舵面损伤在线故障模式预测及故障检测

On-Line Fault Mode Prediction and Fault Detection for Control Surface Damage

舵面损伤的传统模型匹配检测方法只适用于线性气动模型 ,并且故障检测和定位能力受操纵和机动状态影响很大。产生这些问题的原因是在故障模式预测和故障检测中忽略了一些实时飞行状态和参数变化。文中将传统的离线故障模式预测改为在线计算 ,在计算中引入实时飞行状态及参数 ,并且在使用解析余度估算正常飞机运动时 ,考虑了实时飞行状态、舵面偏转、质量、惯量等参数变化 ,由实时估算值与实测值比较获得残差。此外 ,采用了故障模式空间最优方向匹配法进行故障隔离以减小残差决策的计算量。对所建立的新算法可以不使用任何简化、假定。使用某机舵面损伤非线性数据进行了分析和仿真 ,结果表明该方法可以适用于多种故障气动模型 ,而且不受机动状态。

The traditional fault detection and isolation (FDI) algorithm in self-repair flight control system for control surface damage is constantly criticized for its limited applicability. The FDI algorithm needs to be improved for different aerodynamic models, control inputs and maneuvering states. Simplifications and assumptions during FDI fault mode calculations cause its limited applicability. We propose to eliminate these simplifications and assumptions by taking four measures: (1) predicting the fault mode on-line with real-time flight state and parameters; (2) estimating the normal aircraft movement by analytical redundancy using real-time maneuvering state, aerosurface deflection and aircraft parameters; (3) obtaining the residual by comparing the measured aircraft movement, which may contain fault message; (4) in order to reduce the computing requirement, using a best-direction matching in fault mode vector space for fault isolation. Analysis and simulation of the nonlinear fault aerodynamic data prove that the new method can work with various aerodynamic models without limitation on flight status and controls, thus satisfying the requirements of engineering applications.