基于极限学习机的航空发电机旋转整流器快速故障分类方法研究

Fast Fault Classification Method Research of Aircraft Generator Rotating Rectifier Based on Extreme Learning Machine

航空发电机在现代多电和全电飞机的发展中将发挥越来越重要的作用,其重要部件的可靠性也是未来研究的重点。该文针对目前航空发电机旋转整流器（aerospace generator rotating rectifier,AGRR）的故障分类问题,提出了一种基于思维进化算法（mind evolutionary algorithm,MEA）的极限学习机快速分类技术。该方法通过MEA算法优化极限学习机的训练参数,以获取优化的识别模型,并将其应用于AGRR的故障分类中,取得了很好的效果。仿真和实验结果表明,经过优化的极限学习机与现有分类方法相比,具有很好的诊断性能和较高的分类速度。因此,该方法适合用于航空发电机旋转整流器的故障快速诊断和定位。

The aerospace generator is playing a more and more important role in the development of modern more-electric and all-electric aircraft. The reliability of important components of aircraft generator will be the focus in the future research. Focusing on the faults classification problem of aerospace generator rotating rectifier（AGRR）, this investigation presented a fast classification technique based on extreme learning machine（ELM）, improved with mind evolutionary algorithm（MEA）. This technique utilized the MEA to optimize the parameters of ELM, and hence, an optimized model of ELM could be achieved and then, applied to rotating rectifier faults classification of aerospace generator. Simulation and experimental results showed that, the optimized ELM could achieve good diagnosis performance and high classification speed. Hence, the presented method can be considered to the application of aerospace generator rotating rectifier faults diagnosis and localization.