论文部分内容阅读
伴随优化方法在飞行器气动外形优化设计方面的应用越来越广泛,但以精细减阻等为目标的优化,面对真实飞行器仍然存在计算量大、计算周期长的问题。将多重网格技术结合并行技术运用于基于非结构混合网格、雷诺平均Navier-Stokes方程的离散伴随方程的求解,提高了离散伴随优化系统的整体效率。给出了伴随方程多重网格计算中的延拓和限制算子,采用V循环比较了不同层数的加速效果和粗网格残差计算方法对目标函数梯度敏感导数的影响。结合Metis分区技术,提出了适合伴随方程的数据并行传递简化方式,使伴随方程在300个并行分区计算时加速比仅比理想加速比低13%。采用发展的高效优化系统,选取了112个设计变量,对DLR F6翼身组合体跨声速状态进行减阻优化,使机翼空间激波得到弱化,阻力减小9counts。模拟验证结果表明,建立的高效的飞行器气动外形优化设计系统在三维真实飞行器外形优化方面,具有良好的应用前景。
Accompanying the optimization method, it is more and more widely used in aerodynamic shape optimization design of aircraft. However, the optimization with the goal of fine drag reduction and so on, still has the problem of large calculation and long calculation period in the face of real aircraft. The multi-grid technology combined with parallel technology is applied to solve the discrete adjoint equation based on the unstructured mixed grid and Reynolds-average Navier-Stokes equations, which improves the overall efficiency of the discrete adjoint optimization system. The continuation and restriction operators in the multi-grid computing are given. The V-cycles are used to compare the effects of acceleration effects and coarse-grid residual calculation methods on the gradient-sensitive derivatives of the objective function. Combined with the Metis partitioning technique, a simplified data parallel transfer scheme is proposed, which makes the adjoint equations accelerate only 13% lower than the ideal speedup ratio in 300 parallel partitioned computing. With the development of an efficient optimization system, 112 design variables were selected to reduce the transonic state of the DLR F6 wing-body combination and to reduce the shock in the wing space, reducing the drag by 9counts. Simulation results show that the established aerodynamic shape optimization design system for aircraft has good application prospects in the aspect of shape optimization of three-dimensional true aircraft.