航空发动机性能仿真的分群量子粒子群算法  

Sup-population Quantum-behaved Particle Swarm Optimization for Aero-engine Performance Simulation

作　　者：逄珊[1] 杨欣毅[2] 张小峰[1] 王增锋[1] 

机构地区：[1]鲁东大学信息与电气工程学院,烟台264025 [2]海军航空工程学院飞行器工程系,烟台264001

出　　处：《系统仿真学报》2013年第6期1176-1182,共7页Journal of System Simulation

基　　金：国家自然科学基金项目(61170161);山东省科技发展计划项目(2011YD04049)

摘　　要：针对传统迭代方法求解航空发动机模型非线性方程组存在受初值影响不易收敛的问题,采用量子粒子群算法求解。为解决算法的局部收敛现象,设计一种分群量子粒子群算法:将种群分为多个分群,每个分群在各自全局极值的引导下搜索解空间不同区域,并对精英解集定期更新。对测试方程组的求解表明分群机制能有效提高量子粒子群算法的搜索性能和收敛速度。运用改进算法对某型混合排气涡扇发动机模型进行仿真求解,得到了满意的结果。When solving aero-engine mathematical model’s equations with traditional iteration methods,it is difficult to attain convergence as these methods are very sensitive to initial values.Therefore quantum-behaved particle swarm optimization（QPSO） is used to solve the model.In order to tackle the local convergence problem of QPSO,a Sup-population Quantum-behaved Particle Swarm Optimization（SQPSO） is proposed.In SQPSO,the whole population is divided into several sub populations and each sub population is assigned a global best particle.Under the guidance of different global best particles,sub populations search different area of solution space.Elite solution set,which composed of global best particles,is updated periodically.Results on test nonlinear equations show that the sub population strategy improves algorithm’s searching performance and convergence speed effectively.Ideal results were obtained using SQPSO when solving a mixed exhaust turbofan engine mathematical model’s equations.

关 键 词：航空发动机 非线性数学模型 量子粒子群 收敛 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]