Conjugate Gradient Method for Evaluating the Anaerobic Wastewater Treatment System in the Prediction of COD

An anaerobic reactor predicts the level of COD by applying a feed forward Neural Network. The organic effluent COD is the most common ecological and process performance indicator for all types of Wastewater Treatment Systems. In this experiment, the prediction is carried out by training the three layered feed forward ANN using Conjugate gradient algorithm. The trial sets are subjected to recognizeand reduce the missing values and outliers using K-Means clustering method and are used for training the input and output data. Thus the assessment results for the prediction of COD using Conjugate gradient function were found satisfactory.