigms for identification of nonlinear chemical processes are proposed. The multi-layer perceptron model and the Kohonen``s Self-Organizing Feature Map are used as encoders for learining nonlinear functional mapping relationships among process states and manipulative actions. Especially, the unsupervised competitive learning paradigm is investigated and applied to training the Kohonen network as a process identifier. The estimation performances of the proposed unsupervised learning neuro-identifier network and that of a conventional multi-layer perceptron network are compared in identification of a nonlinear continuous stirred tank reactor system. The proposed neuro-identifier based on the self-organizing feature map shows the capability of mapping a huge input data space into a reduced output space in a shorter training time than required by the multi-layer perceptron network. In Chapter 3, a neural model predictive control strategy combining the neural network for plant identification and the nonlinear programming algorithm for solving nonlinear control problems is presented. The proposed controller combines the nonlinear model predictive control approach and the robust multi-step neural model predictor network. A constrained nonlinear optimization using successive quadratic programming combined with the neural identification network is used to generate the optimum control law for complex continuous chemical reactor systems that have inherent nonlinear dynamics. The neural model predictive controller (NMPC) shows good performances and robustness. In Chapter 4, a fuzzy dynamicl learning controller is proposed and applied to control of time delayed, non-linear and unstable chemical processes. The proposed fuzzy dynamic learning controller can self-adjust its fuzzy control rules using the external dynamic information from the process during on-line control and it can create t...