Automobile manufacturer
Process transparency through analytical visualization of additive manufacturing
Starting point
- Different sensors monitor each individual layer of the Laser Powder Bed Fusion 3D manufacturing process
- The sensor data is available in various formats and resolutions
- Based on this data, efficient real-time analysis and quality monitoring are not possible
- Without analytical processing of the data, detecting issues and errors during the ongoing process is not feasible
- AI methods for optimization cannot be integrated
Procedure
- Requirement engineering and design of an efficient solution
- Definition of a cloud solution and the tool stack
- Design of an interactive dashboard
- Definition of agile user stories for implementation
- Development of real-time algorithms and data processing
- 3D & 2D visualization of sensor data in Python Dash
- Persistence of sensor data on Amazon S3
Features/Project outcome
- Intuitive visualisation of consistently processed sensor data
- 'Near real-time' validation of production results based on statistical analysis
- Processing, adjusting, and persisting sensor data
- Visualisation of layer-wise inconsistencies based on data science model predictions
- Integration of AI models into the visualizsation
Customer benefits
- Transparency of production results through advanced analytics and visualization of sensor data
- 'Near real-time' quality control for additive manufacturing
- Persisting sensor data for long-term optimisation of production processes and analysis of production outcomes
- AI-driven prediction of production quality

Erich Holzinger | Senior Manager / Authorised Officer