IoT-Based Predictive Condition Monitoring System for CNC Machines Using MATLAB

This project presented the design and implementation of an IoT-based predictive
condition monitoring system for CNC machines, specifically the MTAB XLTurn lathe.
CNC machines are prone to failures due to excessive vibration, temperature rise, and
misalignment, which can lead to reduced productivity and increased maintenance costs.
The objective of this work was to develop a cost-effective system capable of real-time
monitoring, early fault detection, and estimation of machine health.

The system utilized an ESP32 microcontroller integrated with ADXL345 (vibration),
LM35 (temperature), and MPU6050 (tilt) sensors for data acquisition. Sensor data was
transmitted to MATLAB through serial communication, where signal processing
techniques such as Root Mean Square (RMS) and Fast Fourier Transform (FFT) were
applied for fault detection and analysis. A multi-sensor threshold-based approach was
implemented to reduce false alarms, and Remaining Useful Life (RUL) was estimated
based on degradation trends. ThingSpeak was used for real-time data visualization, and
Twilio was used for alert notifications.

The results demonstrated accurate fault detection, improved monitoring efficiency, and
reduced false alarms compared to single-parameter systems. The proposed system
provided a scalable and economical solution for predictive maintenance in industrial
applications.