Aerodynamic Noise Performance Analysis And Optimization Design Of Low Specific Speed Centrifugal Impeller

The low specific speed centrifugal compressor has the characteristics of compact structure,fast response,and good adaptability at low speed.It is the core component of the fuel cell system that improves power density and efficiency through air boosting.With the rapid development of hydrogen energy and fuel cell electric vehicle technology in recent years,the design of low specific speed centrifugal impellers faces urgent needs such as high efficiency,low noise,stability and reliability.The multi-objective optimization design of centrifugal compressor impellers has become a hot issue at present.The compressor itself affected greatly by geometric errors consumes a lot of power which is the largest noise source of fuel cell electric vehicles.Involving multiple design parameters and target responses,and the complex threedimensional flow,the rotor design belongs to a high-dimensional nonlinear optimization problem.Moreover,the relevant design knowledge and experience are relatively lacking.The impact mechanism of design parameters on performance needs to be further explored,which brings challenges to the current optimization design.Aiming at solving the above-mentioned problems,taking a low specific speed centrifugal compressor impeller as the research object,based on the three dimensional numerical simulation,the BP neural network,the sparse grid method and the multi-objective genetic algorithm,this paper investigates the optimization design of centrifugal impeller aerodynamic performance and noise.The self-organizing mapping method is employed to perform deep data mining on data of the optimization design process,the optimization design rules and decision-making knowledge is explored,and a multiobjective optimization method for low-specific speed centrifugal compressor impellers is formed,the multi-objective trade-off relationships,the correlations and redundancy of design parameters are obtained.It provides reference for the optimal design of centrifugal impeller at low specific speed.The main research contents of this paper are as follows:1.A three-dimensional flow field and sound field numerical analysis model of the centrifugal impeller was established,and the correctness of the model method was verified based on the LSCC impeller case experiment.The influences of various design parameters such as tip clearance,inlet angle,outlet angle,and outlet width on centrifugal force with low specific speed were explored.The influence of compressor impeller pressure ratio,efficiency and noise performance,and the influence mechanism of each design parameter on the impeller performance was analyzed.Based on the established analysis model and the Latin hypercube sampling method,the design parameter values and pressure ratio,efficiency and noise response values were established The Pearson coefficient was used to quantify the correlation between the geometric parameters and the performance response,and the order of the correlation strength was obtained,so as to select the strong correlation variables as subsequently optimization design variables.2.Based on BP neural network,a prediction model for aerodynamic noise performance of centrifugal impeller was constructed,and the prediction results of the neural network were compared with the numerical simulation results of the impeller flow field to verify the accuracy of the prediction model.Based on the optimization mathematical model,combined with the multi-objective genetic algorithm,the Pareto optimization frontier solution was obtained,and the optimization results were verified based on the three-dimensional flow field analysis.Using the self-organizing mapping method,a data mining on the Pareto frontier solution set generated during the optimization process was perform,and the impeller multi-objective trade-off relationship,the objective mechanisms of influence with design variables were obtained.3.Considering the influence of random errors of impeller geometric parameters on the aerodynamic robustness of centrifugal impellers,taking the pressure ratio,efficiency and standard deviation as optimization objectives,and taking noise as constraints,based on numerical simulation of flow field and sound field,combined with the sparse grid method,the BP neural network and the multi-objective genetic algorithm,the aerodynamic robust optimization solutions were obtained.Using self-organizing mapping method,the data mining of this optimization solution set was carried out,which provides reference for the multiobjective design of low specific speed centrifugal impeller.