AI-based Predictive Maintenance – Reality or Fiction if you don’t have Failure Data?

Remi Duquette (Maya HTT Ltd)

Abstract:

At Maya HTT, we have a highly skilled team of professionals, including System Integration Architects, Process Engineers, Mechnical Engineers, Industrial Data Engineers, Industrial Data Scientists, AI Scientists, ML-Ops engineers, and both Front-end and Back-end software developers. The presentation will explore the challenges and possibilities of implementing predictive maintenance using a combination of artificial intelligence and synthetic simulation of onset of failures when failure data is limited or absent in the real-time operations. The presentation addresses how AI models rely on historical data for accurate predictions, and examines alternative strategies, data sources, and techniques, such as synthetic simulation of onset of failure data and transfer learning to overcome this barrier and achieve practical results in industrial settings. Some practical implementations and use cases will be shown.