CAE Modeling of HPDC Aluminum - A Correlation Study on Material Properties and Simulation Accuracy through experimental validation on a component level

Abstract

This thesis seeks to improve the correlation between the results of physical testing and simulations for a HPDC aluminum component loaded until failure. Physical experimental data was gathered through a three point bending test of cast beams and further validated with tensile tests using coupons from the same casting component. Explicit time-stepping simulations were carried out using LS-DYNA and failure models GISSMO and CrachFEM were used for solid and shell elements. Analysis from physical test results revealed the need to incorporate porosity into the model, both through pore meshing and scaling on different levels. Scaling was justified with representative volume elements. Quantitative improvements were made, significantly reducing error. Two different macro-mapping methods reduced the error both in peak force, displacement at failure and absorbed energy. Error went from 66.2%(force), 25.0%(displacement) and 77.3%(energy) to 7.70% & 7.74% (force), 12.5% & 31.28%(displacement) and 12.5% & 46.97% (energy). Additionally, time for meshing with hexahedral elements was reduced from one week to mere minutes, while maintaining accuracy comparable to tetra elements. This method shows promise for reducing shear locking. The study found no major differences between first and second order solid elements for both tetrahedral and hexahedral meshes other than computational cost. The micro-mapping methodology developed in this investigation using CT-scan data and casting process simulation data proved insufficient in accounting for the disparities between experiment and simulation results and needs further development. The thesis also explores the potential of Isogeometric Analysis (IGA) as a future method, noting its current limitations and immaturity for practical application. Furthermore, the study investigated alloys with increased iron and vanadium content, imitating the use of recycled material, to determine if there were any significant differences between them. However, no clear differences were observed, which may be attributed to larger sources of error overshadowing the potential variations between the alloys.