Neutron transport parallel computing method and computing device for optimizing transverse leakage item processing

Abstract

The invention discloses a neutron transport parallel computing method and computing device for optimizing transverse leakage item processing, and belongs to the field of nuclear power. According to the neutron transport parallel computing method for optimizing transverse leakage item processing, a boundary flow weighted by a second-order Legendre polynomial is introduced in the computing process of a single segment, adjacent segments are updated in combination with the boundary neutron flow, all the segments are traversed through parallel computing, and updated segment neutron flux, eigenvalue keff and neutron source items are obtained. And substituting the updated data back to the calculation model for iteration to obtain a final calculation result. According to the method, the high-order expansion coefficient of the transverse leakage item of the node is obtained based on the single node model, and the calculation stability is improved while the calculation precision is ensured.

Description

Technical Field

[0001] The invention belongs to the field of nuclear power, and in particular relates to a neutron transport parallel calculation method and a calculation device for optimizing lateral leakage term processing. Background Art

[0002] In the calculation of reactor reactivity and whole-reactor fine power distribution, accurate calculation of neutron transport behavior is the basis for obtaining accurate calculation results. The transverse integration nodal method is widely used in neutron transport calculations. This method requires accurate consideration of the transverse leakage term of the nodal block to improve the calculation accuracy. In the high-resolution neutron transport calculation of nuclear reactors based on grid element homogenization, the radial grid size is about 1 cm, with a high resolution, and the calculation accuracy can be further improved by discontinuity factors or super equivalent homogenization factors; the axial grid is about 10 cm, with a large grid size. In the neutron transport calculation program system based on the transverse integration nodal method, it is necessary to accurately consider the transverse leakage term to improve the calculation accuracy of the axial power distribution, and then obtain accurate axial power offset (AO), hot spot factor (F Q ) and other key physical parameters.

[0003] Regarding the lateral leakage term in the neutron transport calculation method based on the transverse integral block method, the existing technical solutions adopt the flat leakage approximation or the high-order leakage term approximation based on the three adjacent block models. The flat leakage approximation method only considers the average value of the leakage term, and the calculation accuracy is low. The high-order leakage term processing method based on the three adjacent block models can obtain higher calculation accuracy, but when performing parallel calculations based on the regional decomposition method, the block boundary flow information at the interface of each parallel region is updated lagging, which will reduce the stability of the parallel calculation and even cause calculation divergence. In order to improve the stability of the neutron transport parallel calculation based on the transverse integral block method, a high-order leakage term processing method based on a single block model is needed. Summary of the invention

[0004] The purpose of the present invention is to provide a neutron transport parallel calculation method for optimizing the lateral leakage term processing, simplify the calculation complexity of neutron transport behavior, and improve the calculation efficiency and accuracy. The present invention also provides a calculation device.

[0005] According to an embodiment of one aspect of the present invention, a neutron transport parallel computing method for optimizing lateral leakage term processing is provided, the method comprising the following steps:

[0006] Step a): providing a simulation model of a reactor neutron transport calculation simulation object, and dividing the simulation model into blocks;

[0007] Step b): color-marking the nodes and making adjacent nodes have different colors, and treating the nodes with the same color marking as a same-color node group to carry out parallel regional decomposition;

[0008] Step c): Allocate the nodes to multiple computing cores according to the parallel domain decomposition result in step b), perform node response matrix calculation, and provide neutron flux, neutron flow, neutron source term and eigenvalue k eff The initial value of

[0009] Step d): calculating the boundary neutron flow, the node average leakage term and the boundary flow weighted by the second-order Legendre polynomial of the node in the node group respectively through each of the computing cores;

[0010] Step e): updating the boundary flow calculated in step d) to the adjacent nodes;

[0011] Step f): the node whose boundary neutron flow, the node average leakage term and the boundary flow weighted by the second-order Legendre polynomial have been calculated is taken as the calculated node, and the remaining nodes are taken as the uncalculated nodes;

