精馏过程的神经网络控制系统的设计

Design of Neural Network Control System in Distillation Process

本文以一个三元精馏塔为研究对象，详细介绍了精馏过程的神经网络控制系统（ＮＮＣＳ）的设计过程。本ＮＮＣＳ主要包括ＮＮ预估器和ＮＮ控制器，前者被训练后用来精确表达精馏塔的输入到输出的非线性映射，后者被训练后用来对该ＮＮ预估器实现解耦控制。所采用的神经网络模型为ＢＰ网络。

This paper takes a triple distillation column as the object and introduces the design of neural network control system(NNCS) in distillation process.The NNCS introduces NN preestimater and NN controller.The NN preestimater is used to express the nonlinear relation between the input and the output.The NN controller is used to realize the decouple control on NN preestimater.BP network is used and realized on a parallel system of 9 Transputers.