A fast shielding calculation method based on real-time optimization of particle sampling positions

Abstract

The present application relates to the field of radiation shielding technology, specifically, to a fast shielding calculation method based on real-time optimization of particle sampling positions, including the following steps: Step 1: Establish a calculation model or import a CAD geometric model; Step 2: Sample particles in the model to determine the real-time information of the position; Step 3: Judge the model based on the real-time information of the particle sampling position; Step 4: Use the Monte Carlo method for transport calculation in areas with complex models, and use the deterministic method for transport calculation in areas with simple models but strong shielding effects; Step 5: Each time particle sampling is performed, the data of the Monte Carlo transport calculation and the data of the deterministic method are converted and interacted in real time; Step 6: Repeat steps 2‑5 until the calculation result of the radiation dose is obtained statistically. The present application realizes the two-way conversion between the Monte Carlo method and the deterministic method, effectively improving the calculation efficiency and accuracy.

Description

Technical Field

[0001] The present application relates to the field of radiation shielding technology, and in particular to a fast shielding calculation method based on real-time optimization of particle sampling positions. Background Art

[0002] Radiation shielding design is an important part of nuclear engineering design. Its purpose is to protect the health and safety of radioactive workers, the public, and the environment. Radiation shielding mainly uses the interaction between radiation and matter to reduce the radiation level in a certain area, thereby reducing the exposure of personnel and radiation damage to materials. In nuclear engineering design, the quality of radiation shielding calculation and design directly affects the life of the nuclear system and the radiation safety of workers and the surrounding environment. Therefore, selecting an appropriate shielding calculation method is one of the keys to ensuring the quality of shielding system design.

[0003] With the development of computer technology in recent years, there are many numerical calculation methods used to solve the problem of particle transport in shielding bodies. The first type of deterministic methods, such as the spherical harmonics method (PN), discrete ordinate method (SN), characteristic line method (MOC), penetration probability method (TPM) and collision probability method (CPM), finite element method, etc., all use numerical methods to solve the Boltzmann transport equation that describes the average behavior of particles to obtain the required physical quantities; the other type of Monte Carlo method is a non-deterministic method. It is a calculation method that randomly simulates the behavior of particles on a computer. Through the observation and analysis of a large number of neutron behaviors, the statistical average method is used to infer the solution of the estimated quantity. At present, these calculation methods have been maturely used in engineering design.

[0004] The advantages of the Monte Carlo method are that it can accurately simulate complex geometric models, use continuous energy cross sections, and have few approximations. However, for problems with thick shielding layers (i.e., deep penetration problems), it is difficult to give reliable calculation results within a reasonable time. Although the deterministic method is suitable for solving deep penetration problems, it has serious defects in dealing with complex geometries and accurate source terms, and cannot solve charged particle transport problems. Its engineering applications are mostly limited to one-dimensional and two-dimensional problems. Although particle transport theory continues to mature and computer hardware continues to develop, due to the above limitations, a single Monte Carlo method or determinism is difficult to meet the requirements of accurate shielding design for large and complex nuclear energy systems. Summary of the invention

[0005] The main purpose of this application is to provide a fast shielding calculation method based on real-time optimization of particle sampling positions, abandoning the disadvantages of deterministic methods and Monte Carlo methods, combining the advantages of both, improving the accuracy and speed of shielding calculations, and solving the shielding calculation problems of various large and complex nuclear facilities.

[0006] In order to achieve the above-mentioned purpose, the present application provides a fast shielding calculation method based on real-time optimization of particle sampling positions, comprising the following steps: Step 1: Establish a calculation model or import a CAD geometric model; Step 2: Perform particle sampling on the model to determine the real-time information of the position; Step 3: Judge the model based on the real-time information of the particle sampling position; Step 4: Use the Monte Carlo method for transport calculation in areas with complex models, and use the deterministic method for transport calculation in areas with simple models but strong shielding effects; Step 5: Each time particle sampling is performed, the data of the Monte Carlo transport calculation and the data of the deterministic method transport calculation are converted and interacted in real time; Step 6: Repeat steps 2-5 until the calculation results of the radiation dose are statistically obtained.

[0007] Furthermore, in step 2, after each sampling, the multidimensional arrays of space, energy group, and angle grids and the position, direction, energy, weight, and particle type information are saved.

