WO2021249374A1 - 特征线追踪方法、堆芯中子物理计算方法和装置 - Google Patents

特征线追踪方法、堆芯中子物理计算方法和装置 Download PDF

Info

Publication number
WO2021249374A1
WO2021249374A1 PCT/CN2021/098806 CN2021098806W WO2021249374A1 WO 2021249374 A1 WO2021249374 A1 WO 2021249374A1 CN 2021098806 W CN2021098806 W CN 2021098806W WO 2021249374 A1 WO2021249374 A1 WO 2021249374A1
Authority
WO
WIPO (PCT)
Prior art keywords
sub
model
intersection
core
characteristic line
Prior art date
Application number
PCT/CN2021/098806
Other languages
English (en)
French (fr)
Inventor
杨睿
胡赟
单浩栋
汪岸
杨文�
徐李
胡长军
李龙
Original Assignee
中国原子能科学研究院
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 中国原子能科学研究院 filed Critical 中国原子能科学研究院
Priority to EP21755885.7A priority Critical patent/EP3951634A4/en
Priority to US17/435,273 priority patent/US20220336116A1/en
Publication of WO2021249374A1 publication Critical patent/WO2021249374A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00Computer-aided design [CAD]
    • G06F30/20Design optimisation, verification or simulation
    • GPHYSICS
    • G21NUCLEAR PHYSICS; NUCLEAR ENGINEERING
    • G21DNUCLEAR POWER PLANT
    • G21D3/00Control of nuclear power plant
    • G21D3/001Computer implemented control
    • G21D3/002Core design; core simulations; core optimisation
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00Computer-aided design [CAD]
    • G06F30/10Geometric CAD
    • G06F30/17Mechanical parametric or variational design
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F2111/00Details relating to CAD techniques
    • G06F2111/10Numerical modelling
    • YGENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02EREDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E30/00Energy generation of nuclear origin

Definitions

  • the invention relates to the technical field of reactors, in particular to design a characteristic line tracking method for a core model, a core neutron physical calculation method and device, and a computer-readable storage medium.
  • a neighbor search method is used in the Monte Carlo program. After the judgment of one point is completed, the judgment starting point of the next point no longer starts from the top-level geometry, but starts from the previous point. In this method If it is not found in the multiple geometry at the position of the previous point, you need to return to the geometry of the previous level for judgment. There are still many repeated calculations, and the efficiency is still low. In addition, due to the Monte Carlo method There is intersection operation in the module process, and the spatial range of the underlying geometry may not be the real spatial range. Therefore, the geometry needs to be pre-processed before searching from the underlying geometry.
  • the present invention is proposed to provide a feature line tracking method for core model, core neutron physics calculation method and device, reactor model, reactor simulation device, which overcomes the above problems or at least partially solves the above problems , And a computer-readable storage medium.
  • a feature line tracking method for a core model the core model being used to divide the space range of the core into a plurality of levels each having a plurality of sub-models, wherein, At least one of the sub-models in the same level of the multiple levels belongs to one of the sub-models in the upper level, and each of the sub-models in the lowermost level includes only one material of the core, each All the sub-models of the upper level are combined and intersected with the upper-level space range to form the upper-level sub-model, the uppermost level is the core model, and the characteristic line tracking method includes: sequentially from top to bottom by level Determine whether the intersection of the sub-model that meets the first preset condition and the characteristic line in each of the levels meets the second preset condition; obtain the sub-model that meets the first preset condition in the lowest level An intersection with the characteristic line that meets the second preset condition, and the segment correspondence relationship between the characteristic line and the material of the core is obtained according to the obtained
  • a core neutron physics calculation method including: a modeling step: constructing a core model, the core model is used to divide the space range of the core into a plurality of sub Multiple levels of models, wherein at least one of the sub-models in the same level of the multiple levels belongs to one of the sub-models in the upper level, and each of the sub-models in the lowermost level includes only one type
  • the characteristic line tracing method includes Tracking step: defining a characteristic line and tracking the characteristic line according to any one of the above-mentioned characteristic line tracking methods; iterative calculation step: performing iterative calculation based on the tracking result of the characteristic line to complete the core Solving the sub-transport equation.
  • a reactor model constructed using the core neutron physics calculation method as described in the second aspect of the present invention.
