Simulation/Experimentation

Challenge: Accelerate utilization of simulation and experimentation.

How: Integrate experimentation and simulation to enable the development of first principles predictive simulation capabilities that are necessary to transition nuclear energy system design and licensing from reliance on experiments to reliance on modeling and simulation.

Background: In the past half century, the nuclear energy industry and regulatory agency approach to nuclear system design and licensing has relied significantly on experimental testing. This conventional paradigm embraces conservative design principles and has ensured nuclear safety, but at the cost of extensive experiments required by the current licensing process to validate modeling and simulation tools currently in use for core design. Additionally, the lengthy and complex software quality assurance process required by the licensing authority prevents many from using newly-available models or tools, thus further delaying the use of new simulation tools that are closer to a true predictive capability. These two issues combined deter licensing authorities from trusting the predictive capabilities of software and increases the reliance on new experiments.

The challenge thus becomes to develop and improve versatile predictive simulation capabilities that can easily integrate new models without a lengthy re-qualification process, while designing and developing a set of broad, challenging, and well-instrumented experiments that can clearly demonstrate the predictive capability of the new simulation tools and identify the areas in which the tools need improvement. Significant computational challenges exist in quantifying the impact of uncertainties on nuclear reactor performance in a multiphysics context.

Software development standards have increased significantly over the years, but the quality assurance process remains uneven. Legacy codes have been grandfathered into the licensing regime, while new codes require a significant quality assurance process, dissuading attempts to integrate advancements. At the same time, experiments have become so cost-prohibitive that laboratories and industry rely on old experiments that often lack the detail and precision needed to validate advanced first principles simulation capabilities.

Addressing this challenge requires a greater trust in first principles modeling and simulation capabilities and the definition of simpler guidelines in the development of quality-assured software. Additionally, high-fidelity software should be used to design a set of broad critical experiments in order to gain support in the construction of such facilities. A new paradigm that closely integrates these experiments and predictive simulations for the design and licensing process is needed.