摘要
将航空发动机润滑油光谱分析的界限值作为动态系统进行研究,对传统的制定界限值的方法进行了对比研究,运用神经网络理论,结合故障诊断思想,建立了动态界限值模型;针对动态界限值模型泛化能力不高的问题,建立遗传神经网络模型,实现了动态界限值模型的优化;利用积累的航空发动机运行的故障数据和实际工作中获得的数据进行分析,验证了动态界限值模型的有效性。
The threhold of oil spectrom analysis in the aero-engine was considered as a dynamical system, and triditional metholds of making threhold were comparatively analyzed. By combining artificial neural network （ANN） and the way of fault diagnosis, the ANN model for dynamically adjusting the threhold was established. Aimed at the problem that the model can＇ t perform well in practise, the Genetic-ANN model was put forward, which can automatically realize performance optimizing by genetic algorithm （GA）. The accumulated data from certain kinds of fault and practical performance data from aero-engines were used as training and analyzing date, and the results verified the correctness of the established model.
出处
《润滑与密封》
CAS
CSCD
北大核心
2009年第6期89-92,113,共5页
Lubrication Engineering
关键词
界限值
动态调整
神经网络
遗传算法
航空发动机
threhold
dynamical adjustment
artificial neural network
genetic algorithm
aero-engine