[0008] Furthermore, in step 5, the data of the transport calculation by the Monte Carlo method and the data of the transport calculation by the deterministic method are intertwined, which is achieved by the mutual conversion of the conditional probability density function to the boundary angular flux density.

[0009] Furthermore, in step 4, the deterministic method transport calculation is called from the spherical harmonics method (PN), discrete ordinate method (SN), method of characteristic lines (MOC), penetration probability method (TPM) and collision probability method (CPM), and finite element method.

[0010] The present invention provides a fast shielding calculation method based on real-time optimization of particle sampling positions, which has the following beneficial effects:

[0011] The present application realizes the two-way conversion between the Monte Carlo method and the deterministic method. The Monte Carlo method is used for transport calculation in geometrically complex areas, and the deterministic method is used for transport calculation in geometrically simple areas but with strong shielding effects, which effectively improves the calculation efficiency and accuracy. At the same time, the parallel calculation of the Monte Carlo module is realized. The angular flux density file can be output at the interface from the Monte Carlo to the deterministic method, and the Monte Carlo source term sampling probability distribution file can be output at the interface from the deterministic method to the Monte Carlo. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The drawings constituting a part of this application are used to provide a further understanding of this application, so that other features, purposes and advantages of this application become more obvious. The schematic embodiment drawings and their descriptions of this application are used to explain this application and do not constitute an improper limitation on this application. In the drawings:

[0013] Figure 1It is a flow chart of a fast shielding calculation method based on real-time optimization of particle sampling positions provided in an embodiment of the present application; DETAILED DESCRIPTION

[0014] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0015] like Figure 1 As shown, the present application provides a fast shielding calculation method based on real-time optimization of particle sampling positions, comprising the following steps: Step 1: According to the problem of radiation shielding calculation, a calculation model is established or a CAD geometric model is imported; Step 2: Particle sampling is performed on the model to determine the real-time information of the position; Step 3: The model is judged according to the real-time information of the particle sampling position; Step 4: According to the types of Monte Carlo method and deterministic method, different user input files are created, different cross-section databases are called, sampling methods are selected according to the sampling position of the particles, and appropriate sampling programs are called. The Monte Carlo method is used for transport calculation in the area with complex model, and the deterministic method is used for transport calculation in the area with simple model but strong shielding effect; Step 5: Each time particle sampling is performed, the data of Monte Carlo transport calculation and the data of deterministic method transport calculation are converted and interacted in real time; Step 6: Steps 2-5 are repeated until the calculation result of radiation dose is obtained statistically.

[0016] Specifically, the fast shielding calculation method based on real-time optimization of particle sampling positions provided in the embodiments of the present application adopts an interface program and a sampling program to automatically generate a source description file, selects a calculation method in real time based on the particle sampling position, and realizes the storage and mutual conversion of data, information, and radiation source items (position, direction, energy, flux, etc.) between Monte Carlo self-loop, deterministic method self-loop, and alternating or doping cycles between Monte Carlo and deterministic methods. In addition, by controlling the interface error and establishing a global communication domain, the advantages of the deterministic method in solving deep penetration problems and the strengths of the Monte Carlo method in simulating complex geometry are maximized, thereby improving the accuracy and speed of shielding calculations, and being used to solve the shielding calculation problems of various large and complex nuclear facilities.

[0017] Furthermore, in step 2, after each sampling, the multidimensional array of space, energy group, angle grid and the position, direction, energy, weight, and particle type information are saved. The multidimensional array of space, energy group, angle grid and the position, direction, energy, weight, and particle type information can realize the bidirectional conversion between Monte Carlo method and deterministic method.

[0018] Furthermore, in step 5, the data of the transport calculation by the Monte Carlo method and the data of the transport calculation by the deterministic method are intertwined, which is achieved by the mutual conversion of the conditional probability density function to the boundary angular flux density.

