改进多元宇宙算法在航空发动机不暖机模型修正中的应用

Application of Improved Multi-Verse Optimization Algorithm in Model Correction of Aero-Engine with Unheating

关键词:  涡扇发动机  暖机  模型修正  多元宇宙优化算法  虫洞机制公式  混沌

Key words:  Turbofan engine  Heating  Model correction  Multi-verse optimization algorithm  Formula of wormhole mechanism  Chaos

摘要:

航空发动机不暖机会产生性能损失，此时原有发动机模型已经不能准确表达发动机性能，因此，需要利用模型修正技术对原有的发动机模型进行修正，以获得发动机不暖机情况下的数学模型。提出了一种改进多元宇宙优化算法（Multi-verse optimization，MVO），并将其应用于发动机不暖机模型的修正研究。在常规MVO算法基础上，修改虫洞机制公式，解决解区间偏离0轴较远时寻优计算易陷入局部最优的问题，并引入混沌思想，通过混沌化初始宇宙和在每一代最优宇宙附近区域内进行混沌搜索，增强了算法的全局搜索能力。将改进后的算法应用于发动机模型的修正研究，并将常规MVO，改进MVO，粒子群算法（Particle swarm optimization，PSO）和遗传算法（Genetic algorithm，GA）四种算法的修正结果进行了对比。结果表明：修正后发动机不暖机模型精度得到很大提高，其中推力误差仅为0.07%，能够满足舰载机起飞动力学模型输入条件的精度要求；改进MVO算法对发动机模型的修正效果优于常规MVO算法，且相比PSO和GA，改进MVO修正效果同样更优。

Abstract:

The original engine model cannot accurately express the engine performance due to the engine performance loss caused by unheating. Therefore， it is necessary to use the model correction technology to modify the original engine model to obtain the mathematical model of the engine with unheating. An improved multi-verse optimization algorithm is proposed， which is applied to correct the model of engine with unheating. On the basis of the conventional multi-verse optimization algorithm， the wormhole mechanism formula is modified to solve the problem that the calculation is easy to fall into the local optimum when the solution range deviates from the 0 axis far away. And the chaos is introduced. The global search ability of the algorithm is enhanced by the chaotic initialization universe and the chaotic search in the region near the optimal universe of each generation. The improved algorithm is applied to the study of engine model correction， and the correction results of conventional MVO， improved MVO， PSO and GA are compared. The results show that the accuracy of the model of the engine with unheating is greatly improved after correction， and the thrust error is only 0.07%， which can meet the input accuracy requirements of the carrier aircraft take-off dynamics model. The results also show that the improved MVO algorithm is better than the conventional MVO algorithm. Compared with PSO and GA， the results of model correction by improved MVO are better as well.