基于NW型小世界人工神经网络的污水出水水质预测

为了预测污水处理出水水质，针对污水处理过程具有多变量、非线性、时变性、严重滞后的特点，提出了基于NW型小世界人工神经网络的污水处理出水水质预测模型。首先根据污水处理系统确定模型输入输出变量个数，然后建立了多层前向小世界神经网络模型，并对网络模型的隐层结构进行了优化研究。借助污水处理过程的历史数据进行了仿真研究，结果表明：和同规模的多层前向人工神经网络相比，小世界神经网络对污水出水水质预测具有较高精度和收敛速度，为污水出水水质的实时预测提供了一种有效的新方法。

Effluent quality prediction of waste water treatment plant based on NW multi-layer forward small world artificial neural networks

A NW multi-layer forward small world artificial neural networks soft sensing model is proposed for the waste water treatment processes, regarding the characteristics of multivariable, nonlinear, Time-varying and Time lag in the treatment process. The input and output variables of the network model were determined according to the waste water treatment system. The multi-layer forward small world artificial neural networks model was builded, and the hidden layer structure of the network model were studied. The waste water treatment process experiments and the training and simulation of the soft sensing model based on the experimental data were conducted, and the results indicate that the soft sensing model proposed has good predictive effect for the water quality. An effective method for water quality prediction was provided.