故障检测滤波器鲁棒性设计新方法

A New Method for Designing Robust Fault Detection Filter

针对航空发动机动态过程中由于模型误差造成故障检测误报的问题,采用参数变化率法建立航空发动机动态线性变参数(LPV)数学模型反映发动机动态特性,同时结合全维非线性未知输入观测器,利用特征结构配置实现故障检测滤波器对于被控对象模型误差和系统不确定性因素的干扰解耦。通过某型涡扇发动机控制系统传感器故障检测仿真表明,减小残差信号与故障信号之间的误差,提高了残差信号对模型误差和未知输入信号的鲁棒性,对发动机过渡过程中故障检测准确度高,实时性好。

Reducing fault detection mistake in aero engine dynamic process is a difficult problem.This paper builds a linear parameter varying model of aero engine by using parameter velocity method to exactly describe aero engine dynamic process characteristics.The simulation results of turbofan engine control system sensor fault show that this method reduces the difference between residual signal and fault signal,enhances the robustness to unknown input,and the detecting accuracy is high.