Requirements for Numeric Models as Sources of Synthetic Data for Predicting Real-World Data Sets in Sheet Metal Forming Processes

Schumann, Markus; Moske, Jonas; Wüst, Antonia; Kersting, Kristian; Groche, Peter

Open Access

Requirements for Numeric Models as Sources of Synthetic Data for Predicting Real-World Data Sets in Sheet Metal Forming Processes

Files

Requirements_for_Numeric_Models_as_Source_of_Synthetic_Data_DATASET.zip (2.92 GB)

Date

2026-03-18

Type

Dataset

Authors

Schumann, Markus

Moske, Jonas

Wüst, Antonia

Kersting, Kristian
Groche, Peter

Description

In the field of forming technology, synthetic data generated by finite element (FE) simulations is increasingly being used to train machine learning (ML) models for quality prediction. However, the predictive accuracy on real-world process data is often limited by the so-called “reality gap” between simulated and measured signals. This study investigates how simulation model complexity influences the suitability of synthetic data for training ML models that generalize to real progressive deep drawing processes, as demonstrated using an symmetric deep-drawn part. Three representative simulation configurations of increasing complexity (L1–L3) are implemented and evaluated against experimental data collected under production conditions using sensor-integrated tools. The analysis covers multiple ML tasks, including classification of pre-process connector cut geometries, detection of process disturbances, separation of subtle geometry variants, assessment of feature transfer robustness, and saliency-based interpretation of signal regions. The results show that simple models (L1) enable robust classification of failures such as material damage within synthetic domains though their transferability to real data is restricted. Although the are computationally more expensive, higher complexity level (L2 and L3) better capture the effects of pre-processing and deformation-history, improving domain alignment and supporting physically meaningful model interpretation. Saliency map analysis reveals that models trained on synthetic data emphasize different signal regions than models trained on real data. This underscores the importance of task-relevant signal fidelity. The findings provide quantitative guidance for selecting an adequate level of simulation complexity in the context of manufacturing processes in the sheet metal working industry and demonstrate that the underlying principles apply across a broader spectrum of modelling configurations. Measurement is always the punch force for the deep drawing state. Time series from simulation and real world experiments. The folder structure is as follows: force curves deepdrawing simulation levels - Level 1 Simulation - Raw Data (CSV) - Augmentation (CSV) - Level 2 Simulation - Raw Data (CSV) - Augmentation (CSV) - Level 3 Simulation - Raw Data (CSV) - Augmentation (CSV) Real force curves 80 SPM - Geometry 1.tdms - Geometry 2.tdms - Geometry 3.tdms - Geometry reference.tdms - channel names.txt Simulation force curves of damaged cups - Raw Data (CSV) - Augmentation (CSV) Simultion force curves of excentric cups - Raw Data (CSV) - Augmentation (CSV)