Predictive maintenance (PdM) is to day considered as an important maintenance strategy to achieve high reliability at a low cost. In contrast to traditional calendar based maintenance the main idea of a predictive maintenance strategy is to utilize component condition, future loads, and opportunity windows to determine a “just in time” plan for maintenance.
PdM is defined as condition-based maintenance carried out following a forecast derived from repeated analysis of known characteristics and evaluation of the significant parameters for degradation of the item (EN 13306:2017).
The era of digitalization taking place these dates is transforming the way systems are operated and maintained as well as the working processes needed to support modern production, transportation, public services and so on.
A leading initiative is the German government Industrie 4.0 strategy for manufacturing industries where a goals is, among others, real-time synchronization of maintenance and operations taking logistic constraints into account.
WP4 under FME Northwind deals with digital twins and asset management. When it comes to maintenance the objective is to develop digital twin (DT) prototypes to realize predictive maintenance strategies.
To accomplish this, a unified real-time framework for integration of weather windows, production profiles, logistic resources, and condition/remaining useful lifetime (RUL) is missing. Hybrid uncertainty models based on physics of failures and data-driven approaches are required.
Predictive maintenance is investigated in many PhD and Post Doc. projects at NTNU. The most important initiatives are: BRU21, SFI-SUBPRO in addition to the FME-Northwind project described here.