Maintenance Decision Support Procedures Based on Machine Learning

In a competitive production environment, a manufacturing company must have plans to improve production performance. To improve production performances depends on various issues such as production efficiency and machine availability. Various preventive maintenance procedures have been developed for efficiently maintaining and repairing machines and equipment in a manufacturing system to maximize machine availability and its readiness for production. In recent years, Artificial Intelligence (AI) technology has been applied in developing maintenance procedures in industries utilizing advanced information technology such as the Internet of Things (IoT). This thesis presents a machine learning model to predict machine failures and maintenance requirements for certain industrial machine tools. Machine learning methods enable manufacturing systems to make smart decisions through communications with humans and machines using sensors. A Logistic regression model is developed in this study to predict machine failures for the purposes of avoiding machine breakdowns and improving system performance. The supervised classification method was incorporated in the developed prediction model. The developed model is tested verified with realistic machine maintenance data. Computational experiments are conducted with results analyzed