Fault diagnosis with matrix analysis for electrically actuated unmanned aerial vehicles

Because of their large operational potential, unmanned aerial vehicles (UAVs) may be required to perform over long periods of time, which might lead to potential degradation or even failure of their electrical or/and mechanical control Surfaces and components. Consequently, the least failure can degrade the performance of the process and might lead to a catastrophic event. Therefore, an efficient mechanism should be capable of making these faults realizable and act accordingly so that a performance index is continuously maintained. However, even when a fault is detected at the monitoring phase, as illustrated in a previous work by Kladis and Economou [1], it is important to determine the cause or combination of causes of that failure through a root-event analysis in order to proceed to its elimination from further expansion. In addition, if the system is not recoverable as in the case of a UAV crashing at a remote area, then post analysis for diagnosing the possible cause(s) of the catastrophic failure is carried Out in order to improve reliability issues. Hence prior knowledge is cumulated in order to define, identify, and evaluate the risk and perform decision making so as to improve future missions. Among many interesting methodologies used for a qualitative and quantitative analysis for fault assessment fault-tree analysis has been widely used. Particularly for complex systems actuators, control surfaces and components are associated in a complicated manner and all signals are gathered in a supervisory control panel. Hence the determination of a cause of a potential failure can be obtained through multiple different investigations of checkpoints. Hence the problem becomes twofold. Not only is it necessary to acquire the cause of a fault but also its investigation Should be performed in the most efficient manner of minimum sense. In this article potential causes that lead to a failure are investigated through a digraph analysis utilizing graph theory tools and pseudo-boolean expressions, motivated by a previous work [2]. The benefits Of Such a formulation afford the ability of handling system interdependencies, redundancies, determining primary causes of failure (cut sets), critical importance of these, and so on. A generic scenario for fault assessment in a UAV application illustrates the methodology proposed. In addition, a quantitative and qualitative analysis is also included, thus describing reliability aspects of the system. The resulting time optimum solution is obtained by the use of shortest path algorithms, due to the means by which the problem is posed and its similarity to routing problems. Then during quantitative analysis, algorithms for determining all fault modes are used. In addition, mathematical formulation and its relevance are also addressed.