Magna is a leading global automotive supplier with operations in 28 countries and a workforce of 164,000 employees. They are dedicated to innovating and delivering reliable, high-quality products that support mobility for all.

Unplanned machinery downtime is expensive. Critical machinery stopping can lead to delays in production, schedule overhauls, and unexpected maintenance. Continuous monitoring of machinery produces early warning signals, cuts cost, sustains production and extends equipment life.

Our ML/AI Pipeline for Condition-Based Maintenance 2.0 uses real sensor readings connected to machinery to detect early warning signals. The 2.0 system expands monitoring beyond vibration alone by adding audio signals and clearer, faster alerts for operators.

Each machine is equipped with a dedicated audio and vibration sensor that captures data. The software continuously monitors the machine’s behavior and flags patterns that suggest developing issues. When readings deviate from a normal baseline, the dashboard highlights the machine and sends an alert. Our system translates complex signals into simple operator views. A dashboard shows clear health status indicators, live trends and alerts so technicians can decide when to inspect or service equipment.

The system studies vibration and audio measurements to capture subtle changes indicating mounting problems. Machine learning models analyze signals to predict when a machine failure is going to occur before it happens, saving money and reducing unexpected maintenance.

The front end is built on Django, while the back end uses Python and PostgreSQL for data storage. The pipeline runs in Python with ROS 2, uses MQTT for publishing and subscribing message payloads, and is containerized with Docker for deployment on an NVIDIA Jetson.