Session: 08-01: New Facility Planning/ Environmental Management (EM)/ Health & Safety

Paper Number: 109865

109865 - Comparison Between Different Source Reconstruction Methods of Atmospheric Leakages 

When a severe nuclear accident occurs, the diffusion and transportation of airborne radioactive materials can cause harmful effects on the public. An accurate source release rate assessment is necessary to evaluate the consequence of the accident. Provided the leakage source location is known, the source inversion method can inversely estimate the release rate by coupling environmental measurements and the atmospheric dispersion model. Since the Fukushima nuclear accident, source inversion has significantly progressed over the past decade. However, the recent accidental radionuclide leakages are characterized by unknown source locations, causing difficulties in source inversion. Consequently, source reconstruction that simultaneously localizes the atmospheric leakage source and retrieves the release rate has become a highly prevailing and challenging topic.

In this study, we introduce and compare three source reconstruction schemes: the source-receptor sensitivity (SRS) statistical method, the cost function method, and the Bayes method. The benchmark SRS data are generated by the backward mode of the FLEXPART model. All methods are validated by the European Tracer Experiment-the first release (ETEX-I). The results indicate the respective advantages and disadvantages in source reconstruction, among which the Bayes method achieves the highest accuracy whereas the SRS statistical method achieves the lowest computational cost. In addition, we compare the source reconstruction performance between the two Bayes methods with the assumption of constant release and time-varied release, respectively. The source release rates in all schemes (if needed) are estimated using the traditional Tikhonov regularization method. The results demonstrate that the Bayes method with time-varied release may contribute to better source reconstruction performance. In summary, the above source reconstruction methods are suitable for different emergency response scenarios or stages, thus having the potential to be integrated into the nuclear emergency and response system.

Presenting Author: Yuhan Xu Tsinghua university

Presenting Author Biography: Yuhan Xu, Ph.D. student, Institute of Nuclear and New Energy Technology (INET) of Tsinghua University in Beijing, China. Supervised by associate professor Sheng, Fang. Recent work focuses on atmospheric dispersion modeling, and spatiotemporal decoupling source reconstruction.