A Dual-Microcontroller IoT-Based Real-Time Monitoring System for Predictive Maintenance of Induction Motors

Induction motors are critical to industrial operations but are prone to mechanical and electrical failures such as bearing wear, overheating, and current imbalance, which can lead to costly downtime. Traditional maintenance approaches often lack real-time insight, while existing IoT-based systems are either cost-prohibitive or insufficiently validated for industrial deployment. This paper presents a low-cost dual-microcontroller Internet-of-Things (IoT) platform that delivers real-time, multi-parameter condition monitoring to support predictive maintenance of three-phase IMs. An STM32F103C8T6 microcontroller acquires line voltage, phase current, surface temperature, and tri-axial vibration, while an ESP32-WROOM-32 transmits 10 Hz data packets to the Blynk cloud and issues Telegram alerts. A 58.9 Hz RC low-pass filter suppresses inverter harmonics, yielding mean absolute errors below 1 V for voltage, 2.5% FS for current, and 1.3° C for temperature. End-to-end latency averages 402 ms, well within supervisory-control requirements. In a 5-h validation, the system discriminated normal motors from those with rear-bearing damage by a rise in total vibration velocity from 0.36 mm/s to 0.94 mm/s (ISO 10816-3 $\text{limit} =0.71 ~\text{mm} / \mathrm{s}$). The complete bill of materials is under USD 35, and the firmware footprint is 57 kB, demonstrating a scalable route to condition-based maintenance in small- and medium-sized plants.