论文部分内容阅读
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。
Aiming at the problem of health parameter estimation under the condition of too few sensors available on aero-engine, a method of pre-classification and post-classification estimation is proposed. The sensor measurement parameters are input to the anomaly monitoring module to monitor the working status of the engine. If the monitoring result is no fault, the diagnosis conclusion without a component fault is directly provided; otherwise, the measurement parameters are input to the least squares support vector machine (LSSVM) Fault classification, Kalman filter based on the classification of the results of the fault component only to estimate the health parameters. Simulation results show that this method can reduce the need to estimate the health parameters and improve the estimation accuracy.