Simulation methods: time-saving mapping of complex processes

Rigid specifications rarely fit agile processes - what applies to organizations also applies to simulation methods. If complex processes such as aquaplaning or the machining of metal are to be mapped virtually, it is not possible to predict all the movements of the components in advance and create them in a suitable calculation grid, as is usually used for simulations.

Meshfree replaces the real test

This was the challenge faced by a research group at the Fraunhofer ITWM more than 20 years ago. "Our very first task in the project team was to simulate the deployment of an airbag during a vehicle crash," reports Dr. Jörg Kuhnert, who was already part of the group at the time. "Apart from cost-intensive crash tests carried out in real life, there was no way of quickly testing the safety of new developments in this area at the time." This is because the more objects that move and interact with each other in a situation, the more difficult it is to reliably reproduce them at a reasonable cost using conventional simulation methods.

Based on Jörg Kuhnert's dissertation, the team developed the grid-free approach - since 2012 also with the collaboration of Dr. Isabel Michel in the area of free jet turbines. In some cases, this makes it possible for the first time to show particularly complex and dynamic situations in the simulation. All of the research results achieved since then have been incorporated into the Meshfree software tool. The result is a simulation tool with a truly unique selling point: no other simulation tool in the world makes the Generalized Finite Difference Method (GFDM) available for industrial use.

Flexible method for dynamic processes

Traditionally, the finite element method is used for simulations: engineers construct a grid suitable for the respective geometry and calculate the changes in each individual element based on this. Even setting up the grid structure is time-consuming; it also has to be repeatedly adapted during the simulation. In contrast, the Meshfree software combines the generalized finite difference method for solving the conservation equations for mass, momentum and energy with efficient algorithms for solving linear systems of equations, which were co-developed by the Fraunhofer Institute for Algorithms and Scientific Computing SCAI - a huge advantage, as the numerical point cloud used can adapt flexibly to moving geometries. There is no need for time-consuming corrections in the computational grid. Dr. Jörg Kuhnert and Dr. Isabel Michel have been awarded the Joseph von Fraunhofer Prize 2024 for their development, which can replace real experiments.

From automotive to process engineering - and beyond

The excellent methodology can be used for a wide range of applications. One current focus is in the automotive sector: in addition to airbag simulation, the researchers have so far been able to support their industrial partners with modeling water crossings or the behaviour of vehicles on sand or gravel, among other things. In process engineering, Meshfree has helped companies to optimize process parameters in the processing of molten glass and the production of plastic parts.

In principle, the method can be used wherever measurements or experiments need to be replaced or only work poorly or not at all. Isabel Michel summarizes: "We are not fixated on the classic applications of numerical fluid mechanics. Meshfree can do much more: the tool is deliberately kept generic." The software has great potential to save costs, time and materials in many other fields of application in the future.