PSI - Issue 75

Laurent Dastugue et al. / Procedia Structural Integrity 75 (2025) 334–343 Laurent Dastugue et al. / Structural Integrity Procedia (2025)

342

9

• Operate within practical hardware and time constraints, • Deliver faster design iterations, and •

Improve the accuracy of life predictions, particularly in components sensitive to high-frequency loading. This case study of a full truck chassis illustrates how the new approach transforms the landscape of large-scale fatigue simulation. By combining faster runtimes, drastically lower disk usage, and the ability to model higher frequency dynamics, it is a game changer for durability engineers in the automotive industry. It unlocks new levels of fidelity in life prediction, now feasible within realistic project timelines and resource budgets. 7. Summary This work presents a significant advancement in the field of fatigue analysis by introducing a fully integrated simulation framework within the industrial software PERMAS finite element analysis (FEA) environment (table 1). By embedding fatigue evaluation directly into the solver, the traditional reliance on external tools, manual data transfers, and interface conversions — particularly for stress data — is entirely eliminated. This streamlined workflow improves not only user efficiency but also data consistency and computational reliability. A cornerstone of this integration is the unified High-Performance Computing (HPC) database architecture, which serves as the backbone for fast, seamless data handling across simulation phases. Furthermore, material data input has been greatly simplified through automatic S-N curve generation based on the established FKM guideline. At the core of the performance gains is the newly introduced Multi-Level Dynamic Reduction (MLDR) Solver, a breakthrough algorithm that drastically reduces matrix sizes and does away with the transformation matrix, leading to substantial improvements in simulation speed and memory efficiency. This innovation proves particularly impactful for large-scale time history analyses where traditional methods have faced limitations in scalability. The industrial case study on a full Daimler Truck chassis highlighted the practical benefits of the integrated approach. In this large and complex model, the MLDR solver enabled significantly faster runtimes and substantially reduced disk space consumption, all while supporting extended modal spaces necessary for analysing higher frequencies, which are increasingly relevant for electric vehicle platforms.

Table 1. Comparison of classic and full integrated fatigue. Classic Approach

Full Integration

Performance

Slow file data transfer

One common database HPC infrastructure

Accuracy

Limited stress gradient result Limited mesh size

Normal stress gradient at any location of surface Fine meshes possible Reliable process by software developer (maintained) Easy retention Integration compatible with most approaches

Usability

Error-prone Manual process Unmaintained scripts

Fatigue approaches

Many available

Industrial big models

Very limited due disk to and runtime limits Unthinkable in industrial applications due to complexity of the process and runtime

Classes larger models Fits in variant workflow Enabler for industrial applications

Optimization

This robust new software framework not only enhances today’s fatigue simulation capabilities but also lays the foundation for future extensions, including integrated damage optimization for industrial-grade FEA models. With its improved computational efficiency and workflow simplification, the solution positions itself as a powerful tool for engineers tackling the growing complexity of modern vehicle design and durability assessment.