Use of COTS functional analysis software as an IVHM design tool for detection and isolation of UAV fuel system faults

Abstract

This paper presents a new approach to the development of health management solutions which can be applied to both new and legacy platforms during the conceptual design phase. The approach involves the qualitative functional modelling of a system in order to perform an Integrated Vehicle Health Management (IVHM) design – the placement of sensors and the diagnostic rules to be used in interrogating their output. The qualitative functional analysis was chosen as a route for early assessment of failures in complex systems. Functional models of system components are required for capturing the available system knowledge used during various stages of system and IVHM design. MADe™ (Maintenance Aware Design environment), a COTS software tool developed by PHM Technology, was used for the health management design. A model has been built incorporating the failure diagrams of five failure modes for five different components of a UAV fuel system. Thus an inherent health management solution for the system and the optimised sensor set solution have been defined. The automatically generated sensor set solution also contains a diagnostic rule set, which was validated on the fuel rig for different operation modes taking into account the predicted fault detection/isolation and ambiguity group coefficients. It was concluded that when using functional modelling, the IVHM design and the actual system design cannot be done in isolation. The functional approach requires permanent input from the system designer and reliability engineers in order to construct a functional model that will qualitatively represent the real system. In other words, the physical insight should not be isolated from the failure phenomena and the diagnostic analysis tools should be able to adequately capture the experience bases. This approach has been verified on a laboratory bench top test rig which can simulate a range of possible fuel system faults. The rig is fully instrumented in order to allow benchmarking of various sensing solution for fault detection/isolation that were identified using functional analysis.