Dutch Prognostics Lab
The Dutch Prognostics Lab is a project initiated by the Dutch Ministry of Defense. It started at the beginning of 2024 and ends at the end of 2025.
The project is carried out by:
Luc Keizers | University of Twente | l.keizers@isa-phm.com
Wouter van Riel | infralytics | wouter.van.riel@infralytics.org
Tiedo Tinga | University of Twente, Ministry of Defense | t.tinga@mindef.nl
Project goal
The goal of the Dutch Prognostics Lab is to uniform data registration for predictive maintenance. This aims to ease data exchange between involved parties, accelerating the development of predictive maintenance.
Involved parties
The following parties are involved.
Project content
Prognostics and Health Management (PHM) is a growing research field, aiming to improve reliability and availability of systems. Especially data-driven diagnostic and prognostic algorithms gained a lot of research attention over the last decades: more and more sensor data is being collected from systems in Industry 4.0, and the idea is that these data show recognizable patterns when systems are close to failures. By training diagnostic and prognostic algorithms on historical data, they can help to detect and predict future failures further in advance.
However, for (safety)-critical assets, preventive maintenance actions are often performed to prevent failures, such that failure data (required to train algorithms) is often unavailable in practical applications. Many algorithms are therefore validated on faults from simulations or experimental set-ups and applications are often tailored to specific data sets.
To train and apply algorithms more universally, it is necessary to collect data in a well-documented and standardized manner. Currently, there is no universally applied data standard for the field of PHM. The aim of the Dutch Prognostic Lab is to propose a data standard, and subsequently promote sharing diagnostic and prognostic data to accelerate development, validation and application of data-driven algorithms.
Data standards are already applied for other research fields. Many fields aim to make data more FAIR, meaning that collected data should be Findable, Accessible, Interoperable and Reusable. One example is the ISA standard, standardizing metadata for life, environmental and biomedical sciences. This ISA standard is community-driven and is coordinated by a working group. Due to its proven concept, this ISA standard is taken as a starting point for the development of a PHM data standard.