[0012] In the uncalculated nodes where the boundary flow is updated, the boundary neutron flow, the node average leakage term and the boundary flow weighted by the second-order Legendre polynomial of each node group of the same color are calculated one by one, and the calculated boundary neutron flow and the boundary flow weighted by the second-order Legendre polynomial are updated to the adjacent nodes;

[0013] Step g): using a second-order polynomial approximation of the lateral leakage term, and taking the average leakage term of each node and the second-order Legendre polynomial weighted boundary flow at the node boundary as constraints, calculating the lateral leakage term expansion coefficient;

[0014] Step h): according to the neutron balance equation, taking the neutron source term and the nodal boundary neutron flux as input, calculating and updating the nodal neutron flux;

[0015] Step i): Calculate the characteristic value k according to the neutron flux of the node calculated in step h). eff , and update to all computing cores;

[0016] Step j): updating the neutron source term according to the boundary neutron flow term of the calculated node, the boundary flow weighted by the second-order Legendre polynomial, the expansion coefficient of the lateral leakage term calculated in step g), and the node neutron flux calculated in step h);

[0017] Step k): according to the updated boundary neutron flow and the boundary flow weighted by the second-order Legendre polynomial, the nodal neutron flux and the eigenvalue k eff, repeat step d) to step j) until the calculation result converges or reaches a given number of iterations, and output the final calculation result.

[0018] When the method performs parallel calculation of neutron transport based on the transverse integral nodal method, the high-order expansion coefficients of the nodal transverse leakage term can be obtained based on a single nodal model, thereby improving the calculation stability while ensuring the calculation accuracy.

[0019] Furthermore, in some embodiments, in the step a), the simulation model includes geometric information, material information, and nuclear reaction cross-section information.

[0020] Furthermore, in some embodiments, in the step b), during the region decomposition process, the load variance of each parallel computing core is minimized, and the surface area to volume ratio of the parallel computing region is minimized.

[0021] When the load variance of each parallel computing core is minimized, the load distribution of each parallel computing core can be balanced; when the surface area to volume ratio of the parallel computing area is minimized, the scale of message communication between each parallel computing core can be reduced.

[0022] Further, in some embodiments, in the step d), the boundary flow weighted by the second-order Legendre polynomial The calculation method is:

[0023]

[0024] Among them, φ(x0,y,z) is the flux value at the coordinate (x0,y,z), and is a second-order Legendre polynomial, Δy k is the width of node k in the y direction, Δz k is the node width of node k in the z direction, and x0 is the coordinate value in the x direction.

[0025] Further, in some embodiments, in step g), the expansion coefficient of the lateral leakage term is The calculation method is:

[0026]

[0027] in, is the average leakage term of node k, is the boundary flow weighted by the second-order Legendre polynomial at the left boundary of node k, is the boundary flow weighted by the second-order Legendre polynomial at the right boundary of node k, Δx k is the width of node k in the x direction.

[0028] Furthermore, in some embodiments, in the step i), the eigenvalue k eff The calculation method is:

[0029] The total number of fission neutrons produced by all the nodes assigned to each parallel computing core is calculated. Where g is the neutron energy group number, is the fission neutron production cross section of the g-th energy group of node k, φ k,g is the neutron flux of the g-th energy group in node k, V k is the volume of node k, the eigenvalue

[0030] Further, in some embodiments, in the step k), the convergence criterion is the neutron flux, the neutron source term, the boundary neutron flow, the boundary flow weighted by the second-order Legendre polynomial and the eigenvalue k in the two rounds of calculation. eff The difference is less than a given threshold.

[0031] According to an embodiment of another aspect of the present invention, there is provided a computing device, comprising a memory and a plurality of parallel computing cores, wherein the memory stores a neutron transport parallel computing program, and when the neutron transport parallel computing program is executed by the parallel computing cores, the neutron transport parallel computing method for optimizing lateral leakage term processing provided in any of the aforementioned embodiments can be implemented.

