非结构网格伴随各向异性自适应音爆近场计算

Adjoint-based Anisotropic Unstructured Grid Adaptation for Sonic Boom Near Field Computation

针对音爆近场CFD计算网格分辨率要求,发展了基于离散伴随的各向异性非结构网格自适应技术。根据指定面上过压信号面积分,采用离散伴随误差估计得到期望的网格尺寸并结合马赫数黑森矩阵构建所需的网格度量张量、用于指导h和r自适应策略优化网格分辨率,并采用二维超音速平板和Seeb-ALR算例进行了验证测试。结果表明,所得自适应网格符合音爆近场计算网格分辨率要求,能以较少网格准确捕捉到激波和膨胀波等流场结构和精确计算近场过压信号。对音爆近场计算网格生成、精确预测和基于CFD的低音爆设计具有积极意义。

Anisotropic unstructured grid adaptation based on discrete adjoint method is developed for the grid resolution requirements of sonic boom near field computation.The presented approach obtains the desired mesh size using the adjoint based error estimate for the surface integral of over pressure.The estimated mesh size is incorporated into the Mach number Hessian matrix to get a better grid metric for improving grid resolution by h and r adaptation.The presented approach is demonstrated in 2D using supersonic flat plate case and in 3D using SEEB ALR case.Results show that the resulting adapted grid meets the resolution requirements for sonic boom near computation and produces accurate over pressure signals.It provides insight to grid generation of sonic boom simulation,and is significant for accurate sonic boom prediction and CFD based low boom design.