Nuclear reactor physical simulation method and device based on singular value decomposition transformation, computer equipment and storage medium

Abstract

The invention provides a nuclear reactor physical simulation method and device based on singular value decomposition transformation, computer equipment and a storage medium, which are used for realizing correlation control on a random sample space and reducing the calculation amount of nuclear data sampling, so that an obtained simulation calculation result reaches expectation based on an obtained final sample. The method mainly comprises the following steps: determining the number of physical corresponding parameters of a nuclear reactor; the number of samples needing to be extracted; obtaining a random sample space corresponding to the number of the parameters; performing singular value decomposition on the covariance matrix of the random sample space to obtain a singular value matrix and a left singular matrix, and establishing a corresponding diagonal matrix; constructing a new random sample space according to the relational expression; checking whether the number of parameters in the random sample space reaches the number of parameters; if yes, calculating a final sample through the relational expression, and outputting the final sample; and inputting the final sample into a simulation system to obtain a nuclear reactor physical simulation result.

Description

technical field [0001] The present application relates to the technical field of data processing, and in particular, to a method, device, computer equipment and storage medium for nuclear reactor physics simulation based on singular value decomposition transformation. Background technique [0002] During sampling in nuclear reactor physics calculations, not only the joint probability density distribution of the input parameters of the analysis program, but also the correlation between the input parameters must be considered, so it is necessary to carry out correlation control in the sampling process. Correlation control technologies on the market mainly include: cyclic Latin hypercube sampling, Rosenblatt transform, Nataf transform and other methods. All of the above methods can achieve the independence of random sample spaces in the sampling process to a certain extent. [0003] The method of cyclic Latin hypercube sampling is cumbersome and the process of cyclic sampling i...

AI Technical Summary

Problems solved by technology

[0003] The cyclical Latin hypercube sampling method is cumbersome and complicated, especially when the sample size is too large, the task load will be very large. It is necessary to consider the rationality of the correlation control standard. The results obtained by this method still have a certain degree of correlation. It is difficult to obtain the joint probability density and cumulative joint probability density of the random vector X under the Rosenblatt transformation, and the Rosenblatt transformation is not unique, so the practicability is poor; the Nataf transformation is often used in practical applications, but it is computationally complex. Still unable to deal with the problem of correlated variables and independent variables

[0004] However, the correlation control of the random sample space cannot be realized through the above method, which will make the distribution type of the simulation calculation results fail to meet expectations, and cause serious deviations in the uncertainty analysis results

Images