Answers are grounded in this paper and cite their sources.
Predictive maintenance has been evolving with increasing adoption of data analytics in the variable frequency drives. Predictive maintenance assisted by artificial intelligence and machine learning (AI/ML) will be the key to reliable drives operation in the future. With a common framework for various components and consistent plant-wide integration, systems with AI/ML based predictive models can dynamically learn and adapt to the drive usage patterns and predict the remaining useful lifetime. This paper reviews the practices and challenges of modern predictive maintenance paradigm. A real use case for Metal Oxide Semiconductor Field-Effect Transistor (MOSFET) condition monitoring is adopted to demonstrate how AI/ML can be utilized in predictive maintenance, including probability based prediction, data compression, pattern clustering and decision-making. Furthermore, a unified probability and statistics based approach is proposed for fault region determination and decision-making based on Gaussian distribution and Gaussian Mixture Model (GMM).
No takes yet. Share an insight, caveat, or question.
Jaremko et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: