DESC0024819

Competitive pressures in many industries including automotive are driving the adoption of advanced manufacturing technologies that promise improved profitability. In turn, these manufacturing changes require new functionality from engineering simulation tools, which are used to enable efficient production processes. In the automotive industry, for example, major carmakers are increasingly adopting the practice of ôgiga-castingö: replacing complex sheet-metal assemblies with large, unified aluminum die-casts. This poses a challenge for crash simulation analysis: existing finite element analysis tools are not optimized to provide crash simulation for cars with these complex largecast part. This project addresses this and related engineering simulation challenges by using a powerful library funded by the Department of Energyĺs Advanced Scientific Computing Research program. This library can accelerate the computational speed and efficiency of a next-generation isogeometric analysis simulation code. Isogeometric analysis technologies are already being investigated intensely by automakers because they dramatically reduce the model preparation requirements that constitute the primary cost of crash and other simulations. While superior in accuracy, robustness, and overall time to solution, isogeometric analysis has remained computationally demanding, requiring relatively more solve time than legacy approaches. Fortunately, a library developed under the Advanced Scientific Computing Research program to support the machine learning community is also ideally suited to accelerate the computation of data produced by isogeometric analysisĺs modern analysis approach. During Phase I, this project will prepare the commercial isogeometric analysis code to use the Department of Energy-funded library, integrate the library with the companyĺs isogeometric analysis code, and demonstrate compatibility through basic benchmarks. Phase II will build on the proof of concept from Phase I and further optimize the isogeometric analysis code using the Department of Energy-funded library. The second phase will exploit the features of the library to accelerate and optimize the isogeometric analysis code to scale to meet industry demands in automotive and other industries. By the end of Phase II, the isogeometric analysis code will be scaled to meet industry demands for large-scale cast parts. Potential benefits of this project include improved automotive safety, accelerated time to market for new automobiles, and improved energy efficiency thought lighter-weight cars. The developed technology will also accelerate time to market of safer products across multiple industries, including aerospace, nuclear energy, and defense.