关键词: 管路布局, 多目标优化, 拉丁超立方, 种群更新机制, 航空发动机

Abstract: To improve the optimization efficiency of aero-engine pipe layout,a multi-objective pipe routing method for aero-engine based on improved Non-dominated Sorting Genetic Algorithm (NSGA-Ⅱ) was presented.To improve the performance of NSGA-Ⅱ with global searching ability,a new population update mechanism was proposed,in which the Latin hypercube method was used to replace the population of some poor individuals in the evolutionary process,while retaining the original elitist strategy so that it could achieve the balance between global search ability and convergence.The 3D convex hull was used for modeling the routing space,and an encoding mechanism based on B-spline curve was designed.Further,pipe length and smoothness were selected as the optimization objectives,and then the improved NSGA-Ⅱ was used to find Pareto solutions of pipe layouts.The feasibility of the proposed method was verified by several numerical computations and pipe routing examples.

Key words: pipe layout, multi-objective optimization, Latin hypercube, population update mechanism, aero-engine