Abstract: A linear combination of health parameters method is proposed for the Kalman filter estimation problem where there are fewer available sensors than gas path health parameters to be estimated in an aircraft engine. An optimal transformation matrix selected by an iterative optimal search routine was used to construct a tuning parameter vector which was a linear combination of all health parameters, and its dimension was equal to the number of available sensors. The estimation value of original health parameters was obtained by the tuning parameter vector, whose dimension was low enough to enable Kalman filter estimation. Results from the application of the proposed method to a turbofan engine simulation are presented and compared with the conventional approach of a subset of the health parameters. When the efficiency and flow capacity of the engine components have 2% or 3% degradation, the new methodology has shown a significant improvement in gas path health parameters estimation accuracy. 

Key words: Aircraft engine; Health parameter; Linear combination; Kalman filter

摘要： 针对发动机上可用传感器个数小于待估计气路健康参数个数时的卡尔曼滤波估计问题,采用一种健康参数线性组合的方法。通过迭代搜索方法优化得到最优变换矩阵,对原健康参数进行线性组合,生成一组维数等于可用传感器个数的调整参数向量,然后采用卡尔曼滤波器对调整参数向量进行估计,最后通过还原变换得到原健康参数的估计值。以涡扇发动机为研究对象的仿真结果表明,在风扇、压气机的效率以及涡轮效率和流量分别或同时蜕化2%到3%时,该方法能够准确估计出所有的健康参数,而估计参数子集法的估计结果有可能出现明显偏差,甚至误判。 

关键词: 航空发动机;健康参数;线性组合;卡尔曼滤波器 