Fidelity Enhancement of Nuclear Power Plant Simulators Utilizing High Fidelity Simulation Predictions

Phase-IIA will focus on adding features that will improve the usability of the engineering grade simulator, which is being developed by integrating RELAP5-3D, NESTLE and CTF, partially derived and verified via the VERA core simulator, during Phase-II. Motivated by Phase-II research results, several features will be added during Phase-IIA that improves the usability and accuracy of simulations. The usability of the engineering grade simulator will be enhanced through the following improvements. To improve core model generation, a process for automatic generation of the coarse mesh CTF model from the VERA model will be established, similar to what is already done for the NESTLE core neutronic model. The process for performing sequences of executions, such as VERA branch cases needed to generate data required to determine the nodal neutronic parameters, will be automated. Interpretation of results, required to assure simulations predict physically expected behavior, better understand transient progression, and verify safety margins, will benefit from work planned to enhance visualization. Calling upon expertise in designing human-machine interfaces and working in partnership with a reactor designer, a multi-monitors, touch screen soft controls immersion workstation will be developed. With coarse mesh CTF taking much longer to execute on a moderate number of cores (e.g. dozen cores) than RELAP5-3D or NESTLE executing each on one core, action will be taken to enhance parallelization effectiveness, improve the solution method, and reduce the geometric detail. Speedup of RELAP5-3D to address the increased computational burden of more detailed reactor vessel models will be pursued by implementation of an improved matrix solver that may employ parallelization. Controlling all three codes making up the engineering simulator, 3KEYMASTERTM will be modified to better manage parallelization in an environment that includes codes using MPI.Accuracy improvements to be implemented include capability to complete thermal-hydraulic analysis using a pin-wise CTF model to evaluate such attributes as fuel temperature and critical heat flux ratio over a core subregion. Required data, such as cross-flows on the subregion’s radial exterior surface, will be generated by the coarse mesh engineering simulator. It has been determined that during transients, pin-power reconstruction may not be sufficiently accurate, so cause of difficulty will be identified and addressed. Additional accuracy improvements to be pursued include improved treatment of partially rodded nodes with regard to the nodal and pin-reconstruction solutions, and treatment of CRUD to determine impact on transient behavior. Finally, accuracy improvements that could not have been anticipated but encountered during Phase-II development include the need for a detailed RELAP5 vessel model as noted above, and impact on nodal neutronic parameters of unrodded nodes from presence of rodded nodes. The resulting product from Phase-IIA will be a more usable and accurate engineering simulator of the nuclear steam supply system, the reactor vessel and core in particular, applicable to engineering design, plant follow and potentially digital twin usages.