一种复合故障预测动态建模方法

A complex fault prediction dynamic modeling method

本文提出一种复合故障预测动态建模方法,将原有故障数据样本经过参数相关性分析,按特征输入支持向量回归机建立相应的故障预测模型,可以很有效预测未来较长时间的数据变化。对航空发动机的飞行参数进行预测的结果表明,该方法的预测精度优于现有的支持向量回归方法。

The aeroengine fault parameter is estimated by the way of SVR which is based on the correlation analysis. The flight data of aeroengine is imported to SVR before its correlation is analyzed as the appliaction object and distilled. When the fault model is achieved, we can get the result under the forecasting method. The method can forecast for a long time and the precision is very high.