基于分布估计算法的涡桨发动机控制研究

An Optimization Control for Turboprop Engine Based on Distribution Estimation Algorithm

涡桨发动机控制系统是典型的多变量耦合控制,为充分发挥涡桨发动机的整体推进效率,需充分考虑螺旋桨和发动机之间的相互关系;分布估计算法是智能优化领域解决非线性、变量耦合的一种有效优化算法。将分布估计算法引入到航空涡桨发动机多变量控制中,并且针对分布估计算法中构造概率分布模型及模型优化估计时间代价巨大的问题,提出改进适应性方案,使寻优时间可接受,解决了应用到发动机控制领域的可行性;建立某型发动机、螺旋桨及控制系统的仿真模型。基于改进后方案,考虑螺旋桨和发动机的实时相关作用,直接在发动机及螺旋桨非线性模型的基础上对控制参数进行整体寻优,并与传统的设计方法数据仿真分析对比,验证了分布估计算法对解决多变量控制的有效性。

An improved adaptive algorithm and its theoretical analysis are proposed for the probability distribution model in the algorithm according to the characteristics of turboprop engine.The time and space cost in the density estimation al-gorithm is solved when it is applied to the engine control.An integrated simulation model of engine,propeller and the control system is established.The control parameters are optimized based on the distributed estimation algorithm and the control ef-fect is com-pared with the traditional method of separate regulation.