Case study: anomaly detection in engineering data for an aviation manufacturer
- −40%analysis time
- 1000+measurements/day
Context
Structural element condition monitoring was based on weight measurement data. As production volume grew, manual oscillogram analysis was no longer scalable.
Task
Develop a system for detecting anomalous patterns in time series, enabling early-stage defect detection.
Implementation
- Multi-channel oscillogram preprocessing and normalization
- Research and testing of multiple anomaly detection approaches
- Quality criteria setup and validation procedure
- Production module development for integration into existing infrastructure
Processing handled thousands of measurements daily.
Results
- Data analysis time reduced by more than 40%
- Reduced dependency on manual expert review
- Solution deployed in production across multiple sites, including international ones
The system is embedded in the manufacturing process and used as a regular diagnostics tool.