[0019] Specifically, in the process of converting the Monte Carlo method to the deterministic method, when the computational domain of the deterministic method is the subsequent computational domain of the Monte Carlo computational domain, the Monte Carlo particle track is converted to the angular flux density in the deterministic method, and the position, direction, energy, weight, particle type, track length and other information in the Monte Carlo calculation are converted to the particle angular flux density in the spatial grid, energy group and direction interval in the deterministic method at the interface. The conversion of the Monte Carlo geometric description to the geometric description data in the deterministic method is the conversion of the computational domain under the CSG representation to the computational domain under the structural grid representation. The material information is extracted from the CSG geometry and set to the corresponding grid of the structural network calculation in the deterministic method. The materials of several CSG network centers are taken as the materials of the grid in the deterministic method. In order to improve the conversion accuracy, the weighted average of the materials of multiple points in the grid is taken as the material of the deterministic grid. The detector provided by the Monte Carlo program generally only gives the average result or the integral result, while the conversion calculation requires complete information on the position, direction, energy, weight and particle type of each Monte Carlo particle track. Use the Write Surface Source Card (SSW) function in Monte Carlo to record the complete information of each Monte Carlo particle track and obtain the complete Monte Carlo particle track information as the input of the Monte Carlo to deterministic data conversion program. According to the correspondence between the space, energy, and angle in the Monte Carlo method and the deterministic method, the contribution of the Monte Carlo particle to the grid flux density in the deterministic method is calculated in real time. The calculation method is as follows: Assume that the track of the Monte Carlo particle on the interface is (weight n ,r n ,E n ,Ω n ), where weight represents the particle weight, r represents the particle position, E represents the particle energy, Ω represents the particle direction, and n represents the track number. The boundary source term in the deterministic method requires the particle angular flux density ψ(r i ,E g ,Ω m ). particle (weight n ,r n ,E n ,Ω n ) contributes to the angular flux density in the grid i, energy group g and direction m:

[0020]

[0021] The particle angular flux density is obtained as follows:

[0022]

[0023] where w m To find the product weight coefficient, N is the number of particle tracks, △S is the area of ​​the surface element, λ n is the cosine of the angle between the particle track and the normal direction of the interface.

[0024] More specifically, in the process of converting the deterministic methodology to the Monte Carlo method, when the Monte Carlo calculation domain is the subsequent calculation domain of the deterministic method calculation domain, the angular flux density in the deterministic method is converted to the Monte Carlo particle source term, that is, the particle angular flux density in the grid, energy group, and direction interval calculated by the deterministic method is converted into the source variable information such as the source term position, direction, energy, weight, and particle type in the Monte Carlo calculation. The solution to the correspondence between the deterministic method and the Monte Carlo space, energy, and angle is the same as the conversion from Monte Carlo to the deterministic method. Secondly, the probability distribution of the Monte Carlo particle source in each grid is calculated based on the angular flux density in the grid in the deterministic method. The calculation method is as follows: Assume that the particle angular flux density in the grid i, energy group g, and direction m interval is ψ(r i ,E g ,Ω m );Directional quadrature weight coefficient is w m The cosine value of the angle between the particle flight direction and the normal direction of the interface surface is λ m ; Grid interval r i The corresponding interface area is S i ; The number of interface grids is I; The number of energy groups is G; The number of discrete directions is M, the probability that the particle position r is within the grid interval i, the particle energy is within the energy group g, and the particle direction is within the direction interval m:

[0025]

[0026] After obtaining the sampling probability of the source variable, first sample the grid where the particle is located, then sample the grid i, energy group g, and direction interval m in turn, and then evenly sample r, E, and Ω in the grid.

[0027] Furthermore, for the conversion of parallel Monte Carlo particle transport to serial deterministic method, a global communication domain is first established, and the Monte Carlo method transport is calculated in parallel in all processes of the global communication domain, while the deterministic method transport is calculated serially on process No. 0 of the global communication domain. Before the calculation starts, the cross-section and other data required for the Monte Carlo method transport are broadcast in the global communication domain. After the calculation of each Monte Carlo transport area is completed, the flux and other data are reduced in the global communication domain and the error is calculated. After the parallel conversion of Monte Carlo transport to deterministic method transport, communication operations such as the reduction of boundary angle flux density are performed. After the serial conversion of deterministic method transport to Monte Carlo transport, communication operations such as the broadcast of the particle source probability density function of Monte Carlo transport are performed.