  • a core neutron physics calculation device including: a modeling module to construct a core model, and the core model is used to divide the space range of the core into a plurality of sub-fields. Multiple levels of models, wherein at least one of the sub-models in the same level of the multiple levels belongs to one of the sub-models in the upper level, and each of the sub-models in the lowermost level includes only one type
  • a feature line tracking module which defines features The characteristic line is tracked according to any one of the above-mentioned characteristic line tracking methods; the iterative calculation module performs iterative calculation based on the tracking result of the characteristic line to complete the solution of the neutron transport equation of the core.
  • a reactor simulation device including the core neutron physics calculation device according to the fourth aspect of the present invention.
  • a computer-readable storage medium wherein the computer-readable storage medium stores instructions for realizing the core neutron physics calculation method described above.
  • the feature line tracking method according to the embodiment of the present invention is relatively simple, reduces the amount of calculation, and improves the calculation efficiency.
  • the tracking method by level also avoids performing geometric preprocessing on the core model.
  • Fig. 1 is a schematic diagram of a core model according to an embodiment of the present invention.
  • Fig. 2 is a schematic diagram of a tracking method according to an embodiment of the present invention.
  • Fig. 3 is a schematic diagram of a basic unit calculation tree according to an embodiment of the present invention.
  • Fig. 4 is an auxiliary function image according to an embodiment of the present invention.
  • Fig. 5 is a schematic diagram of a sub-solution set intersection or union operation method according to an embodiment of the present invention.
  • FIG. 6 is a schematic diagram of the Takeda problem in the prior art
  • Fig. 7 is a schematic diagram of the C5G7 problem in the prior art.
  • This embodiment first provides a feature line tracking method for a core model.
  • the core model is used to divide the space range of the core into multiple levels with multiple sub-models, that is, there are multiple sub-models in the core model. There are multiple levels, and each level has multiple submodels.
  • At least one sub-model of the same level belongs to a sub-model in the upper level, that is, in a level, after a sub-model or multiple sub-models are combined, they are intersected in the upper level space. You will get the sub-model of the next level.
  • Each sub-model at the lowest level includes only one material of the core. It can be understood that through the above calculations, as the levels continue to rise, more and more core materials will be included in each sub-model. Therefore, all sub-models at the top level can correspond to the space range of the core and include all the materials of the core after calculation.
  • modeling is mainly to describe the material information of each position in a three-dimensional space.
  • This sub-model can also be called a primitive, abbreviated as (U, m), it is the simplest model unit.
  • the spatial range it represents is mutually exclusive, then the simultaneous f 1 , f 2 ,..., f n can form a new mapping f:
  • This kind of operation is called union operation, Called the union of G 1 , G 2 ,..., G n , it is equivalent to assembling several sub-models of the next level together.
  • intersection operation U′ ⁇ G is called the intersection of the space range U′ and the geometric body G.
  • the essence of the intersection operation is to use a space range to cut the graphics, and the actual effect is the space range of U' ⁇ U.
  • the spatial extent have:
  • the specific implementation is the same as the Monka program, using two data structures, Cell and Universe.
  • Cell represents a geometric model
  • the user is required to give a space range U and material index m, that is, the user needs to set the lowest level sub-model that includes only one material:
  • the Universe requires the user to give a list of several geometric models ⁇ Cell down, i ⁇ , which represents the union of these models, that is, represents the union operation of multiple sub-models:
  • Figure 1 shows a simple schematic diagram of the core model.
  • the parallelogram box represents the material
  • the rounded box represents the spatial range U
  • the box represents the sub-model
  • the hexagon box represents the universe
  • the solid line represents the union
  • the dashed line represents the intersection.
  • the fuel, cladding, and moderator of the lowest level sub-models include only one material; these lowest level sub-models generate and aggregate cell 3, and intersect with the spatial range of cell 3 to form the sub-models of the upper level.
  • Model cell 3 that is, constitutes a sub-model of the cell level.
