Abstract
There has been a lot of focus within the Dutch Defence in recent years on smarter maintenance of (weapon) systems. This aligns with the desire for more Information-driven Operations (IGO) and is facilitated by the increasingly presence of sensors and data recorders in (weapon) systems. Additionally, it can help address challenges related to smaller crews on naval vessels, scarcity
of technical personnel, and the desired higher readiness of equipment because of the current concerning developments in the world. Based on the Vision of the Dutch Defence, visions and roadmaps have been developed within various departments for the transition to (more) predictive maintenance. Predictive Maintenance and Logistics are also focal points within the Defence Data Science and AI strategy [1], with an accompanying developed roadmap. However, it has been observed that this transition is a challenging process, requiring the resolution of numerous underlying challenges. Moreover, full Predictive Maintenance is the ultimate goal, which may only be achieved in about 10 years,
but there are also more achievable interim steps, such as automatic fault detection, automated diagnosis, and (on-board) health assessments of systems. This vision resulted in the development of the “Data Driven Smart Maintenance (DDSM)” project. In current literature numerous Artificial Intelligence (AI) models and emerging technologies exhibit promising capabilities in facilitating DDSM [2]. However, the seamless implementation of these AI models and especially Predictive Maintenance (PdM) with real-world data poses inherent chal-
lenges [3]. Tiddens, et al. [4] found that companies often have a high ambition level of preventive maintenance but that this level is not feasible because of, for instance, the lack of data. There is a gap in the current literature regarding the different problems companies have with implementing DDSM. The discussion above shows the need for DDSM but it also shows that there are several problems with the implementation. Currently there is no clear view on what problems occur when companies try to implement DDSM. Especially the big difference between research on DDSM and the implementation of it is concerning. This concern calls for more research towards the problems
and issues there are regarding the implementation of DDSM. After identifying these problems, the focus can be shifted to solving them.
of technical personnel, and the desired higher readiness of equipment because of the current concerning developments in the world. Based on the Vision of the Dutch Defence, visions and roadmaps have been developed within various departments for the transition to (more) predictive maintenance. Predictive Maintenance and Logistics are also focal points within the Defence Data Science and AI strategy [1], with an accompanying developed roadmap. However, it has been observed that this transition is a challenging process, requiring the resolution of numerous underlying challenges. Moreover, full Predictive Maintenance is the ultimate goal, which may only be achieved in about 10 years,
but there are also more achievable interim steps, such as automatic fault detection, automated diagnosis, and (on-board) health assessments of systems. This vision resulted in the development of the “Data Driven Smart Maintenance (DDSM)” project. In current literature numerous Artificial Intelligence (AI) models and emerging technologies exhibit promising capabilities in facilitating DDSM [2]. However, the seamless implementation of these AI models and especially Predictive Maintenance (PdM) with real-world data poses inherent chal-
lenges [3]. Tiddens, et al. [4] found that companies often have a high ambition level of preventive maintenance but that this level is not feasible because of, for instance, the lack of data. There is a gap in the current literature regarding the different problems companies have with implementing DDSM. The discussion above shows the need for DDSM but it also shows that there are several problems with the implementation. Currently there is no clear view on what problems occur when companies try to implement DDSM. Especially the big difference between research on DDSM and the implementation of it is concerning. This concern calls for more research towards the problems
and issues there are regarding the implementation of DDSM. After identifying these problems, the focus can be shifted to solving them.
| Original language | English |
|---|---|
| Number of pages | 31 |
| Publication status | Published - 28 Jan 2025 |