活性污泥过程MISO神经网络模型
Neural Network MISO Model for the Activated Sludge Process
-
摘要: 用误差反向传递 ( BP)神经网络模拟了活性污泥过程 ,发现使用多输入多输出 ( MIMO)神经网络来模拟活性污泥过程变量间关系时 ,线性部分误差较小 ,非线性部分误差较大 ;但若用数个多输入单输出 ( MISO)神经网络的组合来模拟时 ,则对活性污泥过程的线性或非线性关系均能很好模拟Abstract: The activated sludge process was simulated by using BP neural network. It is found that errors occur in the non linear parts of the ASP when using MIMO neural network, and that these errors can be reduced when the MIMO neural network is replaced by a group of MISO neural networks.