No abstract

No abstract

performance metrics may be monitored by working within a recently-developed data management system called the ARMI. Through ARMI, the response of each design perturbation may be evaluated not only for the flux and reactivity, but also for reactivity coefficients, thermal hydraulics parameters, economics, and transient performance. Considering the parameters available, an automated optimization framework is designed and implemented. A non-parametric surrogate model using the Alternating Conditional Expectation (ACE) algorithm is trained with many design perturbations and then transformed through the Physical Programming (PP) paradigm to build an aggregate objective function without iteratively determining weights. Finally, the design is optimized with standard gradient-based methods. Through the power of ACE and the transparency of PP, the optimization system allows users to locate designs that best suit their multiobjective preferences with ease.

The Advanced Reactor Modeling Interface (ARMI) code system has been developed at TerraPower to enable rapid and robust core design. ARMI is a modular modeling framework that loosely couples nuclear reactor simulations to provide high-fidelity system analysis in a highly automated fashion. Using a unified description of the reactor as input, a wide variety of independent modules run sequentially within ARMI. Some directly calculate results, while others write inputs for external simulation tools, execute them, and then process the results and update the state of the ARMI model. By using a standardized framework, a single design change, such as the modification of the fuel pin diameter, is seamlessly translated to every module involved in the full analysis; bypassing error-prone multi-analyst, multi-code approaches. Incorporating global flux and depletion solvers, subchannel thermal-hydraulics codes, pin-level power and flux reconstruction methods, detailed fuel cycle and history tracking systems, finite element-based fuel performance coupling, reactivity coefficient generation, SASSYS-1/SAS4A transient modeling, control rod worth routines, and multi-objective optimization engines, ARMI allows “one click” steady-state and transient assessments throughout the reactor lifetime by a single user. This capability allows a user to work on the full-system design iterations required for reactor performance optimizations that has traditionally required the close attention of a multi-disciplinary team. Through the ARMI framework, a single user can quickly explore a design concept and then consult the multi-disciplinary team for model validation and design improvements. This system is in full production use for reactor design at TerraPower, and some of its capabilities are demonstrated in this paper by looking at how design perturbations in fast reactor core assemblies affect steady-state performance at equilibrium as well as transient performance. Additionally, the pin-power profile is examined in the high flux gradient portion of the core to show the impact of the perturbations on pin peaking factors.