Country : Italy
Client : Michele Cortina, Ai.ON
This project involved creating an anomaly detection model to detect the condition of the gear box in large wind turbines in a wind farm.
The key condition was that only the existing PLC data could be used by the model. No extra equipment or sensors could be installed. The only sensor in the gearbox was a temperature sensor.
With three years of operational data and the maintenance records, we were able to successfully create a set of features that could train a machine learning model to indirectly detect the condition of the gearbox and give indications before a failure occurred.
“Richard immediately understood what to do and went straight to the goal. He knows very well the world of industry 4.0 and sensors’ signals and machine learning techniques and tools to work on them. Highly recommended”
★★★★★