基于混合卡尔曼滤波器组的航空发动机双通道传感器故障检测

Aircraft Engine Sensor Fault Diagnostics Through Dual-Channel Sensor Measurements Based on a Bank of Hybrid Kalman Filters

基于机载非线性模型与分段线性卡尔曼滤波器混合组成的混合卡尔曼滤波器组,结合双通道传感器的特点,建立了民用航空发动机传感器故障诊断系统;给出故障诊断原理及算法的同时,将该系统应用于民用涡扇发动机传感器常见典型故障进行了仿真;仿真结果表明,诊断系统可以在发动机发生健康蜕化后,通过只简单更新机载模型的蜕化因子,而保持线性卡尔曼滤波器的参数不变,便能准确地检测和隔离各类传感器故障而不发生误报;该更新过程可以在线自动完成,省时省力,易于工程实现。

Based on a bank of hybrid Kalman filters which are hybrids of a nonlinear on-board engine model（NOBEM） and piecewise linear Kalman filters,a civil aircraft engine sensor fault diagnostics system which utilizes dual-channel sensor measurements is developed.Principles and algorithms of sensor fault detection,isolation and accommodation are given.By this system applied to some typical civil turbofan engine sensor faults,simulation results show that the diagnostic effectiveness of the system is maintained to avoid false alarms as the health of the engine degrades over time through a simple process： by feeding the health degradation values into the NOBEM and not changing the parameters of the linear Kalman filters.The update process,which can be completed automatically online to save time and effort,is feasible in the real application environment.