Towards a Neuro-Fuzzy System for time series forecasting in Maintenance Applications.

International audienceIn maintenance field, industrial and research communities take a growing interest in the "undesstanding of the degradation phenomenon". Within this frame, the general purpose of the paper is to explore the way of defining a prognostic system that is able to approximate and predict the degradation of equipment. Prognostic is defined and the evolution of developments on forecasting methods is studied. A neuro-fuzzy system for failure prediciton based on the ANFIS model is closely studied and a pre-treatment of data is proposed to perform short term accurate and mid-term reliable predictions