CFD Simulation-Based Aerodynamic Performance Assessment And Sensitivity Analysis Of Aircraft Wings Under Hybrid Uncertainties

Computational fluid dynamics(CFD)has widespread applications in the aerodynamic performance evaluation of modern aircraft wings.Nevertheless,the current CFD simulation cannot thoroughly reflect the real flow since the manufacturing deviation and abundant assumptions of the model parameters in CFD simulation.Therefore,aleatory and epistemic uncertainties inevitably coexist in CFD simulation.These uncertainties couple to formulate hybrid uncertainty which has a great impact on the evaluation results.Traditional CFD simulation methods are formulated by deterministic methods or only consider a single type of uncertainty,which cannot fully reflect multitypes uncertainties in real aerodynamic performance.Therefore,it is necessary to research CFD simulation under hybrid uncertainty.The probability box(P-Box)can characterize hybrid uncertainty,aleatory uncertainty,epistemic uncertainty,and constant.To date,this thesis is going to develop a wing aerodynamic performance evaluation method via CFD under the P-Box framework.Besides,a novel sensitivity analysis of P-Box variables will be proposed to analyze the influence of wing structure and flow field variables on wing aerodynamic performance.The contents and innovations are summarized as follows:(1)Development of a wing aerodynamic performance by CFD under parametric and non-parametric P-Box.The parametric and non-parametric P-Boxes are introduced to quantify the hybrid uncertainty of wing structure variables.The proposed method uses Gaussian Process(GP)model to replace the time-consuming CFD model.Besides,we introduce the P-Box convolution sampling uncertainty propagation method to evaluate the wing aerodynamic performance.Two case studies show that the developed method can thoroughly reflect subjective and objective uncertainty information.(2)Development of a separating sensitivity analysis(SSA)method under parametric P-Box.An area metric and a maximum variance metric are developed to formulate the epistemic and aleatory SSA indices,respectively.A numerical example is exemplified to demonstrate the feasibility of SSA.Meanwhile,an engineering case provides the effects of wing structure variables on wing aerodynamic performance,which shows the necessity of separating the effects of aleatory and epistemic uncertainties.(3)Development of an SSA method with non-parametric P-Box inputs.This method is a further extension of SSA on non-parametric P-Box.Two methods,i.e.,grid point method and expectation method,are developed to separate the input aleatory and epistemic uncertainties in input P-Box variables,respectively.The engineering case demonstrates the impacts of flow field variables on wing aerodynamic performance.The examples show the proposed SSA can differentiate the aleatory and epistemic effects of non-parametric p-boxes and provide important evidence for further wing aerodynamic design optimization.