The Excel file our client used contained nearly 24,000 simulations, each described by around sixty parameters. This volume made it difficult to access information, find similar cases, or identify parameter ranges that had already been explored.
This structuring also facilitates the complex queries required for our client's studies, particularly to quickly answer technical questions or validate hypotheses. Using a simple query language allows querying all simulations, filtering according to business criteria, and immediately obtaining the necessary results. The goal was to provide our client with a reliable and coherent model to leverage existing data and prepare for future analyses. This database will serve as a foundation for advanced design support tools and guidance for simulation campaigns.
We proposed a relational database to structure this dataset, avoid redundancy, and improve the consultability of the historical data. The new architecture aligns the data around several business entities: material characteristics, simulation parameters, associated projects, and results (stress, temperatures, spacing, etc.).
