Healthy-State and Time-To-Failure Assessment of Induction Motors and Pumps for Maintenance Scheduling Using Cloud-Enabled Fault Prediction and Reporting System with Machine Learning Scheme

Abstract

Every industry relies heavily on electrical machines. When these machines fail without giving a signal or warning, there is unplanned downtime and unanticipated deployment of maintenance staff which may affect or lower production if machines are serving as important assets. Some failures can lead to greater damage, instead of one machine or item being replaced a lot of resources will be needed to repair once a machine fails and assets will have a shorter life expectancy. There are also safety issues associated with running a machine to failure. Running a machine to failure also increases labour costs and safety costs. Therefore, this paper proposes a Cloud-Enabled Fault Prediction and Reporting System with Machine scheme for Healthy-State and Time-To-Failure Assessment of Induction Motors and Pumps for Maintenance Scheduling. This system uses sensors to extract data (current, speed, temperature, vibrations). The extracted data is then conditioned and compared to the values related to the healthy state of the equipment using machine learning algorithms, manufacturer specifications, and historical readings to identify performance non-conformities or items that require action and display the asset health report which will help schedule a proper maintenance procedure based on analysis.