Parametric spline “mesh” generation for nuclear engineering

The global warming crisis requires non-carbon-burning power generation to play a more significant role in the energy landscape. Nuclear energy efficiently generates electricity without releasing green- house gasses; however, nuclear systems take years and significant expense to design and construct. The work proposed in this SBIR Phase I will improve the quality and speed of nuclear design and simulation, reducing time to market and product safety risk. The highly qualified management has previously built and sold a company in a similar space and is capable of commercial success. Idaho National Laboratory’s open-source Multiphysics Object Oriented Simulation Environment (MOOSE) was written to allow modeling and finite element analysis (FEA) to be more accessible to a broad array of scientists. However, like all legacy codes, MOOSE is limited to running FEA on faceted approximations of computer-aided design (CAD) data, which inefficiently requires a significant number of superfluous elements to accurately represent shape and is disconnected from the CAD workflow. The three objectives proposed in this project will return more accurate nuclear simulation results, streamline the CAD design/FEA simulation workflow, and accelerate simulations by leveraging iso-geometric analysis (IGA). In previous work, IGA paradigms successfully have produced simulations orders of magnitude faster than FEA. The proposed work will enhance MOOSE with smooth spline IGA elements to bring this new approach to simulation to the nuclear sector. IGA enhances FEA by using a smooth CAD description directly in the analysis. This eliminates data translation and improves accuracy. The key to successfully using IGA in this manner is the company’s patent-pending breakthrough, U-splines. U-splines are a geometric representation suitable for both CAD and FEA. During this SBIR Phase I project, U-splines will be leveraged to build a parametric analysis-suitable CAD model of a light water reactor fuel cell, with an order of magnitude fewer elements than FEA. The resulting simulation improvements will be studied. The developed toolset will shorten the time required to complete nuclear simulations, and increase confidence in their results, because discretization error will be removed from the simulations. It will also make the MOOSE code, and especially mesh refinement studies, more accessible to less experienced users. This will lead directly to the potential for greater nuclear safety and innovation. This proposal builds on continued work and other SBIR projects previously awarded to bring the power of IGA to the automotive, defense, and mining industries. These industries share requirements for shape representation and contact algorithms with nuclear, and differences will be addressed in this proposal.