Aluminum foundry
Integration and development of an advanced analytics platform
Starting point
- Aluminum wheels are produced using the low-pressure die casting process, and after casting, they undergo an X-ray inspection
- During the casting process, around 130 sensor data points are recorded at 300 timestamps per wheel produced
- Currently, the sensor data is not used for error analysis or for the earliest possible removal of defective products
Procedure
- Explorative analysis and visualization of sensor data
- Transformation and statistical analysis of sensor and casting performance data
- Development of multiple use cases and scenarios for testing on a casting machine
- Designing a hybrid data architecture for persisting sensor data
- Developing a dashboard for visualizing various key metrics
Features/Project outcome
- Tableau dashboard for visualizing various key metrics and evaluations
- R/Rshiny analytics platform
- Big data integration
- Evaluation of different error models
- Customer-specific training offerings
- Independent analysis and visualization of results by the specialist departments using Tableau
Customer benefits
- Reduction of defective wheels through new opportunities to optimize the casting process
- Reduction of inspection costs through error prediction immediately after the casting process
- Visualisation and analysis of casting performance for casting machines, wheel types, and production shifts
- Daily retrospective of casting performance for further root cause analysis of errors
- Data discovery to identify which sensors are relevant for production errors

Erich Holzinger | Senior Manager / Authorised Officer