基于多信号流图的大气数据系统故障诊断

针对传统无人机大气数据系统故障诊断依赖专家经验、故障定位不准确的问题，提出了基于多信号流图的大气数据系统故障诊断方法。通过分析大气数据系统的组成结构和故障模式影响与危害度分析，采用多信号流图方法对某型无人机大气数据系统关键部件进行测试性分析，得到大气数据系统的故障-测试相关性矩阵。针对传统实时测试性工程与维护系统(Real-Time Testability Engineering and Maintenance System，TEAMS-RT)算法诊断速度较慢的问题，对TEAMS-RT算法进行优化研究，提出了一种基于TEAMS-RT的矩阵分解优化(简称Tree-RT)算法，在此基础上对D矩阵分别采用Tree-RT算法和测试一致性算法对大气数据系统进行故障注入试验。试验结果表明，Tree-RT算法对故障注入试验的诊断率为94.62%，高于基于测试一致性算法85.73%的诊断率，证明了大气数据系统多信号流图和Tree-RT算法的有效性，为无人机大气数据系统的故障诊断相关研究提供了技术支撑。

Fault Diagnosis of Air Data System Based on Multi-Signal Flow Graph

In order to solve the problem that the fault diagnosis of traditional UAV air data system depends on expert experience and the fault location is inaccurate,a fault diagnosis method of air data system based on multi-signal flow graph is proposed.By analyzing the composition structure,fault modes effects and criticality analysis of the air data system,the testability analysis of the key components of the air data system of a certain UAV is carried out by using the multi-signal flow diagram method,and the fault-testing correlation matrix of the air data system is obtained.In order to solve the problem of slow diagnosis speed of the traditional real-time testability engineering and maintenance system (TEAMS-RT) algorithm,the TEAMS-RT algorithm is optimized.A matrix decomposition optimization (also known as Tree-RT) algorithm based on TEAMS-RT is proposed.On this basis,the fault injection test of air data system is carried out by using Tree-RT algorithm and testing consistency algorithm for D matrix respectively.The experimental results show that the diagnostic rate of the Tree-RT algorithm for the fault injection test is 94.62%〖KG-*5〗,which is higher than the diagnostic rate of 85.73% based on the testing consistency algorithm.It proves the effectiveness of the multi-signal flow graph and the Tree-RT algorithm of the air data system,and provides technical support for the fault diagnosis related research of the UAV air data system.