Font Size: a A A

Some Applications And Comparative Studies Of Adjoint Optimization Method In Internal Flow

Posted on:2023-06-27Degree:MasterType:Thesis
Country:ChinaCandidate:S J ZhangFull Text:PDF
GTID:2532306830973949Subject:Power engineering
Abstract/Summary:PDF Full Text Request
This paper mainly adopts Fluent adjoint method and CAESES-based parametric optimization method to optimize the aerodynamic shape design of internal flow components,including numerical calculation and analysis of flow field,solution of adjoint equations,construction of parametric optimization platform and comparison with parametric optimization.In the adjoint method,the steepest descent method and the mesh deformation method based on Bernstein polynomial and B-spline are used to achieve the adjoint optimization.Then a 3D parametric fully automatic optimization platform was built based on the parametric optimization software CAESES and compared with the adjoint methods.The specific work carried out is divided into the following aspects.(1)A detailed study of the adjoint method for internal flow components is carried out.The adjoint optimization process based on Fluent Adjoint Solver and the optimization algorithm and mesh deformation steps used in the process are briefly introduced.The effectiveness of the adjoint method is verified by applying the adjoint method to a certain type of supersonic air inlet.The adjoint sensitivity of total pressure loss coefficient and outlet airflow distortion is solved by using discrete adjoint equations,and the area of mesh deformation is determined according to the inlet surface sensitivity to achieve single-and multi-objective optimization of the aerodynamic shape of the supersonic inlet.(2)A full 3D parametric optimization platform is built by combining parametric design,numerical calculation and parametric optimization algorithm.A CAESES-based parametric optimization design of a supercritical CO2 inlet volute is carried out.The global sensitivity analysis Sobol algorithm is used for sampling,and a model library is established from which the optimal model is selected,based on which a tangent search method is used for local optimization.(3)Adjoint and parametric optimization were carried out for an S-type subsonic inlet and a type of mixed mass centrifugal compressor discharge volute,and the results obtained from the two optimization methods were compared and analyzed.After 300 steps of optimization based on the global search method combined with the tangent search method in the parametric optimization platform,the total pressure recovery coefficient increased by 0.186%,and the performance of the inlet tract flow field improved from the model,Mach number cloud diagram and total pressure cloud diagram before and after optimization.The results of the exhaust volute optimization showed that the total pressure recovery coefficient of the exhaust volute increased by 0.577% after 65 steps of tangent search method,while the total pressure recovery coefficient increased by 1.614% after the adjoint optimization,and the pressure expansion capability of the volute was also significantly improved.After comparison,the adjoint optimization makes the aerodynamic shape of the air inlet and volute change more obviously,and the aerodynamic performance is improved more significantly,and the convergence solution is also reached faster to find the optimal model.
Keywords/Search Tags:Intake, Volute, Discrete adjoint equations, Parametric design, Aerodynamic shape optimization
PDF Full Text Request
Related items