Performance Estimation of Cooling Towers Using Adaptive Neuro-Fuzzy Inference

This paper describes an application of adaptive neuro-fuzzy inference (ANFI) to predict the performance of a cooling tower. In order to gather data for training and testing the proposed ANFI model, an experimental cooling tower was operated at steady state conditions. Utilizing some experimental data for training, an ANFI model based on a standard back propagation algorithm was developed. The performance of the ANFI predictions was tested using data not employed in the training process. The predictions usually agreed well with the experimental values with the coefficients of multiple determinations in the range of 0.995-0.9999, and mean relative errors in the range of 0.69%-3.74%. The ANFI approach shows high accuracy and reliability for predicting the performance of cooling towers.