CN-122019966-A - Method for estimating medium-small scale nuclear accident source intensity of coupling numerical forecasting product

CN122019966ACN 122019966 ACN122019966 ACN 122019966ACN-122019966-A

Abstract

The invention discloses a medium-small scale nuclear accident source intensity estimation method of a coupling numerical forecasting product, which comprises the steps of obtaining radiation monitoring data and monitoring point position information in a simulation area, identifying available numerical weather forecasting product types, decoding weather data corresponding to the simulation area to obtain decoding files, obtaining space-time resolution encrypted wind field data files, generating wind field files consistent with radiation monitoring frequency through diagnosis wind field modes, calculating concentration fields of each nuclide according to the wind field files by calling particle diffusion modes, outputting concentration field files consistent with the radiation monitoring frequency, extracting diffusion factor time sequence data corresponding to monitoring points according to the monitoring point position information, constructing inversion matrixes corresponding to radiation monitoring data and diffusion factors one by one, inputting the inversion matrixes into a data assimilation algorithm, and carrying out iterative inversion by adopting unscented Kalman filtering to realize source intensity estimation.

Inventors

ZHAO DAN
LI YU
ZHANG JUNFANG
LI MINGYE
ZHANG FANG
ZHAO DUOXIN
LV MINGHUA
HUANG SHA
HE XIN
QIU ZHIXIN

Assignees

中国辐射防护研究院

Dates

Publication Date: 20260512
Application Date: 20251229

Claims (10)

1. The medium-small scale nuclear accident source intensity estimation method of the coupling numerical forecast product is characterized by comprising the following steps of: Acquiring radiation monitoring data and monitoring point position information in an analog region, performing validity check on the radiation monitoring data, removing invalid data and completing unit conversion; identifying available numerical weather forecast product types, and decoding weather data corresponding to the simulation area to obtain a decoding file; Acquiring a wind field data file with encrypted space-time resolution through a numerical weather forecast mode based on the decoding file, and generating a wind field file consistent with the radiation monitoring frequency through a diagnosis wind field mode by combining a simulated region terrain elevation file; According to the wind field file, a particle diffusion mode or commercial atmospheric diffusion numerical simulation software is called to calculate the concentration field of each nuclide and output a concentration field file consistent with the radiation monitoring frequency; Extracting diffusion factor time sequence data corresponding to the monitoring points from the concentration field file according to the position information of the monitoring points, and constructing an inversion matrix corresponding to the radiation monitoring data and the diffusion factors one by one; And inputting the inversion matrix into a data assimilation algorithm, performing iterative inversion by adopting unscented Kalman filtering, outputting an accident release point release rate time sequence, and entering the calculation at the next moment based on the condition that adjacent step errors meet preset convergence conditions at the evaluation moment, so as to realize the continuous source intensity estimation.
2. The method of claim 1, wherein the validity check of the radiation monitoring data includes checking a data file to identify source item information and radiation monitoring data information, marking corresponding monitoring point data as invalid data when no data is returned by the monitoring instrument or the monitoring value is background data, checking monitoring point position data and eliminating monitoring point data beyond a calculation area, and outputting a reminding file.
3. The method of claim 1, wherein the unit conversion comprises converting gamma-ray air absorption dose rate data to nuclide activity concentration data satisfying the relationship: Wherein, the The activity concentration of a certain nuclide in the air; the air absorption dose rate is gamma rays; Is the conversion coefficient of dose rate to activity concentration.
4. The method according to claim 1, wherein after identifying the type of the numerical weather forecast product, the preprocessing decoding module PreData is invoked to decode the range of the simulation area, so as to obtain a decoded file for processing the subsequent numerical weather forecast mode and the diagnostic wind field mode.
5. The method of claim 1, wherein the numerical weather forecast mode is a WRF mode that utilizes the decoded file to obtain a spatio-temporal resolution encrypted wind field data file.
6. The method of claim 1, wherein the diagnostic wind pattern is a CALMET pattern that calculates a simulated regional wind pattern file consistent with the radiation monitoring data frequency based on the simulated regional terrain elevation file and the radiation monitoring frequency.
7. The method of claim 1, wherein the particle diffusion mode calculates the diffusion process of each nuclide in the simulation area, outputs a concentration field file consistent with the frequency of the radiation monitoring data, and extracts diffusion factor time sequence data of the corresponding position of the monitoring point according to the grid point position of the radiation monitoring instrument in the concentration field data.
8. The method of claim 1, wherein the unscented kalman filtered data assimilation inversion comprises establishing a state equation and an observation equation, and performing inversion calculations based on the array of monitored data to estimate the accident release source intensity, the state equation and the observation equation satisfying: Wherein, the The time is the moment; is the state variable at time k +1, A state variable at time k; observing the quantity at the moment k; is a function of a state equation; Is an observation equation function; Is process noise; To observe noise.
9. The method of claim 1, wherein the convergence criterion comprises calculating adjacent step errors during the evaluation time And (3) entering the next moment calculation when the following conditions are met, otherwise, continuing iteration until the convergence conditions are met: wherein the threshold satisfies: Wherein, the To evaluate adjacent step errors in time; is the convergence threshold.
10. A medium and small scale nuclear accident source intensity estimation system of a coupled numerical forecast product, comprising: The monitoring data preprocessing module is used for acquiring radiation monitoring data and monitoring point position information in the simulation area, and carrying out validity check on the radiation monitoring data, removing invalid data and unit conversion; The weather product identification and decoding module is used for identifying the type of the numerical weather forecast product and calling PreData the corresponding weather data of the simulation area to decode so as to obtain a decoding file; the wind field generation module is used for calling the WRF mode to generate a space-time resolution encrypted wind field data file, and calling the CALMET mode to generate a wind field file consistent with the radiation monitoring frequency by combining the simulated region terrain elevation file; the diffusion calculation module is used for calling a particle diffusion mode or commercial atmospheric diffusion numerical simulation software to generate a concentration field file consistent with the radiation monitoring frequency; the factor extraction and matrix construction module is used for extracting a diffusion factor time sequence from the concentration field file according to the position information of the monitoring points and constructing an inversion matrix; And the assimilation inversion and convergence judging module is used for carrying out iterative inversion on the inversion matrix by adopting unscented Kalman filtering, outputting an accident release point release rate time sequence, and entering the next time calculation when the adjacent step error meets a preset threshold value.