[0032] Furthermore, in some embodiments, the computing program includes a modeling and initialization module, a node response calculation module, a high-order leakage term calculation module based on a single node model, a source term and eigenvalue solution module, and a parallel scheduling module, wherein:

[0033] The modeling and initialization module is used to read the input information of the calculation object, perform node division, node coloring, parallel region decomposition and calculation result initialization;

[0034] The node response calculation module is used to define the boundary flow weighted by the second-order Legendre polynomial, and according to the node response matrix, the neutron source term and the neutron flow data in the previous iteration are used as input to solve the node boundary neutron flow, the node average leakage term and the boundary flow weighted by the second-order Legendre polynomial;

[0035] The high-order leakage term calculation module based on the single-node model is used to solve the expansion coefficient of the lateral leakage term by taking the average leakage term of a single node and the boundary flow weighted by the second-order Legendre polynomial at the node boundary as constraint conditions;

[0036] The source term and eigenvalue solving module is used to solve the neutron flux, eigenvalue k according to the neutron balance equation, taking the node boundary neutron flow, the average boundary flow weighted by the second-order Legendre polynomial, and the lateral leakage term expansion coefficient as input. eff and neutron source terms. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 A schematic diagram of a computing object structure in an embodiment;

[0038] Figure 2 This is a schematic diagram of the block coloring in one embodiment;

[0039] Figure 3 A schematic diagram of node block allocation in an embodiment;

[0040] Figure 4 A schematic diagram of a parallel computing process in an embodiment;

[0041] Figure 5 Schematic diagram of a parallel computing program module in one embodiment.

[0042] The purpose of the above drawings is to explain the present invention in detail so that those skilled in the art can understand the technical concept of the present invention, but it is not intended to limit the present invention. For the sake of simplicity, the above drawings only schematically illustrate the structures related to the technical features of the present invention, and do not strictly follow the actual proportions to draw the complete structure and all details. DETAILED DESCRIPTION

[0043] The present invention is further described in detail below through specific embodiments in conjunction with the accompanying drawings.

[0044] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present invention. The phrase appearing in various locations in the specification does not necessarily refer to the same embodiment, nor is it limited to mutually exclusive independent or alternative embodiments. Those skilled in the art should be able to understand that the embodiments herein may be combined with other embodiments without causing structural conflicts.

[0045] In the description of this article, the terms "first", "second" and the like are only used to distinguish different objects, and cannot be understood as indicating relative importance or limiting the quantity, specific order or primary and secondary relationship of the described technical features. In the description of this article, the meaning of "plurality" is at least two.

[0046] In order to solve the defects of low calculation accuracy and poor calculation stability of transverse leakage term in existing neutron transport calculation, an embodiment of the present invention provides a neutron transport parallel calculation method for optimizing transverse leakage processing. The method adopts a high-order second-order parabola approximation of the transverse leakage term, which can improve the calculation accuracy of key nuclear characteristic parameters such as eigenvalues, power distribution, and power peak factor; at the same time, by introducing the boundary flow weighted by the second-order Legendre polynomial, the high-order expansion coefficient of the leakage term can be calculated based on a single node model, thereby improving the stability of the parallel calculation of neutron transport in nuclear reactor nodes based on the regional decomposition method.

[0047] like Figure 4 As shown, the steps of the method are as follows:

[0048] Step 1: Provide a computational model for the nuclear reactor neutron transport computation object, such as Figure 1 As shown, the computational model specifically includes geometric information (i.e., geometric structure and size characteristics), material information (material type and distribution), nuclear reaction cross-section information, etc. of the computational object, while providing parallel computing information and control parameters of the system for parallel computing; the computational object is divided into nodes (grids) according to the input information.

[0049] Step 2: If Figure 2 As shown in the figure, the nodes are color-coded from 1 to M to ensure that the colors of adjacent nodes are different; according to the parallel computing information input data, the parallel region decomposition is carried out to make the load of each parallel computing core tend to be balanced, specifically by controlling the load variance of each computing core to reach the minimum value. In order to reduce the scale of message communication between each parallel computing core, the principle of regional decomposition is to minimize the surface area to volume ratio of the parallel region as much as possible. The nodes with the same color mark are regarded as a same-color node group.