  • component 2 which is a sub-model in the component level, that is, the top-level sub-model in the model, and the top-level sub-model finally constitutes the entire core Model
  • component 2 is a sub-model in the component level, that is, the top-level sub-model in the model, and the top-level sub-model finally constitutes the entire core Model
  • the feature line tracking method includes: Whether the intersection of the sub-model that meets the first preset condition and the characteristic line in the hierarchy meets the second preset condition; obtain the sub-model and the sub-model that meet the first preset condition in the lowest layer The intersection of characteristic lines that meets the second preset condition, and the segment correspondence relationship between the characteristic line and the material of the core is obtained according to the obtained intersection.
  • the traversal of the core model is only carried out to the lower layer, that is to say, in the entire feature line tracking, the segment-by-section tracking algorithm of the prior art is avoided The step of returning to the upper-level node for traversal in the middle, thereby saving the number of calculations and improving efficiency.
  • the segment-by-segment tracking algorithm in the prior art needs to store segment information when performing real-time tracking. Even if the feature line tracking method of the present invention performs real-time tracking, it only needs to double the calculation time, which avoids The storage of segment information saves a lot of storage space.
  • the first preset condition corresponding to the uppermost level is: being located at the uppermost level; the first preset condition corresponding to the non-uppermost level is: conforming to the intersection of the characteristic line in the upper level of the non-uppermost level
  • the sub-model corresponding to the second preset condition corresponds.
  • Those skilled in the art can understand that when judging the top level, all the sub-models of the top level are judged. When judging the non-top level, it is the part of the sub-models in the non-top level. This part of the sub-model is NAND.
  • Certain spatial ranges of the upper level of the hierarchy correspond to certain spatial ranges, and the certain spatial ranges refer to the sub-models in the upper level that is not the uppermost level whose intersection with the characteristic line meets the second preset condition.
  • each lowest-level sub-model whose intersection with the feature line meets the second preset condition will be obtained. Since each lowest-level sub-model only includes one material, the feature can be obtained. Correspondence between lines and core materials.
  • the second preset condition is that the intersection is not an empty set, that is, in the judgment process of each level, sub-models that do not have an intersection with the characteristic line will be excluded. For example, when calculating the component level in Figure 2, If the intersection of sub-model component 1 and the characteristic line is an empty set, no calculation is performed on the sub-model corresponding to component 1 at the next level.
  • the core model of Figure 2 when tracking the core model of Figure 2, first calculate the sub-models in the top level that meet the first preset condition. For the top level, all sub-models meet the first preset. Condition, the intersection of the feature line and sub-model component 1, sub-model component 2, sub-model component 3..., and determine whether the intersection meets the second preset condition, that is, filter out the sub-models in the hierarchy that are not empty at the intersection with the feature line Model. Further, after the calculation of the component level is completed, for example, the result is that the sub-model component 2 meets the second preset condition, and the calculation of the next level is performed.
  • the sub-model that meets the first preset condition refers to its corresponding
  • the sub-model of the upper level meets the second preset condition, that is, the sub-model cell 1, cell 2, etc. corresponding to the sub-model component 2 meet the first preset condition. In this level, only these The cell makes a judgment. Similarly, if only the intersection of cell 3 and the characteristic line is not an empty set, then calculate the sub-model fuel, reflective layer, and cladding at the next layer of sub-model cell 3, because the fuel and reflective layer core cladding are the lowest After the calculation of the hierarchical sub-model is completed, the corresponding relationship between the characteristic line and the material can be obtained.
  • the obtained intersections that are not empty sets and the sub-models of these intersections can be sorted according to predetermined rules.
  • this predetermined rule can be set according to the position of the elements in these intersections on the feature line.
  • a more preferred method of calculating the intersection of a characteristic line and a sub-model is used, and the calculation method includes: constructing an auxiliary function, and the auxiliary function is used to calculate the corresponding sub-model in the Multiple boundary points on the characteristic line; calculate the left limit of each boundary point on the auxiliary function; when the left limit satisfies the third preset condition, determine that the boundary point and the adjacent and the The line segment between the other boundary points corresponding to the left limit is in the sub-model.
  • the derivation process of the auxiliary function will be explained below.
  • Sub-models are usually more complicated, and the complex spatial scope can be described by the following simple set operations between basic units by using the structural solid geometry method.
