Geometrically Versatile Numerical Method For Robust Imaged-Based Simulations

Our proposed commercial software will enable the automation of 3D scan-to-simulation capabilities, addressing the escalating demand for improved efficiency and robustness in engineering simulations. The product will be designed to meet the needs of diverse industries, including manufacturing, mining, biomechanics, aerospace, medicine, and defense. Running simulation based on scanned ‘as-built’ parts often requires manual intervention to ensure high-quality discretizations with well-shaped elements. Our tool focuses on automating workflows through a geometrically versatile numerical method (GVNM) which automatically handles poor meshes. Phase I involves developing proof-of-concept software, demonstrating the capability to construct robust and accurate implicit dynamic simulations from three-dimensional images, such as CT scans. Building on the conforming reproducing kernel (CRK) method, our tool promises superior performance and expanded applicability. The subsequent Phase II will align with customer needs and include: porting to graphical processing units (GPUs) for improved throughput; uncertainty quantification (UQ) of scan-to-segment-to-mesh-to-analysis workflows; improved efficacy of contact, friction, and material response; and an optimization-based framework for material calibration and improving system performance. In summary, our software offers an innovative solution to make scan-based analysis accessible, providing engineers and scientists with efficient tools to make informed decisions faster in the rapidly evolving digital engineering landscape.