[0050] Step 3: If Figure 3 As shown, according to the parallel domain decomposition result in step 2, each parallel computing core performs node response matrix calculation on the node to which it is assigned, and gives the initialized neutron flux, neutron flow, neutron source term and eigenvalue k eff Specifically, the initial value can be experimental data, historical simulation calculation data or other artificially given initial values.

[0051] Step 4: For the nodes marked with the first color, each parallel computing core calculates and updates the node boundary neutron flow, node average leakage term, and boundary flow weighted by the second-order Legendre polynomials based on the node response matrix in step 3 and the initial value (in the first iteration) or the neutron source term and neutron flow of the previous iteration (in the second and subsequent iterations) for the nodes assigned to it.

[0052] Among them, the boundary flow weighted by the second-order Legendre polynomial of node k is The calculation method is:

[0053]

[0054] Among them, φ(x0,y,z) is the flux value at the coordinate (x0,y,z), and is a second-order Legendre polynomial, Δy k is the width of node k in the y direction, Δz k is the node width of node k in the z direction, and x0 is the coordinate value in the x direction.

[0055] Step 5: Each parallel computing core carries out message communication between the first color node group and its adjacent nodes at the interface of the parallel region for the node assigned to it, and updates the latest node boundary neutron flow and second-order Legendre polynomial weighted boundary flow information to the adjacent nodes.

[0056] Step 6: Each parallel computing core, for the nodes assigned to it, performs boundary neutron flow, node average leakage term, and boundary flow weighted by second-order Legendre polynomial calculations for the same-color node groups of color 2 to color M that have not been calculated, and updates the boundary neutron flow and boundary flow weighted by second-order Legendre polynomial calculation results to the adjacent nodes. This continues until all the same-color node groups of all colors have performed boundary neutron flow, node average leakage term, and boundary flow weighted by second-order Legendre polynomial calculations.

[0057] Step 7: Each parallel computing core uses the second-order polynomial approximation of the lateral leakage term for the node to which it is assigned, and uses the average leakage term of a single node k and the boundary flow weighted by the second-order Legendre polynomial at the node boundary as constraints to solve the expansion coefficient of the lateral leakage term

[0058] in, is the average leakage term of node k, is the boundary flow weighted by the second-order Legendre polynomial at the left boundary of node k, is the boundary flow weighted by the second-order Legendre polynomial at the right boundary of node k, Δx k is the width of node k in the x direction.

[0059] Step 8: Each parallel computing core calculates and updates the neutron flux of the node to which it is assigned, based on the neutron balance equation as a constraint, and with the neutron source term and the node boundary neutron flow as input.

[0060] Step 9: Each parallel computing core calculates the total number of fission neutrons produced for all nodes assigned to it. Where g is the neutron energy group number, is the fission neutron production cross section of the g-th energy group of node k, φ k,g is the neutron flux of the g-th energy group in node k, V k is the volume of node k, the eigenvalue The eigenvalue k is transmitted through message communication eff Broadcast to all computing cores for updates.

[0061] Step 10: Each parallel computing core updates the neutron source terms for the nodes assigned to it, including data such as fission sources, scattering sources, and total source terms, based on the boundary neutron current, second-order Legendre polynomial weighted boundary current, neutron flux, and transverse leakage term expansion coefficients calculated in steps 4 to 9.

[0062] Step 11: Update the boundary neutron flux, the boundary flux weighted by the second-order Legendre polynomial, the nodal neutron flux and the eigenvalue k eff As input, repeat steps 4 to 10 for iterative calculation until the iterative calculation results converge or the iterative calculation reaches a given number of times.

[0063] The convergence criteria include that the calculation deviation of neutron flux, neutron source term, neutron flow and eigenvalue is less than the specified value. If there is still no convergence after the specified number of iterations, it means that the calculated state itself may be divergent.

[0064] The final calculation results are output to complete the neutron transport calculation.

