Hydraulic components’ matching optimization design and entropy production analysis in a large vertical centrifugal pump

The large vertical centrifugal pump is the core power equipment for long-distance water division and large-scale irrigation projects. The power of the matching motor can reach 40 MW. In view of reducing the operating energy consumption of this kind of pump, the efficiency under the design condition was taken as the optimization objective, and a matching optimization on hydraulic components was proposed in this research. The optimization process was divided into two stages. The first stage focused on improving the configuration of the vane diffuser. In the second stage, the Plackett-Burman test design was used to screen out the optimization design variables of the vane diffuser and the volute, the optimal Latin hypercube sampling method was adopted to generate 106 sample cases, and an automatic numerical simulation optimization platform was built. Then, different approximate models were employed to construct the relationship between the optimization design variables and the objective function, and their fitting accuracy and robustness were compared. Finally, the optimal design parameters were determined by particle swarm optimization, and entropy production theory was used to analyze the internal flow pattern of the model before and after optimization. Results showed that D 3 , b 4 , β 3 , and S 8 have the greatest impact on the hydraulic efficiency of the pump. The multilayer backpropagation neural network has a higher fitting accuracy and better robustness compared with the other three approximate models. The efficiency of the optimized model under the design condition is increased by 4.22 %, reaching 90.82 %. Reducing the number of layers and vanes of the diffuser and improving the matching of hydraulic components can dramatically improve the hydraulic performance and hump characteristics of large vertical centrifugal pumps. Entropy production theory is a reliable approach to visualizing flow loss. The turbulent flow dissipation in the vane diffuser can be reduced the most after optimization.