A generic diagnosis and prognosis framework: application to permanent magnets synchronous machines

Garance Vinson (1) , Pauline Ribot (1) , Thomas Prado (2) , Michel Combacau (1)

1 LAAS-DISCO - Équipe DIagnostic, Supervision et COnduite
2 Messier-Bugatti-Dowty

In the context of electrification of aeronautical actuation systems, aircrafts manufacturers intend to more and more supervise the health state of equipment during their service life. The objective is to increase the operational availability of systems, and to organize efficient maintenance management for safety and economic reasons. In this purpose diagnosis and prognosis are performed to determine the current and future health state of mechanical, electrical, and hydraulic equipment. A generic modeling framework is first proposed. It is made of two parts. First a structural and functional model enables representing nominal, faulty, abnormal, or unknown behavior of a complex system. It also describes the functional interaction between its components. Then an ageing model allows representing the system degradation through time thanks to the knowledge available about its ageing, from physical laws or experience. It also shows the effect of damaged components on others, and the propagation of ageing-related performance losses through the system. The evolution of the machine health state in nominal case and during faults initiation and progression is modeled. A health assessment method is then proposed based on this generic modeling framework. It is meant to evaluate the health state of any complex system at any time of its life. This method comprises two modules. The detection and diagnosis module assesses the system current health state. The prognosis module assesses the system future health state and computes the remaining time before failure. Permanent Magnet Synchronous Machines (PMSMs) are used as an illustration. These machines are used in crucial applications including landing gears systems, Messier-Bugatti-Dowty core business. They are subject to critical progressive faults such as short-circuits and rotor demagnetization that are successfully modeled by the proposed framework. The health assessment method is then implemented on healthy and faulty simulated PMSMs and developed algorithms succeed in assessing current and future health state of the machines.

Domaines: Automatique / Robotique

Garance Vinson, Pauline Ribot, Thomas Prado, Michel Combacau. A generic diagnosis and prognosis framework: application to permanent magnets synchronous machines. Prognostics and System Health Management Conference 2013, Sep 2013, Milan, Italy. pp.1039-1044. ⟨hal-00925514⟩