Multi-objective optimization and energy loss characteristics analysis of multistage centrifugal pump as turbine based on response surface methodology and MOPSO algorithm

As a key equipment for high-pressure fluid energy recovery, the operational performance of a multistage centrifugal pump as turbine directly affects industrial energy-saving benefits. To address operational limitations such as low average efficiency, narrow high-efficiency ranges, and uneven inter-stage energy distribution, a cooperative optimization method based on the response surface method and multi-objective particle swarm optimization is proposed in this paper. This methodology enhances system performance by simultaneously optimizing the impeller and guide vane geometries. Based on entropy generation theory, the energy loss mechanisms before and after optimization are investigated to reveal the reasons for performance improvement. Results show that at the design flow rate, the efficiency increased by 9.84 percentage points, shaft power rose by 1.47 kW, and the head decreased by 36.25 m, with a significantly wider high-efficiency zone. The optimized guide vane mitigates inlet vortex formation, creating more uniform flow, while the impeller optimization improves flow stability and alleviates the jet-wake phenomenon at the blade trailing edge.