¹û¶³´«Ã½

from the research group Human Performance Management (HPM) defended his PhD on April 15 with the thesis titled ‘Human-Centric Predictive Maintenance’. This research addresses the societal relevance of predictive maintenance (PdM) and how this technology can be optimized to increase employee acceptance and successful adoption in industry.

Machine Failures Increasingly a Problem

In recent years, the complexity and costs of machines in various industries have increased. As a result, unexpected machine failures have become more problematic, as the downtime caused by these failures is very costly. Predictive maintenance offers a technological solution by using sensors to monitor the health of machines and applying statistical models to estimate their remaining lifespan. This enables organizations to propose maintenance actions just before a failure would occur, reducing unnecessary preventive actions and the waste of healthy components.

Challenges and Solutions

Despite the promise of PdM, many organizations struggle to implement these systems successfully. A significant part of the challenge is that employees often hesitate to trust or follow the recommendations of PdM technologies. Van Oudenhoven's dissertation investigates why employees accept or reject PdM systems and how their design can be improved to increase acceptance. A human-centric approach is employed, focusing on the needs and perspectives of employees.

Trust and Presentation

The research begins with an overview of existing literature on how employees respond to data-driven decision support tools. It identifies key factors influencing acceptance, such as trust in the system and how decisions are shared between technology and employees. Interviews with maintenance personnel reveal that while employees hope PdM will make their work easier, they also worry about losing their skills and autonomy due to the introduction of automated systems.

Van Oudenhoven further investigates how the presentation of predictions—whether as ranges or specific numbers—affects employees' trust in PdM solutions. It turns out that predictions presented as ranges are perceived as more accurate than specific numbers, especially when employees are asked to focus on the system's accuracy.

Impact of False Alarms

Additionally, the dissertation examines the impact of false alarms, or incorrect recommendations from PdM systems. It shows that unnecessary actions can diminish employees' trust in future recommendations. However, if a recommendation is accompanied by a high probability of machine failure, employees are more likely to overlook past errors.

The research demonstrates how the implementation of predictive maintenance can influence employee behavior at work. It sheds light on the factors that can improve employees' acceptance of and trust in technology, ultimately aiming to create better-designed PdM systems that support employees in their work.

Bas van Oudenhoven defended his thesis on April 15, 2025. Title of the thesis: ‘’. Supervisors: Rob Basten, Evangelia Demerouti, and Philippe van de Calseyde.