Embedded Cognitive Bayesian Prognostic Health Management for RPA Propulsion Systems

ABSTRACT: The Air Force is facing severe budget cuts which force the need to cut operating costs of aircraft and manpower. One way to do this is to increase reliability of systems and components which would reduce maintenance and increase operational availability, which reduces the number of aircraft needed to complete missions. The Air Force desires a new prognostic health management (PHM) approach that applies advanced diagnostic and prognostic algorithms to systems and components of vehicles operated under harsh or off-design conditions typical of military platforms. In Phase I, MSI developed and prototyped a PHM system that will provide accurate prognosis of propulsion system health and, as much as possible, preempt loss of control, in-flight shutdowns and cut down on false removals and maintenance time caused by inadequate information from the aircraft. The demonstration included real time data collection, signal conditioning, cognitive processing, and data offload from the EPIC; and probabilistic decision support information for maintenance troubleshooting assists. The RTS operating engineer flew simulated missions and seeded faults that typical of hard to isolate problems experienced in RPA propulsion systems. All the seeded faults introduced by the Rolls-Royce RTS engineer were successfully identified. BENEFIT: The COSMOS PHMS three tier approach, implemented with passive embedded processing nodes offers cost effective advantages that go beyond data recording to provide in-flight diagnostics and prognostics and also support O-level troubleshooting and data mining of fleet data. Being on-board and real time COSMOS can programmed to alert the maintainers not only of the diagnosed cause of a red tag events but also predict probabilistically the progress of degradation of components to indicate those what actions can be taken to preempt future failures. The probabilistic information from the COSMOS PHMS will also provide actionable decision support during maintenance by providing diagnostics and backup data. COSMOS will provide root cause analyses to wiring electrical or other cause which will reduce No Fault Found. COSMOS PHMS will improve existing logistic support processes by providing information not currently available. All this translates to reduced maintenance time, lower operating costs and higher operational availability.