Predicting BOD5 exist of domestic wastewater treatment plan by using artificial neural network

Saving water as a very precious capital and resource is very important. Iran, being situated in a dry and semi-dry region and due to water supply shortages in one hand, and for its in creasing population in other hand, will encounter water halter, and recycling sewagesystems and returning it into hydrological cycle is of high significance. Using mathematical models, similar patterns and computer programs are expected to be issues of predicting. But in issues pertaining to neural resources, to choose the models to be closed to the river basin as for as possible would be and is a difficult case. In recent years, there has been a great tendency to use and apply artificial networks in anticipating hydrological models. These networks are parallel processors which save and transfer the input data to network structure during learning. In this study, MATLAB software has been used for predicting BOD5 exite of wastewater treatment plan in Miyandowab city and BOD5and discharge input data. The used network was of Multi Layer Perceptron with Back Propugation learning rule. The obtained results of used model, according to (R square) and (R mse) error coefficient indicate high accuracty of Neural Network in predicting BOD5 exit of domestic wastewater.

Keywords:

BOD5 , DomesticWastewaterTreatmentPlan , ArtificialNeuralNetwork , MultiLayerPerceptron , BackPropagation