[0065] The method provided in the above embodiment can be implemented by a parallel computing device. The parallel computing device includes a memory and a plurality of computing cores, wherein the memory stores a neutron transport calculation program, and when the neutron transport calculation program is executed by the computing core, the neutron transport parallel calculation method for optimizing the lateral leakage term processing provided in the above embodiment can be implemented.

[0066] Specifically, the neutron transport calculation program is as follows Figure 5 As shown, it includes a modeling and initialization module, a node response calculation module, a high-order leakage term calculation module based on a single node model, a source term and eigenvalue solution module, and a parallel scheduling module.

[0067] Among them, the modeling and initialization module is used to read the input information of the calculation object, perform node (grid) division, node coloring, parallel region decomposition, and calculation result initialization.

[0068] The node response calculation module defines a boundary flow weighted by a second-order Legendre polynomial. Based on the node response matrix, the module takes the neutron source and neutron flow of the previous iteration as input to solve the node boundary neutron flow, the node average leakage term, and the boundary flow weighted by the second-order Legendre polynomial.

[0069] The high-order leakage term calculation module based on the single-node model solves the lateral leakage term expansion coefficient with the average leakage term of a single node and the boundary flow weighted by the second-order Legendre polynomial at the node boundary as constraints.

[0070] The source term and eigenvalue solving module solves the neutron flux, eigenvalue and neutron source term according to the constraints such as the neutron balance equation, with the block boundary neutron flow, the boundary flow weighted by the second-order Legendre polynomial, and the expansion coefficient of the lateral leakage term as input.

[0071] Parallel scheduling module: carries out parallel scheduling, message communication, result collection and broadcasting of nuclear reactor neutron transport calculations.

[0072] The purpose of the above embodiments is to further explain the present invention in detail in conjunction with the accompanying drawings so that those skilled in the art can understand the technical concept of the present invention. Within the scope of the present invention, the optimization or equivalent replacement of the method steps involved, and the combination of the implementation methods in different embodiments without conflict of structure and principle, all fall within the protection scope of the present invention.

Claims

1. A neutron transport parallel computing method for optimizing lateral leakage term processing, characterized in that: The following steps are involved: Step a): providing a simulation model of a reactor neutron transport calculation simulation object, and dividing the simulation model into blocks; Step b): color-marking the nodes and making adjacent nodes have different colors, and treating the nodes with the same color marking as a same-color node group to carry out parallel regional decomposition; Step c): Allocate the nodes to multiple computing cores according to the parallel domain decomposition result in step b), perform node response matrix calculation, and provide neutron flux, neutron flow, neutron source term and eigenvalue k eff The initial value of Step d): calculating the boundary neutron flow, the node average leakage term and the boundary flow weighted by the second-order Legendre polynomial of the node in the node group respectively through each of the computing cores; Step e): updating the boundary flow calculated in step d) to the adjacent nodes; Step f): the node whose boundary neutron flow, the node average leakage term and the boundary flow weighted by the second-order Legendre polynomial have been calculated is taken as the calculated node, and the remaining nodes are taken as the uncalculated nodes; In the uncalculated nodes where the boundary flow is updated, the boundary neutron flow, the node average leakage term and the boundary flow weighted by the second-order Legendre polynomial of each node group of the same color are calculated one by one, and the calculated boundary neutron flow and the boundary flow weighted by the second-order Legendre polynomial are updated to the adjacent nodes; Step g): using a second-order polynomial approximation of the lateral leakage term, and taking the average leakage term of each node and the second-order Legendre polynomial weighted boundary flow at the node boundary as constraints, calculating the lateral leakage term expansion coefficient; Step h): according to the neutron balance equation, taking the neutron source term and the nodal boundary neutron flux as input, calculating and updating the nodal neutron flux; Step i): Calculate the characteristic value k according to the neutron flux of the node calculated in step h). eff , and update to all computing cores; Step j): updating the neutron source term according to the boundary neutron flow term of the calculated node, the boundary flow weighted by the second-order Legendre polynomial, the expansion coefficient of the lateral leakage term calculated in step g), and the node neutron flux calculated in step h); Step k): according to the updated boundary neutron flow and the boundary flow weighted by the second-order Legendre polynomial, the nodal neutron flux and the eigenvalue k eff , repeat step d) to step j) until the calculation result converges or reaches a given number of iterations, and output the final calculation result.

