The COD predictive technique based on neural network

A new online COD predictive technique is proposed in this paper for sewage treatment plants. The technique utilizes the BP neural network method and Elman neural network method to build a model and adopts the real operating data of a chemical industry to establish the model for training and simulation. The results of the simulations indicate that the process variables can be achieved through the establishment of the network model and a reasonable choice of the auxiliary input variable in the complex systems of online prediction.