  • the so-called basic unit refers to some basic geometric bodies that can be expressed by mathematical formulas, such as cylinders, spheres, cubes, etc. These geometric bodies can be used to describe the relevant structure in the core, and the basic units can be expressed as ⁇ r
  • the equation of the basic unit used in the sub-model is generally less than or equal to quadratic, and its unified form can be written as:
  • a ⁇ B ⁇ r
  • the complement operation the complement of the spatial range closure is used, denoted as Among them, cl is the symbol of closure, and the above form is used to ensure that the calculation result is still an open interval, such as and Furthermore, the complement operation here satisfies:
  • the calculation process can be drawn as a tree diagram according to the order of the operations.
  • the tree diagram in Figure 3 represents the operation as
  • leaf nodes in Figure 3 are a basic unit, and the other nodes are a kind of operators.
  • the complement operation has one subtree, and the intersection operation has two subtrees, that is, it needs to be performed by two basic units.
  • l is the length of the characteristic line segment
  • r 0 is the starting point of the characteristic line segment
  • is the direction vector of the characteristic line segment
  • the Rvachev function is introduced to further describe the sub-model, and the sub-model based on the geometric description of the construction entity is described as an inequality form of implicit numbers. Specifically, for a single basic unit, it can be expressed as
  • R(r)>0 means in U
  • R(r) ⁇ 0 means in Within
  • sgn is a symbolic function, that is, the left and right ends of t must have different signs. In this way, one end must be inside the body and one section outside the body, thus forming a boundary point. To do this, define an auxiliary function:
  • this auxiliary function greatly simplifies the difficulty of the transfer and complement operation.
  • this is a quadratic equation of one variable, which is easy to solve and expressed as ⁇ and ⁇ t 1 ,...,t n ⁇ .
  • segment ⁇ (U 1 ⁇ U 2 ) (segment ⁇ U 1 ) ⁇ (segment ⁇ U 2 )
  • segment ⁇ (U 1 ⁇ U 2 ) (segment ⁇ U 1 ) ⁇ (segment ⁇ U 2 )
  • intersection operation it is necessary to make a judgment, that is, first use the above auxiliary function to solve the intersection of the characteristic line and multiple basic units, obtain multiple sub-solution sets, and then calculate these based on the intersection operation relationship between these basic units
  • the sub-solution sets perform corresponding intersection or union operations until the final solution set is obtained.
  • the sub-model may be obtained by multiple operations of multiple basic units. In these cases, it is necessary to perform operations on the corresponding sub-solution sets in the order of operations performed by the basic units. Operation, for example, referring to the calculation tree in Figure 3, to obtain the solution set corresponding to sub-model Y, you need to perform operations on A and B to obtain the sub-solution set corresponding to I, and perform operations on C and D to obtain the sub-solution corresponding to Y Set, and so on, until all calculations in the calculation tree are completed to obtain the solution set corresponding to the sub-model Y.
  • the third line of the first term in (2) is to ensure
  • a function value is provided.
  • a method for judging all points at one time includes: setting the first pointer and the second pointer to point to the elements on the leftmost side of the number line in the two sub-solution sets; When the element pointed to by the first pointer or the second pointer meets the fourth preset condition at the left limit of the corresponding auxiliary function, the element pointed to by the first pointer or the second pointer on the relatively left side of the number line is included in the solution of the intersection or union operation Set, and move the first pointer or second pointer on the relatively left to the right to point to the next element; when the element pointed to by the first pointer and the second pointer, calculate the corresponding auxiliary element in the two sub-solution sets.
  • the left limit of the function when the left limit satisfies the fifth preset condition, the element is included in the solution set of the intersection or union operation, and the first pointer or the second pointer is moved to the right of the number axis to point to The next element; repeat the above-mentioned solution set obtaining step until the solution set obtained by the intersection or union operation of the two sub-solution sets is obtained.
  • this judgment method takes advantage of the orderliness of ⁇ t 1,i ⁇ and ⁇ t 2,j ⁇ .
  • the first pointer P1 and the second pointer P2 are set, and initially they point to The first element of two sequence of ordinal numbers, if the sequence is empty, it points to infinity; use and Indicates the left limit of the elements pointed to by the pointers P1 and P2 on the corresponding auxiliary function, and the initial values are ⁇ 1 and ⁇ 2 .
  • the above solution set acquisition step ends after P1 and P2 move to infinity, that is, the judgment of all elements is completed.
