Exposé

• Titre : Dynamic prognostics and health management
• Intervenant : Jie LIU (CentraleSupelec)

Résumé : Prognostics and health management (PHM), gaining much attention of the researchers and industrial practitioners, aims at detect the fault, isolate the fault and failure prognostics of the industrial equipment. In the presentation, we focus on prognostics, i.e. predicting the Remaining Useful Life (RUL) during which the equipment can still carry out the assigned task. PHM has been successfully used in many domains. Recent research shows that the fixed prognostic  models may not work well for online prognostics, especially in a nonstationary environment and dynamic operation condition. The prognostic models need to be dynamic to adapt to the change in the environment and operation condition. This presentation focuses on data-driven approaches and physics-of-failure-based approaches for prognostics and the strategies to update the model. Two real problems from the industry are considered to show the effectiveness of the introduced strategies.