非线性系统的改进BP神经网络状态观测器设计

A State Observer Design of Improved BP Neural Network in Nonlinear System

针对活性污泥污水处理非线性系统中,由于某些状态变量不能在线检测时状态观测器的设计问题,本文采用了一种基于改进激励函数的变步长BP神经网络状态观测器设计方法,设计了改进BP神经网络状态观测器,并进行了容错性实验以验证该观测器的实用性和可靠性。

While nonlinear state observer is designed, bat there still exists the problem that some state variable can not be detected bylinear checking in treating activated sludge and sewage process. A state observer design method of varied step BP neural network based on improved activation function is applied. An improved BP neural network state observer is designed, and fault tolerance experiment is done to verify its practicability and reliability.