  • the time complexity of a single merge operation is reduced to O(n).
  • the calculation time of the complement operation can be ignored.
  • the total number of intersection operations is related to the calculation tree of the basic unit. Assuming that the intersection and merge operations have a total of b layers, the maximum number of calculations is when the binary tree is full and no points are deleted. The number of calculations O((b-1) ⁇ 2 b-1 ).
  • the level Cell is the lowest level Cell (U, m), that is, each sub-model in the level Cell only includes one material, then:
  • intersection of the characteristic line segment and the spatial range U down,i of a sub-model Cell down,i in the level Cell that meets the first preset condition can be obtained by performing an intersection operation:
  • J i is the space range U down of the characteristic line segment quilt model , and the number of line segments formed after i is cut.
  • Cell is a level
  • I is the number of sub-models in the cell of the level that meet the first preset condition.
  • segment ⁇ Cell U(segment i ,m i )
  • the main calculation time is concentrated on the calculation of the line segment of the characteristic line and the sub-model.
  • the number of calculations is equal to the sum of the number of all sub-model lists passed by the characteristic line. If the entire calculation is
  • the model can be represented by a binary tree, so the number of calculations can be estimated to be twice the number of segments. If it is considered that there are 4 basic units in each of the lowest-level sub-models on average, it is equivalent to the E index calculation time of the number of segments ⁇ 8 times.
  • This embodiment also provides a method for calculating core neutron physics, including:
  • Modeling step constructing a core model, which is used to divide the space range of the core into a plurality of levels each having a plurality of sub-models, wherein at least one of the same levels of the plurality of levels
  • the sub-models constitute one of the sub-models in the upper level
  • each of the sub-models in the lower level includes only one material of the core, and all the sub-models in the upper level are combined to form the reactor.
  • Tracking step defining a characteristic line and tracking the characteristic line according to any one of the above-mentioned characteristic line tracking methods
  • Iterative calculation step iterative calculation is performed based on the tracking result of the characteristic line to complete the solution of the neutron transport equation of the core.
  • This embodiment also provides a core neutron physics calculation device, including:
  • a modeling module a component core model, which is used to divide the space range of the core into a plurality of levels each having a plurality of sub-models, wherein at least one of the same levels of the plurality of levels
  • the sub-models constitute one of the sub-models in the upper level
  • each of the sub-models in the lower level includes only one material of the core, and all the sub-models in the upper level are combined to form the reactor.
  • a characteristic line tracking module which defines a characteristic line and tracks the characteristic line according to any one of the above-mentioned characteristic line tracking methods
  • the iterative calculation module performs iterative calculation based on the tracking result of the characteristic line to complete the solution of the neutron transport equation of the core.
  • the core neutron physics calculation method and device of the present invention use any one of the above-mentioned characteristic line tracking methods, thereby saving calculation steps and improving calculation efficiency.
  • This embodiment also provides a reactor model.
  • the reactor model can be used to calculate and describe the state of various physical quantities in the reactor.
  • the neutron physics calculation method is mainly used to calculate and describe the neutron flux density and reactivity related to neutronics. And reaction rate distribution and other physical quantities.
  • the construction of reactor model generally includes modeling and numerical processing. Modeling is to input the reactor model to the computer through a given method, and numerical processing is to convert the input model into the form required for calculation.
  • the reactor model of the embodiment of the present invention uses any of the above-mentioned core neutron physical calculation methods to complete the numerical processing steps after the modeling, so as to save the calculation amount in the modeling process and improve the calculation efficiency.
  • This embodiment also provides a reactor simulation device.
  • the reactor simulation device can be used to simulate the actual operating conditions of the reactor. The more important one is to simulate and calculate the neutron physics of the core. Therefore, this implementation provides a core that includes any of the above
  • the reactor simulation device of the neutron physics calculation device can save calculation steps and improve calculation efficiency when the reactor simulation device is used for reactor simulation.
  • This embodiment also provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions, and the instructions are used to implement the above-mentioned core neutron physics calculation method.
  • has the same sign as A;
  • the specific embodiment also provides a tracking method.
  • the specific embodiment also provides that the tracking method according to the present invention is tested and calculated on the benchmark questions Takeda and C5G7.
