Application of a multi-layered neural network to system identification

A study on the application of the neural network method to estimate chemical oxygen demand (COD) in a wastewater treatment process is introduced after presenting the back propagation algorithm, one of the most typical algorithms of neural networks, in detail. Several types of simulation have been tried in order to select suitable neural network parameters. The results of simulation to estimate real sampling COD data for a specific real-world wastewater treatment process shows that our method is promising.