基于遗传算法的涡喷发动机身份证模型建立

Establishment of Turbojet Engine Identification Model Based on Genetic Algorithms

为建立发动机身份证模型,针对发动机个体差异和寿命期内性能蜕化引起的发动机模型与真实发动机之间的失配问题,提出1种基于遗传算法的发动机部件级模型自动修正方法。在发动机部件级模型中,引入能反映个体发动机部件特性的性能调节因子。根据发动机试车数据与待修正部件级模型输出数据,以发动机关键测量参数残差最小为优化目标,采用遗传算法获得不同换算转速下的特性修正系数,建立发动机身份证模型。以某型涡喷发动机为对象进行试验验证,结果表明该方法能有效提高发动机部件级模型精度,适用于建立发动机身份证模型。

In order to establish the identification model of a turbojet engine, based on the mismatch between the real engine and the engine model caused by individual differences and performance degradation in the life time, an automatic modification of component-level model from genelic algorithm （GA） was proposed. Performance adjustment factor which can reflect the individual characteristics of engine components was introduced into the engine component-level model. The optimal target was to minimize the errors of key measurements based on engine test data and the output of the component-level model. The correction coefficients were obtained using genetic algorithm and the identification model of the engine was buih. The demonstration test was conducted on a turbojet engine. Results show that the model precision of the engine components was effectively improved by this method and the method is suitable for building up certificate turbojet engine model.