基于融合EKF的航空发动机气路性能健康预测

针对集中式滤波算法存在计算效率不高、容错性差,引入融合滤波的思想,提出采用非线性融合的联邦式扩展卡尔曼滤波器进行发动机气路健康性能预测.子滤波器根据量测参数完成发动机部分健康性能的局部估计,主滤波根据子滤波器估计参数完成融合滤波估计,并将状态估计值和协方差反馈至子滤波器用于下一步健康预测.通过某型涡扇发动机仿真表明:融合EKF滤波器能准确地预测发动机的健康状态,估计稳定收敛时间短、计算时间短、效率高.

Gas Path Performance Health Prediction for Aircraft Engine Based on Fusing EKF

In order to improve the computational efficiency and fault tolerance of the conventional centralized filtering algorithms,a novel federated extended Kalman filter with the nonlinear fusing architecture is developed to predict the engine gas path health.The local filter is employed to estimate the local health states by using the measurement subsets,and the global estimations are reached with fusing strategy in the master filter.The current state estimations and covariances are fed back to the local filters for the next health prediction.The simulation is carried out on a turbofan engine,and the results reveal that the F-EKF can predict the gas path performance health accurately.Besides,the proposed methodology has the advantages of less convergence time and computational time,and higher efficiency.