Session: 05-07: System Performance and Safety Enhancements

Paper Number: 135061

135061 - Evaluation and Sensitivity Study of Uniform Kernel Method in Atmospheric 41ar Dispersion Against the Belgian Field Experiment 

Abstract: 

Abstract

The Lagrangian Particle Dispersion Model (LPDM) is an effective tool that comprehensively and accurately estimates the transport trajectory of radionuclides and determines their concentration to assess environmental contamination levels. However, the application of LPDM often presents a steep increase in computational costs, constrained by the mechanism of the traditional concentration calculation method (Box counting) that necessitates a substantial number of released particles and the need to track each of them. This shortcoming would pose challenges for nuclear emergency response, particularly in the early stages of the nuclear power plant (NPP) areas. To overcome this limitation, another concentration calculation method has been proposed recently, named kernel density estimator, which assumes that the mass of the particle is spread in space following density distribution. This method allows for a reduction in the number of particles and conserves computational resources compared to the traditional concentration calculation in LPDM.

The Uniform kernel is one of the effective kernel density estimators, serving as a valuable and precise tool to ensure effective nuclear emergency response in the early stages. However, the corresponding application of it in the NPP scenarios has not been adequately investigated. In this paper, based on the small-scale Belgian field dispersion experiment, the atmospheric dispersion modeling is developed by coupling the California Meteorological Model (CALMET) and the Lagrangian Particle Dispersion Model (LAPMOD) to simulate the ⁴¹Ar concentration field by utilizing the Uniform kernel. And the ⁴¹Ar gamma dose rate field is subsequently obtained using the dose conversion coefficient. The effectiveness of the Uniform kernel is systematically evaluated by comparing simulations with measurements from eight ⁴¹Ar air gamma dose rate monitoring sites deployed over two days (October 3 and October 4, 2011) near the BR1 research reactor during the Belgian field experiment. In addition, sensitivity studies are performed on the relevantly vital parameters, including the bandwidths multiplication factors, the horizontal grid size, and the released particle numbers. The results demonstrate satisfactory agreement between simulations and measurements, with acceptable statistical metrics on October 3 (Fraction of predictions within a factor of two of observations: 0.48, Fractional mean bias: 0.53, Normalized mean-square error: 1.15, Normalized absolute difference: 0.38) and on October 4 (Fraction of predictions within a factor of two of observations: 0.37, Fractional mean bias: 0.34, Normalized mean-square error: 1.09, Normalized absolute difference: 0.42). The sensitivity analysis highlights the significant impact of bandwidths multiplication factors and horizontal grid size on the Uniform kernel. Among them, an increase in the bandwidths multiplication factors is found to considerably expand the width of the plume and decrease the simulated values of the central plume. Additionally, the grid size exhibits a noteworthy change in the detailed characteristics of the plume. The appropriate parameters are summarized for the application of the Uniform kernel in such a small-scale dispersion of the NPP scenario.

Keywords: Atmospheric dispersion; Uniform kernel; CALMET; LAPMOD; Belgian field database; Sensitivity analysis

Presenting Author: Xinpeng Li North China Electric Power University

Presenting Author Biography: Xinpeng Li, lecturer at North China Electric Power University, whose main research area is the
evaluation of nuclear accident consequences

Authors:

Yeshuai Sun North China Electric Power University
Yujie Zhang North China Electric Power University
Zhaoyang Wang North China Electric Power University
Li Yang North China Electric Power University
Xinpeng Li North China Electric Power University
Yixue Chen North China Electric Power University
Xinwen Dong Tsinghua University
Sheng Fang Tsinghua University