Industrial data management strategy towards an SME-oriented PHM

The fourth industrial revolution is derived from advances in digitization and prognostic and health management (PHM) disciplines to make plants smarter and more efficient. However, an adapted approach for data-driven PHM process implementation in small and medium-sized enterprises (SMEs) has not been yet discussed. This research gap is due to the specificities of SMEs and the lack of documentation. In this paper, we examine existing standards for implementing PHM in the industrial field and discuss the limitations within SMEs. Based on that, a novel strategy to implement a data-driven PHM approach in SMEs is proposed. Accordingly, the data management process and the impact of data quality are reviewed to address some critical data problems in SMEs (e.g., data volume and data accuracy). A first set of simulations was carried out to study the impact of the data volume and percentage of missing data on classification problems in PHM. A general model of the evolution of the results accuracy in function of data volume and missing data is then generated, and an economic data volume notion is proposed for data infrastructure resizing. The proposed strategy and the developed models are then applied to the Scoder enterprise, which is a French SME. The feedback on the first results of this application is reported and discussed.