神经网络在污水处理过程中的应用

Neural Network Application in Sewage Treatment Process

摘要:

针对活性污泥法在污水处理过程中,水量会发生巨大的变化,以及运行所带来的强耦合、非线性、大滞后性等影响了污水处理的控制过程；利用神经网络技术特点对污水处理厂进行模拟实验,建立BP神经网络模型的仿真,用试凑法确定隐含层节点的个数,为了避免建立的网络过大,在训练网络过程中,避免出现网络“过训练”,建立合适的网络模型；通过对污水处理过程中输入数据的水质的变量参数,来预测未来某一时间输出某一水质变量参数；结果表明,BP神经网络可以应用于活性污泥法水处理过程中对水质参数进行模拟仿真和预测的效果。

Abstract:

In activated sludge sewage treatment process, the water will change dramatically, and run strongly coupled, nonlinear, big lag, the establishment of the BP neural network model simulation, using trial and error method to determine the number of hidden layer nodes, avoiding the establishment of too large network. In the process of training the network, the establishment of an appropriate network model to avoid excessive training. Through the input data of water quality in the sewage treatment process of variable parameters, to predict the future output of a water quality in a certain time variable parameters.The results show that, BP neural network can be applied to the activated sludge treatment process simulation and prediction of water quality parameters.