Structural Optimization of Multistage Centrifugal Pump via CFD and Machine Learning Method

Abstract To implement energy savings in multistage centrifugal pumps, a return channel is utilized to replace the origin inter-stage flow structure, and then single-objective optimization work containing high-precision numerical simulation, design variable dimensionality reduction, machine learning conducted obtain optimal geometric parameters. The reduction process based on Spearman correlation analysis method. influence of 15 variables impeller investigated, seven them with high-impact factors are selected as final variables. Thereafter, genetic algorithm-backpropagation neural network (GA-BPNN) model used create surrogate high-fitting performance by employing GA optimize initial thresholds weights BPNN. Finally, multi-island algorithm (MIGA) employed maximize hydraulic efficiency under nominal condition. findings demonstrate that optimized model’s increased 4.29% at 1.0Qd, deterioration pump overload conditions effectively eliminated (the maximum increase 14.72% 1.3Qd). Furthermore, internal indicates scheme can improve turbulence kinetic distribution reduce unstable structures pump.