Simulative-experimental approach to the digital twin Efficient multi-scale simulation of material properties with experimentally calibrated models of a glass fiber non-crimp fabric

Multi-scale approach enables a precise and efficient determination of material characteristics of technical textiles with an experimentally calibrated digital twin. The experimental effort is significantly reduced compared to conventional characteristic determination, which saves time and costs in component and process design.

Fiber reinforced polymer (FRP) with textile reinforcement have a significant share in the FRP market. In order to determine the material properties for the simulative component and process design, numerous tests which cover different fiber volume contents (FVC) and structural variations are necessary. A new simulative-experimental multi scale approach reduces the experimental effort and thus time and costs. Based on easy-to-determine input parameters, the material properties of textiles are determined precisely and efficiently using an experimentally calibrated digital twin.

The workflow integrated in GeoDict is used for efficient determination of permeability, which quantifies the impregnation process by a matrix system. The multi scale permeability determination begins with the determination of the micro permeability within the rovings. This is followed separately by meso modeling of individual layers consisting of solid rovings with direction-dependent micro permeabilities. This approach significantly minimizes the computational effort. The individual layers are randomized within a given parameter space in order to map stochastic structural variation. Several individual layers are stacked with random offset and virtually compacted in order to represent realistic roving deformations as well as nesting and ondulation in the digital twin.

A simple FFT solver using the Stokes-Brinkman equation calculates the permeability of the digital twin. The simulation results are compared with experimentally determined values at 50% FVC to calibrate the single layer model. The digital twin can then be virtually compacted to higher FVCs. During the investigation of a glass fiber non crimp fabric "Hacotech G300U 1270" the calibration is carried out at 50% FVC, afterwards the digital twin is compacted to 55% and 60% FVC. The simulated permeabilities show deviations from experimental results of -30% to +42% and -35% to +8% respectively for these FVCs. Thus, the deviations are in the same order of magnitude as in experimental comparative studies.

The digital twin is also used to determine the mechanical properties of the epoxy resin impregnated non crimp fabric. At micro level, the results of a pure resin tensile test are combined with the mechanical properties of the fibers to simulate the properties of a roving and to assume them as homogeneous properties of the solid rovings of the digital twin. At meso level, a tensile test is carried out, both in GeoDict and in laboratory. Thus, deformation and failure can be predicted precisely.

Contact: Martina Hümbert¹, Tim Schmidt², Aaron Widera¹, David May², Nicole Motsch²

¹Math2Market GmbH
Richard-Wagner-Straße 1
67655 Kaiserslautern
martina.huembert@math2market.de
Phone: 0631 205 605 28

²Institut für Verbundwerkstoffe GmbH
Erwin-Schrödinger-Str. 58
67663 Kaiserslautern
tim.schmidt@ivw.uni-kl.de
Phone: 0631 31607 32