FAGOR Arrasate – Anomaly Detection in Servo Drives
FAGOR Arrasate
Sector: Industry
Business Case
At times, the drives in our automation systems deteriorate due to factors like wear, loosening, or changing working conditions. The objective of this project is to establish proper operational cycles based on performance data and identify any unusual cycles.
Objectives
Identify abnormal cycles in our automation systems, with the flexibility to classify failures as they occur.
Use case
Learn how motor torque changes based on operating parameters (strokes, angles, gpm). Notify when a cycle deviates from the standard pattern and evaluate the extent of the deviation.
Infrastructure
Edge Cloud
Technology
Machine learning and deep learning
Data
StrokeData: dataset with normal automation conditions _x000B_DegreeData: dataset with torque evolution_x000B_ValidationData: validation dataset
Resources
Data science team and domain expert.
Difficulties and learning
KPIs (business impact and metrics of the model)
Learn how motor torque changes based on operating parameters (strokes, angles, gpm). Notify when a cycle deviates from the standard pattern and evaluate the extent of the deviation.
Funding
Collaborators, Partners
INETUM BAIC
