Nuclear power plant atmospheric diffusion prediction correction method and system, medium and electronic equipment

Abstract

The invention discloses a nuclear power plant atmospheric diffusion prediction correction method and system, a medium and electronic equipment. The method and the system are suitable for nuclear powerplant accident leakage. The method comprises the following steps of: S1, predicting the current pollutant spatio-temporal data through a combined prediction model according to the current environmentinformation and leakage information; S2, acquiring current radiation monitoring spatio-temporal data actually monitored during accident leakage; S3, calculating to obtain a correction model accordingto the predicted current pollutant spatio-temporal data and the actually monitored current radiation monitoring spatio-temporal data; and S4, predicting future pollutant spatio-temporal data throughthe combined prediction model according to the current environment information and the leakage information, and inputting the future pollutant spatio-temporal data into the correction model for correction to obtain corrected future pollutant spatio-temporal data. Therefore, more accurate prediction is made, practical and feasible nuclear emergency decision-making technical support is provided fora decision maker, and the system and the method have important significance in taking emergency protection actions in a short time and reducing radioactive consequences of nuclear accidents.

Description

technical field [0001] The invention relates to the technical field of computer software, in particular to a nuclear power plant atmospheric diffusion prediction and correction method, system, medium and electronic equipment. Background technique [0002] Atmospheric diffusion needs to be predicted when a nuclear power plant accident leaks, but this prediction often has deviations, so the prediction results need to be corrected. The existing prediction and correction method is to combine model simulation data with observation data to improve the prediction results of the model, which originated in the 1950s. Its development has gone through several stages. Empirical algorithms were developed in the 1950s and 1960s, mathematical algorithms were developed in the 1970s and 1980s, simple intelligent algorithms were developed after the 1980s, and intelligent algorithms were developed after the 21st century. Data assimilation methods are of great significance in numerical predict...

AI Technical Summary

Problems solved by technology

[0004] However, due to the nonlinear change of the state quantity in practical applications, and the dimension is usually very high, the variational algorithm requires a lot of calculation iterations to solve

The four-dimensional variational method needs to solve the adjoint mode and tangent linear equation, but the adjoint mode is difficult to write, and the calculation cost is very high, especially considering the change of background field information over time, which further increases the complexity of the adjoint mode

[0006] However, in the practical application of this algorithm in the assimilation of atmospheric diffusion data, the atmospheric diffusion model needs to be called multiple times in each calibration step to generate the probability density function of the state quantity, which takes a lot of time and also has the problem of filter divergence

[0007] Therefore, the existing prediction models all require very strict input, and there are a lot of input data. In the early stage of the accident, it is impossible to obtain those data accurately.

Images