Instruments & industry / Project story
Predictive maintenance system
My work at SEW-EURODRIVE connected data acquisition, ML-assisted analysis and maintenance interfaces within a platform using neural networks.
01 / Context
Context
At SEW-EURODRIVE, I worked on a predictive maintenance platform for industrial gear units. The system uses vibration, temperature, oil and load data alongside neural networks to identify unusual patterns that can inform maintenance decisions.
02 / Requirements
Requirements
A model finding is useful only when a maintenance team can interpret it and fit it into an operational decision.
03 / Implementation
Implementation
I developed components for data acquisition, analysis logic and the operator-facing application. Dashboards, trends and configurable alerts bring model findings into the workflows of maintenance teams.
The platform combines C#, Angular and Python with cloud services and existing production systems. My work covered several stages of that path, from acquiring data to making analysis results understandable in the application.
04 / Outcome and project status
Outcome and project status
Dashboards, trends, and configurable alerts bring the analysis into the application used by maintenance teams. My contribution covered acquisition, analysis logic, and the operator-facing application.
Project images