Prediction of relative efficiency reduction of centrifugal slurry pumps: empirical- and artificial-neural network-based methods

Abstract

This paper has focused on the predictive methods for relative efficiency reduction of centrifugal pumps handling slurries based on empirical- and artificial-neural network (ANN) approaches. A new correlation has been developed to predict the relative efficiency reduction of the centrifugal slurry pumps, and the range of validity of the present correlation has been verified using the data available in the literature. Then, the applicability of ANNs for the same purpose has been investigated using a total of 315 data. The comparisons of both methods showed that the present correlation produced the lowest deviation among some recent correlations in the literature, and, if properly constructed, ANNs could be used as a predictive tool with higher accuracy than the conventional empirical methods.