Multi-objective Optimization design of the multiphase pump boosting cell

Based on the boosting cell parameters of the third generation YQH-100 multiphase pump independently developed, the orthogonal experimental design was used to design the optimization scheme, using relative head and efficiency as the evaluation targets for the multi-objective optimization design. The software FLUENT was used to predict relative head and efficiency of the multiphase pump. BP neural networks and PSO-BP neural networks were adopted to construct the response relation between the design variable and the objective function. The predict results show that the prediction performance of PSO-BP neural network was much better than BP neural network. A multi-attribute decision making method based on fuzzy preference information was used to select the best multi-objective solution. The optimization results show that relative head and efficiency of the multiphase pump were increased by 0.81 % and 0.77%, respectively.