基于状态空间模型的飞机 APU 在翼 RUL 预测方法

为解决利用飞机辅助动力装置(APU)在翼监测数据难以表征其性能状态而造成的性能评估以及剩余使用寿命预测(RUL)难的问题,本文提出一种基于状态空间模型(SSM)与卡尔曼滤波融合的APU在翼RUL预测方法。首先,通过在翼监测数据构造含噪声的性能指标(PI)来表征APU的性能状态,借助维纳过程与建立的含噪声的PI构建状态方程,来描述APU性能衰退过程。然后,将卡尔曼滤波状态估计和预测方法应用于SSM,通过对APU在翼性能状态的估计,达到预测其RUL的目的。最后,采用国内航空公司运营的APU在翼监测数据进行方法的综合验证和评估。实验结果表明,与ELM和Optimized ELM相比,本文方法的预测绝对百分比误差分别减少了72.1%和67.9%。此外,与其它3类方法的实验结果对比,本文方法的预测绝对百分比误差至少减少了69.2%。该方法可以有效地实现在翼APU的RUL预测,可为运维人员合理规划维护维修提供参考,更为重要的是在一定程度上可以提高旅客的舒适性和飞机的安全性。

On-wing RUL prediction method of aircraft APU based on state space model

The on-wing monitoring data of aircraft auxiliary power unit (APU) are difficult to characterize its performance states, which will lead to the difficulty that the performance evaluation and remaining useful life (RUL) prediction of the APU is difficult to carry out. To solve this problem, a performance evaluation and RUL prediction approach of APU is proposed based on the state space model ( SSM) and Kalman filter (KF). Firstly, a performance indicator (PI) containing noise is constructed from the on-wing monitoring data to characterize the performance state of the APU. The performance degradation process of APU is described by state equation, which is constructed with the help of the Wiener process and the constructed PI with noise contained. Then, the KF state estimation and prediction method is applied to the SSM. Through estimating the on-wing performance state of APU, the purpose of predicting RUL is achieved. Finally, the APU on-wing monitoring data from the operation of an airline company in China are adopted to conduct the comprehensive verification and evaluation of the proposed method. Experiment results show that compared with ELM and Optimized ELM, the prediction absolute percentage error of the proposed method is reduced by 72. 1% and 67. 9% , respectively. In addition, compared with the experiment results of other three kinds of methods, the prediction absolute percentage error of the proposed method is reduced by 69. 2% at least. The proposed method can effectively predict the RUL of an on-wing APU, which can provide a reference for the operation and maintenance personnel to plan maintenance and repair reasonably. More importantly, the method can improve the comfort of the passengers and the aircraft safety to a certain degree.