发动机磨损状态的改进LS-SVM监控研究  

Monitoring of Aviation Engine Friction and Wear Based on LS-SVM

作　　者：黄永武[1] 王伟平[1] 韩孟涛[1] 

机构地区：[1]空军工程大学工程学院航空自动控制工程系,陕西西安710038

出　　处：《仪表技术》2009年第6期54-56,59,共4页Instrumentation Technology

摘　　要：介绍最小二乘支持向量机对航空发动机滑油系统铁元素浓度的变化趋势进行预测;并采用遗传算法对最小二乘支持向量机的参数进行优化。通过与时间序列分析的预测结果相比较,仿真实验结果表明:得到的最小二乘支持向量机的预测精度高,具有很好的泛化能力和学习能力。That LS - SVM forecasting the change trend of the iron element density in the aviation engine slippery oil system is introduced in this paper. In the paper, the hyper-parameters of LS - SVM could be optimized by genetic algorithm. The results show that the model has excellent learning ability and can provide more accurate data prediction compared with Time Series. The errors of the model are very little and the method is effective.

关 键 词：最小二乘支持向量机 遗传算法 时间序列 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]