AI for Industry: Transforming the Daily Lives of Maintenance Operators
Résumé
In a highly constrained industrial environment, AI is gradually moving into a field where it was not expected: that of maintenance assistance. In the following, we will discuss this low-key innovation, the conditions of its acceptance and the dialogical processes in which it is involved. It is the fruit of several years of industrial cooperation and innovation.
Maintenance support has long been a key area for AI. This applies to expert systems and, more specifically, diagnostic support systems, event memorization and retrieval systems, and trend interpretation systems for predictive maintenance. AI focuses on the exploitation of large masses of data, or on "specialized", targeted expert knowledge, depending on the era, research fields or technological opportunities. Hybrid AI has yet to be developed in this field.
Our innovation, which combines expert knowledge and the latest AI techniques, is a real case study and the fruit of a dynamic R&D program involving several industrial partners: EDF R&D, FRAMATOME, SPIE Nucléaire and TECHNICATOME, Boost Conseil and ASSYSTEM. The results are tangible thanks to a demonstration by SPIX Industry.
Its history is part of a panorama of successful projects, some of which are cited in Mercier-Laurent (Mercier-Laurent 2011), with:
Chapter written by Anne DOURGNON, Eunika MERCIER LAURENT and Alain ANTOINE.