Interpolation Influence on the Fast Fourier Transform Based Calculation of Three-Dimensional Dose Rate Field

Abstract

The calculation of three-dimensional dose rate fields plays a key role in radiation dose rate estimation and the service for the nuclear emergency. The recent fast calculation method based on the Fast Fourier Transform (FFT) method can greatly speed up the calculation without losing accuracy, which is promising for operational usage in nuclear emergency response systems. But it can only be used for a uniform grid. Unfortunately, most atmospheric dispersion models use a non-uniform grid, which prevents the direct application of FFT-based calculation. Therefore, interpolation is required beforehand to use the Fourier transform, which may introduce errors and affect computing efficiency. In this paper, an atmospheric dispersion modeling case of a typical nuclear power plant (NPP) is used to investigate the efficiency of different interpolation methods, which are based on a non-uniform grid. These methods are linear interpolation and nearest-neighbor interpolation. The sensitive analysis of grid resolution is investigated in the slices of x, y, and z at typical positions, which confirms the smooth-out and speed-up effects in rough grids. A grid size over 10 m at any slice commonly causes losses of change details of dose rate fields. Given the same resolution of 50 m × 50 m × 50 m, the nearest neighbor performs a 717 times calculation faster than the linear method, which preserves more change details of dose rate fields as well. For complex calculation tasks, e.,g., non-uniform NPP buildings, the nearest neighbor interpolation method is recommended with a resolution of 10 m × 10 m × 10 m to make a good balance between accuracy and speed.