Solving the Heterogeneous VHTR Core with Efficient Grid Computing

Abstract

This paper uses the coarse mesh transport method COMET to solve the eigenvalue and pin fission density distribution of the Very High Temperature Reactor (VHTR). It does this using the Boltzmann transport equation without such low-order approximations as diffusion, and it does not simplify the reactor core problem through homogenization techniques. This method is chosen as it makes highly efficient use of grid computing resources: it conducts a series of calculations at the block level using Monte Carlo to model the explicit geometry within the core without approximation, and compiles a compendium of data with the solution set. From there, it is able to solve the desired core configuration on a single processor in a fraction of the time necessary for whole-core deterministic or stochastic transport calculations. Thus, the method supplies a solution which has the accuracy of a whole-core Monte Carlo solution via the computing power available to the user. The core solved herein, the VHTR, was chosen due to its complexity. With a high level of detailed heterogeneity present from the core level to the pin level, and with asymmetric blocks and control material present outside of the fueled region of the core, this reactor geometry creates problems for methods which rely on homogenization or diffusion methods. Even transport methods find it challenging to solve. As it is desirable to reduce the number of assumptions necessary for a whole core calculation, this choice of reactor and solution method combination is an appropriate choice for a demonstration on an efficient use of grid computing.