流场分析和设计的并行计算

网络连接机群是一种有效的并行计算工具，讨论了在此环境下流场分析和设计的并行计算问题。流场解中采用了Euler方程作为主控方程，并用有限体积方法和时间隐式方法进行求解。在MPI／PVM环境下用分区方法作了二维翼型和三维机翼绕流的并行计算。数值算例表明流场计算的正确性和并行计算的有效性，并讨论了影响加速比和并行效率的各种因素。用耦合流场解和并行遗传优化算法做了二维翼型和三维机翼的单目标／双目标数值优化。算例表明：使用的适应函数优于传统线性组合法构成的，遗传算法计算三维优化问题时必须并行化。

Parallel Computation of Flow Field Analysis and Design

A network of computers connected via ethernet is one of effective parallel computation tools. Parallel computation of flow field analysis and design on this kind of network are carried out. Euler equations are used as governing equations in the flow solver, which are solved by using finite volume method and time implicit method. The flow fields of an airfoil and simplified aircraft are calculated parallelly under MPI/PVM environmcent. Numerical results show the correction of calculation and effectiveness of parallel computation. Some parameters influencing speed up and efficiency are discussed. Mono-objective/bi-objective optimization of an airfoil and a wing are made by coupling the flow solver and a parallel genetic algorithm. Numerical results show that present fitness function is better than that constructed by linear weighted combination and that 3D optimization using genetic algorithms has to be paralleled.