Management of complex systems requires analysis and fusion of a diverse type and quantities of data and models of subsystems to provide inference on the state of health and decisions on maintenance actions that need to be taken. The data and models are often held in distributed environments and so a service oriented framework can be used to integrate them with the analysis methods. The data analytics tools suite will include intelligent signal processing based novelty detection, fault diagnosis, remaining life time estimation and learning for decision making. The decision making tools will embody methods for robust optimization and architecture selection that involve multiple criteria such as cost, performance and other functional specifications. The projects are carried out in the RR-UTC and in collaboration with other University of Sheffield Departments.
The aim of this project is to develop a distributed decision support framework and associated analytics tools that can be operated in an autonomous manner or be used interactively to discover and identify hitherto unidentified faults, remedial actions and modify systems design. It will build on the group’s track record in developing such a framework for complex systems such as management of a fleet of aircraft and analysis of the impact of joint replacement in humans. The framework and analytics tools will be based on an array of techniques spanning particle filtering, case based reasoning, multiobjective optimization, statistical inference and cloud computing.