Prognostics on energy efficiency performance for maintenance decision-making: Application to industrial platform TELMA

Industrial plants are facing today with new challenges to well optimize performance of their operation and maintenance. For example, the sustainability paradigm is introducing new requirements to be taken into account in the decision-making process. In that way, energy consumption (EC) and energy efficiency (EE) are two critical performances impacting severely the plant effectiveness mainly with regards to its life cycle cost. Although there are models for following these two performances at the component level, there is a real need for modelling them at the function or system levels not only to support strategic decisions (and not only operational one) but also to forecast them to make decisions in advance for better optimization. Thus, the principles of a generic approach, which is focused on EE performance (EEP) and built on the modelling of this EEP at functional level, and its prognostics to calculate a Remaining Energy-Efficient Lifetime (REEL), are proposed in this paper. The REEL should integrate future mission profiles and operation conditions. The prognostics model is developed from a data-driven approach by using a nonlinear regression method. This generic approach is instantiated and validated on the TELMA platform (a motor-driven system) which is simulating a real industrial plant addressing unwinding metal bobbins. So, models are built from field data of two independent motors (the component level and electrical energy) in addition to data on the function supported by means of these two motors. It leads to prognostics models usable to predict the EE evolution - REEL (the input of the decision-making module) both at the component level from the relationships between speed performance (motor output), bearing deterioration (Gamma process) and EE, and at the functional level from the relationships between productivity performance (functional output on the product delivered), components deterioration level and EE.