Blade Parameterization And Aerodynamic Performance Optimization Of Centrifugal Compressor Impellers

Compared with axial compressor,centrifugal compressor has more compact structure,higher single-stage pressure ratio and better operation stability,therefore it is widely used in aerospace,energy and other fields.The impeller is the main power component of centrifugal compressor,and its performance greatly affects the performance of the whole machine.The traditional ’modify-verify-modify’ optimization design method has low efficiency and depends on the experience of the designers.In order to achieve fast and efficient optimization design of centrifugal impeller,this paper studies the parametric modeling method and aerodynamic performance optimization design method of centrifugal blade,mainly including the following parts:1.According to the geometric characteristics of centrifugal blade,this paper developed a method,combining parameterizing the two-dimensional section through Bezier curve with coordinate transformation,to realize the three-dimensional modeling of centrifugal blade.On the basis,the modeling program and its GUI interface were developed independently.Combined with the modeling program and optimization algorithm,a method of parametric fitting modeling for the existing blade was proposed,and the Eckardt impeller was used as an example to verify it.CFD numerical simulation results of original impeller and parameterized impeller show that the modeling program can realize the parameterized fitting modeling of the centrifugal blade with high precision,which provides a geometric tool for the further construction of optimization platform.2.Based on Python language,an optimization platform for the aerodynamic performance of centrifugal blade was established by combining the modeling program,CFD numerical simulation,design of experiments,approximate model and optimization algorithm.The optimization platform firstly parameterized the original blade by modeling program,and determined optimization parameters and geometric constraints.After that,three-dimensional samples were sampled by experiments.Using NUMECA software,sample library was constructed through CFD simulations of the samples.The data in sample library was used to construct the approximate model between the optimization parameters and the aerodynamic performance parameters based on the artificial neural network,and the iterative optimization of the blade was realized by the multi-objective genetic algorithm.3.In order to improve the optimization efficiency,the effect of lean angle of blade outlet on aerodynamic performance was studied through CFD numerical simulation,which helped establish the geometric constraint conditions for optimization process.4.The above optimization strategy and optimization platform was carried out to optimize the aerodynamic performance of a centrifugal compressor impellers with a flow coefficient of0.08 and 0.18 respectively.The verification results of CFD numerical simulation show that compared with the original impeller,the isentropic efficiency and total pressure ratio of the optimized impeller at the design condition are respectively increased by 0.63%,5.59 %(0.08)and 1.34 %,5.4 %(0.18).Through the analysis of the internal flow field,it can be seen that the energy loss is decreased and the flow uniformity is improved after optimization.In summary,the parametric modeling method and the optimization platform developed in this paper have certain rationality and practicability in the optimization of centrifugal impeller,which can provide reference for related research.