Fraunhofer Institute for Computer Graphics Research IGD
Fraunhofer Institute for Computer Graphics Research IGD

RISTRA – Rapid Structural Simulation: FEM on GPUs

RISTRA is a GPU-optimized FEM library that enables CAE software providers and users to significantly accelerate their structural simulations.

The simulation of a wheel rim for the RONAL GROUP is up to 93× faster than before.

With our high-performance structural analysis software, RISTRA, CAE software providers and technology partners can specifically enhance existing solutions to offer their customers faster simulations than competitors and clearly differentiate themselves technologically. RISTRA significantly reduces simulation times without compromising accuracy, enabling faster design cycles.

Engineers can also use RISTRA as a standalone solution to benefit directly from its advantages.

Our Solution: GPU-Accelerated FEM Simulation

RISTRA accelerates simulations through the consistent use of GPUs (Graphics Processing Units).

The software is based on the established finite element method and combines it with high-performance GPU-based numerics. This allows simulation times to be drastically reduced without sacrificing result quality.

RISTRA is available both as standalone software and as a library for integration into existing CAE systems, allowing flexible integration into existing development workflows.

Challenge of Structural Simulation

Structural simulation is a core component of modern engineering processes. However, computation times increase significantly for complex models, numerous variants, or iterative optimizations. In practice, the performance of traditional CPU-based FEM solvers is becoming increasingly limiting.

Simulation results are often only available after hours or days, significantly slowing down design and decision-making processes. As a result, simulation is often used as a downstream step rather than accompanying the design process.

With RISTRA, results are available within a short time. Simulation can be integrated directly into the design process, workflows become more efficient, and design decisions can be validated earlier.

Up to 89× faster: The GPU-accelerated structural mechanics solver RISTRA significantly reduces simulation times.

Use Case

Honda Research Institute Europe

In a joint project with Honda Research Institute Europe, RISTRA was used to accelerate topology optimization. The goal was to analyze the crash behavior of components more efficiently and to explore more design concepts in early development stages. By using RISTRA, computation time per optimization run was reduced from five hours to 45 minutes, significantly shortening simulation times and greatly expanding the design space.

Our Technology and Expertise

RISTRA is based on many years of research at Fraunhofer IGD in the areas of:

  • GPU programming and GPU-accelerated numerics
  • Development of new numerical methods for structural simulation
  • High-performance finite element methods (FEM)
  • Visualization and processing of CAD and simulation data

The core functionalities of the software are fully developed at Fraunhofer IGD and form the basis of RISTRA’s technological uniqueness.

Services

  • Software licensing (standalone or library)
  • Integration into existing CAE system landscapes
  • Development of proof-of-concepts
  • Benchmarking of existing simulation processes
  • Training, workshops, and consulting

Further contents

Interactive simulation in practice: Insights into the software

Expand all Close all

 

Manufacturing and Mobility

We address the entire mobility industry – from development and production to after-sales. Our focus: intelligent, practical solutions for a more efficient industry.

Learn more

Further Literature

Weber, D.; Mueller-Roemer, J. S.; Altenhofen, C.; Fellner, D.: Deformation simulation using cubic finite elements and efficient p-multigrid methods, Computers & Graphics, Vol. 53, Part B, pp. 185–195, 2015.

Mueller-Roemer, J. S.; Stork, A.: GPU-based Polynomial Finite Element Matrix Assembly for Simplex Meshes, Computer Graphics Forum, Vol. 37, No. 7, pp. 443–454, 2018.

Altenhofen, C.; Loosmann, F.; Mueller-Roemer, J. S.; Grasser, T.; Luu, T. H.; Stork, A.: Integrating interactive design and simulation for mass-customized 3D-printed objects – a cup holder example, in: 28th Annual International Solid Freeform Fabrication Symposium – An Additive Manufacturing Conference 2017, Austin, Texas, 2017.

Ströter, D.; Halm, A.; Krispel, U.; Mueller-Roemer, J. S.; Fellner, D.: Integrating GPU-accelerated tetrahedral mesh editing and simulation, Lecture Notes in Networks and Systems, Vol. 601, pp. 24–42, 2023.

Weber, Daniel; Grasser, Tim; Stork, André: Rapid interactive structural analysis, in: 2020 NAFEMS DACH Regional Conference, 4 p., 2020.