The use of artificial neural networks for prediction of discharge capacity of a spillway with a breast wall

Abstract

     The accuracy of discharge estimation is very important from an operational, environmental and economic point of view in irrigation or hydroelectric projects. This accuracy affects the safety in the activities and operation of the entire work systems. This paper describes the application of Artificial Neural Networks (ANNs) and Adaptive Network Fuzzy Inference System (ANFIS) models to predict the discharge capacity of a breast wall spillway. The performance of these models is compared to the conventional non-linear regression (NLR) and multi-linear regression (MLR) models based on experimental data. Root mean square errors ( ), average error ( ), average absolute deviation ( ) and correlation coefficient ( ) statistical parameters are used as comparing criteria for the evaluation of these model's performance. The comparison result indicates that these neural networks could be employed successfully in the discharge prediction of the spillway with a breast wall. The results also show that the performance of the ANFIS and FFBP model are found superior to those of the MLR, NLR and FFCC models with the lowest error and the highest correlation coefficient.