Simulation Of Flow-Acoustic Characteristics And Optimization Design For Noise Reduction Of Counter-Rotating Axial Flow Fan

Compared to single-blade radial flow fans with the same radial dimensions,counterrotating axial flow fans can provide higher static pressure and stronger airflow,thereby enhancing heat dissipation.They have been widely employed in electronic device cooling applications.However,the associated noise pollution during operation poses severe challenges,affecting the reliability of electronic components and the well-being of personnel.Consequently,it is of paramount importance to optimize the noise reduction of such fans without compromising their aerodynamic performance.This study focuses on a counter-rotating axial flow fan used in servers,investigating its flow field and acoustic characteristics through numerical simulations.Additionally,noise reduction optimization designs for the fan are conducted.The research in this paper is structured as follows:(1)Research on numerical simulation methods for the flow field and acoustic characteristics of the fan.The simulation grid is established through model simplification,computational fluid domain creation,and grid generation.The simulation process for obtaining the P-Q curve and noise values of the counter-rotating axial flow fan is determined.The analysis and selection of turbulence models,solvers,and boundary conditions are performed.The credibility of the simulation method is verified by comparing the simulation results with experimental data.(2)Study on the influence of single-factor variations of the fan on the flow field and acoustic characteristics.The flow field and acoustic characteristics of the prototype fan are analyzed to identify the significant factors contributing to noise.The effects of blade quantity and serrated trailing edge on the flow-acoustic characteristics are explored.The mechanisms behind these effects are analyzed,and the variations in the P-Q curve and noise values are summarized.(3)Research on a noise reduction optimization method based on surrogate models of blade parameters.The optimization of blade parameters,specifically static blade bending features and installation angles,is achieved through blade parameterization.The optimized Latin hypercube sampling method is employed to determine random sampling points,and a task submission service program for high-performance computing is designed to enable parallel simulation tasks for a large number of sample points.RBF neural networks are utilized to fit the sample data,constructing surrogate models for the P-Q curve and noise values.A multi-objective optimization problem for fan noise reduction,based on the surrogate models,is formulated.The genetic algorithm is employed to solve this problem,and solutions from the non-dominated set are selected as optimization results that exhibit consistent performance in the P-Q curve with the prototype fan while reducing noise.Finally,the noise reduction mechanism of the optimized fan is analyzed,and the results are validated through physical testing and comparison.