摘要
针对航空发动机上可用传感器测量参数偏少情况下的健康参数估计问题,提出1种先分类后估计的方法。将传感器测量参数输入异常监测模块,对发动机工作状态进行监测,若监测结果为无故障则直接给出无部件故障的诊断结论;否则将测量参数输入最小二乘支持向量机(LSSVM),对部件故障进行分类,卡尔曼滤波器根据分类结果只对故障部件的健康参数进行估计。仿真结果表明:该方法可以减少需要估计的健康参数,提高估计精度。
The fault classification and health parameter estimation method was proposed for the problem where there were fewer available sensors in an engine. The sensor measurements were inputted to the anomaly detection module for monitoring engine operation. If the detection result was no fauh, no component fault conclusion can be made, otherwise the measurements were then inputted to the least squares support vector machine （LSSVM） for fault classification. The Kalman filter was only used to estimate the health parameters of the fault components. The simulation results show that the method can reduce the number of health parameters to be estimated, and yield a significant improvement in health parameter estimation accuracy.
出处
《航空发动机》
2012年第1期47-50,共4页
Aeroengine