Description

Method for estimating medium-small scale nuclear accident source intensity of coupling numerical forecasting product Technical Field The invention relates to the field of nuclear safety, in particular to a medium and small scale nuclear accident source intensity estimation method of a coupling numerical forecasting product. Background In the existing nuclear accident real-time on-line source item inversion method, the model has higher dependence on a weather observation station of a calculation area due to the fact that a single weather data source is relied on. The dependence is reflected in two aspects, firstly, a model starting condition needs to be provided with a tower layer weather observation station and a ground station weather observation station in a simulation area, the representativeness of weather data to real-time weather conditions in the simulation area is ensured, and the accuracy of inversion of a source item can be directly affected by inaccurate data. Second, due to the lack of coupling capability with a variety of numerical weather forecast products, existing methods have difficulty in efficiently integrating weather data from different sources, and when a data transmission channel fails, the mode will lose inversion computing capability due to the lack of startup data. This limitation makes it difficult for the model to provide reliable source term estimates in the face of sudden nuclear incidents, thereby affecting the effectiveness of emergency response and decision support. Disclosure of Invention To achieve the above object and other related objects, the present invention discloses a method for estimating medium-small scale nuclear accident source intensity of a coupled numerical forecast product, comprising: Acquiring radiation monitoring data and monitoring point position information in an analog region, performing validity check on the radiation monitoring data, removing invalid data and completing unit conversion; identifying available numerical weather forecast product types, and decoding weather data corresponding to the simulation area to obtain a decoding file; Acquiring a wind field data file with encrypted space-time resolution through a numerical weather forecast mode based on the decoding file, and generating a wind field file consistent with the radiation monitoring frequency through a diagnosis wind field mode by combining a simulated region terrain elevation file; According to the wind field file, a particle diffusion mode or commercial atmospheric diffusion numerical simulation software is called to calculate the concentration field of each nuclide and output a concentration field file consistent with the radiation monitoring frequency; Extracting diffusion factor time sequence data corresponding to the monitoring points from the concentration field file according to the position information of the monitoring points, and constructing an inversion matrix corresponding to the radiation monitoring data and the diffusion factors one by one; And inputting the inversion matrix into a data assimilation algorithm, performing iterative inversion by adopting unscented Kalman filtering, outputting an accident release point release rate time sequence, and entering the calculation at the next moment based on the condition that adjacent step errors meet preset convergence conditions at the evaluation moment, so as to realize the continuous source intensity estimation. Preferably, the validity check of the radiation monitoring data comprises the steps of checking a data file to identify source item information and radiation monitoring data information, marking corresponding monitoring point data as invalid data when monitoring instruments do not return data or monitoring values are background data, checking monitoring point position data, removing monitoring point data exceeding a calculation area, and outputting a reminding file. Preferably, the unit conversion includes converting gamma-ray air absorption dose rate data into nuclide activity concentration data satisfying the following relation: Wherein, the The activity concentration of a certain nuclide in the air; the air absorption dose rate is gamma rays; Is the conversion coefficient of dose rate to activity concentration. Preferably, after the type of the numerical weather forecast product is identified, the preprocessing decoding module PreData is called to decode the range of the simulation area, so as to obtain decoding files for processing the subsequent numerical weather forecast mode and the diagnosis wind field mode. Preferably, the numerical weather forecast mode is a WRF mode, and the WRF mode obtains the space-time resolution encrypted wind field data file by using the decoding file. Preferably, the diagnosis wind field mode is a CALMET mode, and the CALMET mode calculates a simulation area wind field file consistent with the radiation monitoring data frequency according to the simulation area topography elevati