Design of a Monitoring System for Predictive Maintenance of an Electromechanical System Using Artificial Intelligence

Abstract

This thesis presents the design and real time implementation of a predictive maintenance system for an electromechanical system using artificial intelligence techniques. The proposed system integrates vibration, temperature, and current sensors connected to an Arduino Uno for real time data acquisition from a single phase asynchronous motor. The acquired data are processed and classified using supervised machine learning algorithms Support Vector Machine (SVM) and Random Forest ( to detect different motor conditions, including overheating, bearing degradation, and normal operation. The trained models are deployed on a Raspberry Pi 5 to enable edge computing and autonomous fault detection without reliance on a central server. The results demonstrate high classification accuracy and system responsiveness, confirming the effectiveness of AI in improving maintenance strategies. This work contributes to the development of intelligent, scalable, and low cost predictive maintenance solutions aligned with Industry 4.0 standards