[0028] Furthermore, at each interface where the Monte Carlo method is converted to the deterministic method, the error of the surface flux of the Monte Carlo method is controlled to ensure that the surface flux error of each grid of the corresponding deterministic method does not exceed 5%; at each interface where the deterministic method is converted to the Monte Carlo method, the accuracy of the angular flux density in the deterministic method is controlled to ensure that the error of the angular flux density in each grid of the deterministic method does not exceed 0.5%.

[0029] Furthermore, in step 4, the deterministic transport calculation is called from the spherical harmonics method (PN), discrete ordinate method (SN), characteristic line method (MOC), penetration probability method (TPM), collision probability method (CPM), and finite element method. According to the actual situation, the deterministic transport calculation can be directly called in the finite element method.

[0030] Specifically, in the embodiments of the present application, by comparing the simulation calculation results of the shielding fast calculation method based on real-time optimization of particle sampling positions and the calculation results of directly using Monte Carlo simulation, the real-time optimization method of particle sampling positions greatly accelerates the calculation convergence, and the error is controlled within 5%; by comparing the simulation calculation results of the shielding fast calculation method based on real-time optimization of particle sampling positions and the calculation results of directly using the deterministic method, the real-time optimization method of particle sampling positions greatly improves the error of the calculation results, and the time cost is relatively low.

[0031] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A fast shielding calculation method based on real-time optimization of particle sampling positions, characterized in that: The steps include: Step 1: Create a computational model or import a CAD geometry model; Step 2: Sample the model particles to determine the real-time information of the position; Step 3: Judge the model based on the real-time information of the particle sampling position; Step 4: For regions with complex models, the Monte Carlo method is used for transport calculation, and for regions with simple models but strong shielding effects, the deterministic method is used for transport calculation; Step 5: Each time a particle sampling is performed, the data of the Monte Carlo transport calculation and the data of the deterministic method transport calculation are converted and interacted in real time; The data of transport calculation by Monte Carlo method and that by deterministic method are intertwined, which is realized by mutual conversion of conditional probability density function to boundary angular flux density. According to the corresponding relationship between space, energy and angle in the Monte Carlo method and the deterministic method, the contribution of the Monte Carlo particle to the grid flux density in the deterministic method is calculated in real time. The calculation method is as follows: Assume that the track of the Monte Carlo particle on the interface is (weight n ,r n ,E n ,Ω n ), where weight represents the particle weight, r represents the particle position, E represents the particle energy, Ω represents the particle direction, and n represents the track number. The boundary source term in the deterministic method requires the particle angular flux density ψ(r i ,E g ,Ω m ); particle (weight n ,r n ,E n ,Ω n ) contributes to the angular flux density in the grid i, energy group g and direction m: The particle angular flux density is obtained as follows: where w m To find the product weight coefficient, N is the number of particle tracks, △S is the area of ​​the surface element, λ n is the cosine of the angle between the particle track and the normal direction of the interface; The probability distribution of the Monte Carlo particle source in each grid is calculated based on the angular flux density in the grid in the deterministic method. The calculation method is as follows: Assume that the particle angular flux density in the grid i, energy group g and direction m interval is ψ(r i ,E g ,Ω m );Directional quadrature weight coefficient is w m The cosine value of the angle between the particle flight direction and the normal direction of the interface surface is λ m ; Grid interval r i The corresponding interface area is S i ; The number of interface grids is I; The number of energy groups is G; The number of discrete directions is M, the particle position r is within the grid interval i, the particle energy is within the energy group g, and the probability that the particle direction is within the direction interval m is: After obtaining the sampling probability of the source variable, first sample the grid where the particle is located, then sample the grid i, energy group g, and direction interval m in turn, and then evenly sample r, E, and Ω in the grid; Step 6: Repeat steps 2-5 until the calculated radiation dose is statistically obtained.

2. The shielding fast calculation method based on real-time optimization of particle sampling position according to claim 1 is characterized in that: In step 2, after each sampling, the multidimensional arrays of space, energy group, and angle grids and the position, direction, energy, weight, and particle type information are saved.

3. The shielding fast calculation method based on real-time optimization of particle sampling position according to claim 1 is characterized in that: In step 4, the deterministic method transport calculation is called from the spherical harmonics method, discrete ordinate method, characteristic line method, penetration probability method and collision probability method, and finite element method.

Images