A Review on Artificial Neural Network Approach for Predicting the Performance of Centrifugal Pump

In this work, centrifugal pump is analysed by using a single-stage end suction centrifugal pump. Two main components of a centrifugal pump are the impeller and the casing. The impeller is a rotating component and the casing is a stationary component. In centrifugal pump, water exits radially, while water enters axially through the impeller eyes. The pump casing is to guide the liquid to the impeller, converts the high velocity kinetic energy of the flow from the impeller discharge into pressure. A mean of centrifugal pump impeller is passed out and analysed to get the best performance point. The design and performance analysis of centrifugal pump impeller are chosen because pump is the most useful mechanical Rotodynamic machine in fluid works which is widely used in domestic, irrigation, industry, large plants and river water pumping system. Also this work aims to model a pump using neural networks and to investigate their performance in predicting its pressure head and shaft power. All the ANN models are trained, validated and tested which predicted the pressure head and shaft power with acceptable accuracy.