Abstract
Predictive maintenance (PdM) allows the prediction of early failures of medical equipment before they occur. It helps to diagnose the defaults of critical equipment in a hospital structure, namely MRI. Founded on the analysis of data collected in real time of the right parameters, thanks to intelligent sensors positioned on the equipment, using Internet of Things (IoT) technology and the practice of machine learning tools. The objective of this techniques is the implementation of algorithms capable to predict an anomaly, which will make equipment and maintenance tools increasingly autonomous and intelligent. Therefore, the idea of this project is to develop a wireless sensor network to ensure continuous monitoring of the state of MRI. The implemented solution includes an IoT monitoring system of the cold head’s cooling circuit. Based on the vibrations at the pump, it allows to monitor the motor circuit, inform the staff at each abnormal state of this system, and protect this device against any future anomalies. Thanks to the CNN algorithm implemented in this solution, the results are very satisfactory, with an accuracy >98%. This solution can be integrated into a general predictive maintenance solution for the most sensitive equipment in a hospital.