摘要
为了解决民用航空发动机定时水洗带来的水洗效果不明显或者未及时水洗的问题,对发动机水洗时机决策方法进行了研究。对水洗效果进行了分析,提出以期望水洗效果作为水洗时机确定的依据。考虑到水洗效果影响因素众多以及水洗样本容量小,将主成分分析与支持向量机结合起来,建立了发动机水洗效果预测模型。基于期望水洗效果以及预测的水洗效果确定合理的水洗时机。验证结果表明,建立的水洗效果预测模型优于多元线性回归模型,也优于单纯的支持向量机模型。
To solve problems that civil aero-engine water wash effect is not obvious or water wash is not timely, which are brought by a hard time water wash, aero-engine water wash time decision-making method is studied. The water wash effect was analyzed, and the expected water wash effect was proposed as the basis for determining water wash time. Taking into account many influencing factors of the water wash effect and small sample size of water wash, the principal components analysis was combined with support vector machine to establish the aero-engine water wash effect prediction model. A reasonable water wash time could be determined based on the expected water wash effect and the prediction of water wash effect. The proposed method was verified using the real data from an airline. Results showed that the proposed water wash effect prediction model was better than multivariate linear regression model and the pure support vector machine model.
出处
《航空精密制造技术》
2016年第3期33-37,共5页
Aviation Precision Manufacturing Technology
关键词
航空发动机
水洗时机
水洗效果
主成分分析
支持向量机
aero-engine
water wash time
water wash effect
principal components analysis
support vector machine