Double-suction pump impeller optimization and performance analysis based on nonlinear programming quadratic Lagrangian algorithm

Abstract

To reduce the energy consumption of the double-suction pump, this study optimizes the impeller of the pump based on the Nonlinear Programming Quadratic Lagrangian (NLPQL) algorithm in OptiSLang. Taking the weighted efficiency as the optimization objective, 28 design parameters of the impeller are selected as the input variables, 300 groups of sample schemes are generated through the Advanced Latin Hypercubic Sampling method, and an automatic numerical simulation platform is setup to optimize the impeller under different working conditions. A metamodel of optimal prognosis is built by OptiSLang based on sample data and system response, seven variables with high sensitivity are screened out, and the NLPQL algorithm is adopted to carry out the optimization design to obtain the optimized model. The optimized model is, then, verified by the numerical simulation, and it is analyzed and compared with the original model from external characteristic curves and internal flow mechanism. The results show that: the match between the optimized design variables and the optimization objective is good; the weighted efficiency of the optimized pump is increased by 4.3%, the high efficiency zone is obviously enlarged and the shaft power is significantly decreased; the flow lines in the optimized impeller are obviously improved, the vortex is eliminated, and the velocity distribution is more uniform; the low pressure zone at the inlet of the vane is obviously reduced.