2. The neutron transport parallel computing method for optimizing lateral leakage term processing according to claim 1 is characterized in that: In the step a), the simulation model includes geometric information, material information, and nuclear reaction cross-section information.

3. The neutron transport parallel computing method for optimizing lateral leakage term processing according to claim 1, characterized in that: In the step b), during the process of performing the regional decomposition, the load variance of each of the parallel computing cores is minimized, and the surface area to volume ratio of the parallel computing region is minimized.

4. The neutron transport parallel computing method for optimizing lateral leakage term processing according to claim 1, characterized in that: In the step d), the boundary flow weighted by the second-order Legendre polynomial The calculation method is: Among them, φ(x0,y,z) is the flux value at the coordinate (x0,y,z), and is a second-order Legendre polynomial, Δy k is the width of node k in the y direction, Δz k is the node width of node k in the z direction, and x0 is the coordinate value in the x direction.

5. The neutron transport parallel computing method for optimizing lateral leakage term processing according to claim 1, characterized in that: In step g), the lateral leakage term expansion coefficient The calculation method is: in, is the average leakage term of node k, is the boundary flow weighted by the second-order Legendre polynomial at the left boundary of node k, is the boundary flow weighted by the second-order Legendre polynomial at the right boundary of node k, Δx k is the width of node k in the x direction.

6. The neutron transport parallel computing method for optimizing lateral leakage term processing according to claim 1, characterized in that: In the step i), the eigenvalue k eff The calculation method is: The total number of fission neutrons produced by all the nodes assigned to each parallel computing core is calculated. Where g is the neutron energy group number, is the fission neutron production cross section of the g-th energy group of node k, φ k,g is the neutron flux of the g-th energy group in node k, V k is the volume of node k, the eigenvalue 7. The neutron transport parallel computing method for optimizing lateral leakage term processing according to claim 1, characterized in that: In the step k), the convergence criterion is the neutron flux, the neutron source term, the boundary neutron flow, the boundary flow weighted by the second-order Legendre polynomial and the eigenvalue k in the two rounds of calculation. eff The difference is less than a given threshold.

8. A computing device, comprising a memory and a plurality of parallel computing cores, characterized in that: The memory stores a computing program, and when the computing program is executed by the parallel computing core, the neutron transport parallel computing method for optimizing lateral leakage term processing as claimed in any one of claims 1 to 7 can be implemented.

9. The computing device according to claim 8, characterized in that The calculation program includes a modeling and initialization module, a node response calculation module, a high-order leakage term calculation module based on a single node model, a source term and eigenvalue solution module, and a parallel scheduling module, wherein: The modeling and initialization module is used to read the input information of the calculation object, perform node division, node coloring, parallel region decomposition and calculation result initialization; The node response calculation module is used to define the boundary flow weighted by the second-order Legendre polynomial, and according to the node response matrix, the neutron source term and the neutron flow data in the previous iteration are used as input to solve the node boundary neutron flow, the node average leakage term and the boundary flow weighted by the second-order Legendre polynomial; The high-order leakage term calculation module based on the single-node model is used to solve the expansion coefficient of the lateral leakage term by taking the average leakage term of a single node and the boundary flow weighted by the second-order Legendre polynomial at the node boundary as constraint conditions; The source term and eigenvalue solving module is used to solve the neutron flux, eigenvalue k according to the neutron balance equation, taking the node boundary neutron flow, the average boundary flow weighted by the second-order Legendre polynomial, and the lateral leakage term expansion coefficient as input. eff and neutron source terms; The parallel scheduling module is used for parallel scheduling, message communication, and result collection and broadcasting of neutron transport calculations.

Images