Research on the simulated dispersion of gaseous radionuclides:Validation and application for splitting puff dispersion model
[Objective]Local-scale atmospheric dispersion modeling of radionuclides is crucial for nuclear emergency response during the early phase.The Lagrangian puff dispersion model excels in accurately and rapidly reproducing radioactive fields at this scale by accounting for natural turbulence and integrating wind fields with spatial and temporal variations.Given that nuclear power plants(NPPs),especially Chinese NPPs,are often located in heterogeneous terrains,which lead to channeling and slope flows,puff splitting in the puff dispersion model is necessary to accurately represent the phenomenon of plume splitting and layer decoupling phenomena.Despite its importance,the threshold values for puff splitting have not been adequately studied.In addition,the complex terrain around NPP sites generates highly complicated flows,necessitating the use of a diagnostic wind field model coupled with the atmospheric dispersion model to improve the accuracy of dispersion simulations.[Methods]To further provide an effective atmospheric dispersion modeling and establish threshold values of puff splitting for the Lagrangian puff dispersion model,the local-scale Lagrangian splitting puff dispersion model(SPUFF)was developed and fully integrated with the California meteorological model(CALMET).Two local-scale dispersion simulations were conducted using the CALMET to drive the SPUFF:one against the Sanmen NPP wind tunnel experiments with east(E)and northeast(NE)wind directions and another to simulate the Fukushima Daiichi nuclear accident.These simulations aimed to validate SPUFF's performance and practicality.Furthermore,a comprehensive sensitivity analysis was performed to determine the credible range of horizontal threshold values for puff splitting.The dispersion results were evaluated using multiple statistical metrics:the fraction of simulations within a factor of two/five/ten of the observations(FAC2/5/10),fractional mean bias(FB),normalized mean-square error(NMSE),normalized absolute difference(NAD),and geometric mean bias(MG).[Results]Validation results indicated that plumes generated by SPUFF effectively covered the majority of measurement sites,with coverage rates reaching 99.60％and 97.54％in the E and NE directions,respectively.All four crucial statistical metrics for SPUFF met acceptable criteria(FAC2:0.52,FB:-0.17;NMSE:0.75,NAD:0.31 in the E direction;FAC2:0.48,FB:0.37;NMSE:1.28,NAD:0.39 in the NE direction),indicating remarkable performance.Practical evaluations demonstrated that SPUFF can reproduce more measurements in the Futaba station compared to the Lagrangian particle model(LAPMOD).SPUFF also successfully captured the concentration peak effects resulting from the reactor events during the Fukushima nuclear accident.Sensitivity analysis suggested that applying no puff splitting module might be sufficient for complex terrains with constant meteorological conditions(constant wind fields).However,puff splitting becomes crucial in complex terrains with variable meteorological conditions.For local-scale dispersion scenarios involving NPPs,the recommended threshold values for puff splitting range between 700 m and 1 100 m.[Conclusions]This paper provides a comprehensive evaluation of the Lagrangian splitting puff dispersion model(SPUFF)and demonstrates its practical application.The results strongly indicate that SPUFF is a valuable tool for future nuclear emergency responses.Additionally,this paper proposes a credible range of threshold values for puff splitting,offering guidelines for applying the puff model in local-scale dispersion scenarios at NPP sites.
万方数据
