Multi-objective Optimization And Blade Parameter Design Of Open Axial Fan Based On Isight

Axial fans are widely used in industrial and household products because of their simple structure,large air volume and high efficiency.Air volume and efficiency are two important indicators for evaluating the performance of axial fans.Improving the performance of axial fans has an important role in saving energy.In this paper,based on an open axial fan used in automobile cooling system,numerical simulation and optimization of airfoil multi-parameter（bend,swept and winglet）are performed.The specific work content is as follows:Using Matlab program,the parametric design of the airfoil blade was carried out.Based on the original blade,the forward bend,forward sweep,and modification parameters of the winglet are defined.The surface geometry of the blade is calculated and output by Matlab,so that the rapid drawing of the blade（skewed swept and winglet）can be achieved.The computational fluid dynamics method was used to verify the accuracy of the computational simulation and check the convergence of the grid.The study uses the steady fluid dynamics equations and RNG k-εturbulence model,a structured grid with single passage,and the rotor-stator interactional surface adopts MRF treatment.The error between the calculated results and the experimental results is small,and the calculation convergence is faster,which is convenient for subsequent optimization work.Based on the Latin hypercube multi-parameter sample setting method,the sample space of the forward bend,forward sweep and winglet parameters was designed.15 sets of parameter combinations were obtained,which not only met the minimum sample number condition,but also achieved a wider coverage of each parameter.For the designed 15 sets of parameters,the blade geometry was plotted and simulation was completed,and the air volume and efficiency of different fans were obtained.Selecting Isight software,using Kriging approximation model and NSGA-Ⅱoptimization algorithm,with wind volume and efficiency as the goal,forward bend,forward sweep and winglet as variable parameters,the multi-objective optimization work was completed.The optimal air volume and efficiency obtained through Isight optimization have been improved by 26.06%and 21.07%compared to the original fan.The optimal parameter combinations C₁,C₂,and C₃ are 0.579,0.941,and 0.704,respectively.The optimal parameter combination was verified by CFD.The internal rule also verified that the optimal parameter combination significantly increased the axial velocity in the near-tip region,and suppressed the tip-stall region through the winglet,thereby significantly improving air volume and efficiency.In the traditional aerodynamic design method of axial fans,there is no design method for forward bend,forward sweep and winglet parameters.In this paper,we try to use feedback neural network method to do this kind of parameter design work.In the established neural network,the wind volume and efficiency are used as input variables,and the forward bend,forward sweep,and winglet parameters are used as output variables.The above 15 groups of samples are used for learning and training.The neural network can converge,which shows that the nonlinear relationship between the air volume,efficiency and these three variables can be established using the feedback neural network method.Finally,through the trained neural network,the design results of forward bend,forward sweep,and winglet are calculated when the optimal air volume and efficiency are calculated.The results are in good agreement with the results of Isight and CFD,which indicated that this design method had certain reliability.This thesis focuses on the multi-parameter optimization of axial fan blades.Fast parameterized geometric design,accurate and fast CFD simulation,the smallest possible sample space,and reasonable optimization strategies are all very important steps.By studying and synthesizing similar work at home and abroad,this thesis achieves a high performance improvement for the axial fan.At the same time,for this type of airfoil parameter without a clear design method,this thesis proposes a feasible design ideas can be used in similar optimization work.