Open-Source Predictive Maintenance System for AC Motors Using Vibration and Temperature Analysis

This work presents the design and experimental validation of a low-cost open-source predictive maintenance system for AC induction motors based on vibration and temperature analysis. The proposed platform integrates wireless sensing nodes built with ESP8266 microcontrollers, MEMS accelerometers, and temperature sensors. Data collected from industrial motors are processed using LSTM neural networks to detect abnormal operating conditions. Experimental validation using controlled mechanical imbalance demonstrates anomaly detection accuracy exceeding 80%.