Gradient-Informed Design Optimization of Select Nuclear Systems

Stochastic Gradient Descent for Optimization of Nuclear Systems

“…ADAM uses the derivate of the objective function presented in the Methodology section to optimize the material density (effectively managing the location) within the system. An innovative aspect of this approach already discussed in previous work [2] is that setting the density to be a continuous variable is one way to approach the design optimization of arbitrary geometry. A penalty term can be added to force the density to converge to either 0 or unity at the end of the optimization creating a solution with discrete density.…”

“…Four challenge problems are proposed to test the use of TSUNAMI to build a gradient for the ADAM method. Challenge problems one, two and three mirror the challenge problems developed in the rst publication [2]. The set-up is a 55 cm x 55 cm 2-dimensional, un-re ected system.…”

“…This algorithm utilizes the derivatives of an objective function to iteratively take steps toward the minimum of the objective function and is designed to work well with a stochastic gradient, such as those estimated by Monte Carlo neutron transport simulations. Previously, this group published the demonstration of gradient descent for the design optimization of neutronics systems with quantitative constraints using the interior point method to optimize the k-eigenvalue, , and reaction rates [2]. The major issue brought up by the previous work was the stochastic nature of the gradients built from Monte Carlo neutron transport simulations.…”

Artificial Intelligence for Multiphysics Nuclear Design Optimization with Additive Manufacturing

“…In traditional light water reactor designs, the tendency for the cosine power distribution is counteracted by a combination of variable fuel and fixed-burnable-poison loading in the axial direction, control rod position, and axial coolant density variation. In an example optimization exercise, the cosine power distribution was counteracted by a variable fuel density in the axial direction [42]. The results show that a significant reduction in the axial power peaking is achievable by varying the density of the fuel and moderator in the axial direction during a manufacturing process, as shown in Figure 4.…”