Uncertainty Analysis of Airfoil Wind Tunnel Experiment Using CFD and Automatic Differentiation
Uncertainties residing in wind tunnel test results are crucial for using the experimental data objectively.In order to acquire some qualitative knowledge addressed to the concern,a reliable computational fluid dynamics (CFD) package in combination with automatic differentiation techniques is invited to approximate the uncertainty analysis procedure in the real world.Uncertainties are quantified by means of a gradient-based first-order moment approach.The required gradients information is calculated via the automatic differentiation (AD) algorithm.By differentiating the CFD program piece by piece,the sensitivities of objectives to given input parameters can be obtained in a single computational course in an efficient and timely manner.The source perturbations are given according to the controlling accuracy of wind tunnels.The methodology is applied to 2D airfoil cases in both subsonic and transonic context.Uncertainties of surface pressure distribution,as well as three typical integral coefficients including lift,drag and pitching moment,caused by perturbations of stream condition variables,such as angle of attack,Mach number and Reynolds number,are exploited in detail,and some practical recommendations about improving experimental accuracy are reported.
Automatic differentiation transonic flow Sensitivity analysis Uncertainty Airfoil Wind tunnel experiment
Xu Lincheng Wang Gang Wu Jie Ye Zhengyin
National Key Laboratory of Aerodynamic Design and Research, NPU, Xi’an 710072, China
国际会议
The 8th International Conference of Computational Fluid Dynamics, (ICCFD8)(第八届国际计算流体力学会议)
成都
英文
1-12
2014-07-25（万方平台首次上网日期，不代表论文的发表时间）