A Systematic Mapping Study on Machine Learning Techniques Applied for Condition Monitoring and Predictive Maintenance in the Manufacturing Sector

Background: Today’s production facilities must be efficient in both manufacturing and maintenance. Efficiency enables the company to maintain required output while reducing effort or costs. With increasing interest process automation Internet of things since Industry 4.0 was introduced, such shop floors are growing complexity. Every component needs continuously monitored, which is basis for predictive maintenance (PdM). To predict when needed, components’ conditions monitored with help a condition monitoring (CM) system. However, this task difficult human employees, as analysis very demanding. overcome this, machine learning (ML) can applied ensure more production. Methods: This paper aims investigate application ML techniques CM PdM sector. For reason, systematic mapping study (SMS) conducted order structure classify current state research identify potential gaps future investigation. Relevant literature considered between January 2011 May 2021. Results: Based on guidelines SMSs previously defined questions, existing publications examined overview domain provided. Conclusions: Techniques reinforcement transfer underrepresented, but increasingly attracting attention. The findings suggest that most promising results belong applications hybrid methods, where set methods combined build powerful model.