Kriging模型在机翼气动外形优化中的应用

针对粒子群等随机优化算法计算量大的缺点,发展了基于Kriging模型的优化方法.采用改进的量子粒子群算法对Kriging模型的相关模型参数进行优化,以提高代理模型预测精度,并与具有双层结构的粒子群算法相结合.采用雷诺平均N-S方程流场求解器与多目标非线性适应值加权方法,对高维度多目标多约束的跨声速机翼进行了优化,设计的机翼具有理想的压力分布,降低了机翼阻力系数,并且有效控制了低头力矩和翼根弯矩,表明该方法具有较强的工程实用性.

Application of Kriging surrogate model for aerodynamic shape optimization of wing

A optimization method based on Kriging surrogate model is developed in order to reduce the number of evaluations of global optimum algorithms such as PSO.A double-layer particle swarm optimization arithmetic based on the Kriging model which enhanced precision by using a modified QPSO to compute the global optimums of correlation model parameters.Reynolds-Averaged Navier-Stokes flow solver and nonlinear weighted sum of multi-objective method is used to optimize transonic wing with multi-objective and multi-r...