  • the filling material in the red part of case1 is empty, and the filling material in the red part of case2 is the control rod.
  • the parameters of the calculation are: the grid size is 0.5cm, the ray density is 0.05cm ⁇ 0.05cm, the Legendre-Chebyshev quadrature group is selected as the integration group, 8 polar angles and 16 azimuth angles.
  • the time ratio refers to the characteristic line tracking time divided by the inner iteration time. In our test, if the Takeda problem is not accelerated, it needs 370 iterations. At this time, the feature line tracking time only takes 1.66% of the time. If the method of generating characteristic lines in real time is adopted, only 61.58% of the original calculation time is needed to avoid all the segment information storage. In order to better illustrate the tracking effect, the number of equal power groups is proposed, that is, the time ratio multiplied by the number of energy groups. Because geometric tracking has nothing to do with the number of energy groups, even if there are more energy groups, the number of equal power groups will not change. So if the number of energy groups for the Takeda problem is increased to 8, the number of equal-efficiency groups is still 1.23, and the time ratio can be further reduced to 15.38%.
  • the tracking efficiency is as follows:
  • the tracking efficiency has declined, mainly because the structure is much more complex and there are quadratic surfaces.
  • it satisfies the estimation that the original tracking time is equivalent to the calculation time within a single iteration, can realize fast feature line tracking, has high tracking efficiency, and supports real-time calculation of the feature line at an acceptable time cost.
  • first and second are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features. Therefore, the features defined with “first” and “second” may explicitly or implicitly include at least one of the features. In the description of the present invention, "a plurality of” means at least two, such as two, three, etc., unless otherwise specifically defined.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Geometry (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Computer Hardware Design (AREA)
  • High Energy & Nuclear Physics (AREA)
  • Plasma & Fusion (AREA)
  • Computational Mathematics (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Pure & Applied Mathematics (AREA)
  • Monitoring And Testing Of Nuclear Reactors (AREA)
  • Image Analysis (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

一种特征线追踪方法、堆芯中子物理计算方法及装置以及计算机可读储存介质,所述特征线追踪方法包括:按层级由上到下依次判断每个所述层级中符合第一预设条件的所述子模型与所述特征线线段的交集是否符合第二预设条件;获得最下层级符合所述第一预设条件的所述子模型与所述特征线线段的符合所述第二预设条件的交集,根据获得的所述交集得到所述特征线与所述堆芯的所有材料的分段对应关系。这种特征线追踪方法减小了计算量,提高了计算效率。

Description

特征线追踪方法、堆芯中子物理计算方法和装置 技术领域
本发明涉及反应堆技术领域,特别是设计一种用于堆芯模型的特征线追踪方法、堆芯中子物理计算方法及装置以及计算机可读储存介质。
背景技术
现有技术中的特征线程序,大多采用逐段搜索的方法来进行特征线的追踪,这样的追踪方法的关键步骤为:1、对某点进行定位,2、求解某点沿固定方向到空间范围边界的距离。这两个步骤均需要耗费大量的时间来进行计算,尤其是对某点进行定位的计算。
在对某点进行定位的时候,一种方法需要先从堆芯模型中,堆芯下层的几何体中搜索,一次进行判断,假设发现某点在一个几何体中,则需要在该几何体的下层几何体中再次进行依次判断,直到找到最底层的反射层图元,完成对该点的定位,在对下一个点进行定位的时候,又需要重复上述的步骤,即,在该方法中需要进行大量的重复计算,计算量大,效率较低。
为了改进该方法,在蒙特卡洛程序中使用一种邻居搜索的方法,在一个点判断完成后,下一个点的判断起点不再从顶层几何体出发,而是从前一点出发,在这种方法中,如果在前一点所在位置上的多个几何体中未能找到,则需要返回上一层的几何体中进行判断,仍然有许多重复的计算,效率仍然较低,另外,由于蒙特卡洛方法的建模过程中存在交运算,下层几何体的空间范围未必是真实的空间范围,所以在从下层几何体开始寻找之前,需要对几何体进行预处理。
发明内容
鉴于上述问题,提出了本发明以便提供一种克服上述问题或者至少部分地解决上述问题的用于堆芯模型的特征线追踪方法,堆芯中子 物理计算方法及装置,反应堆模型,反应堆模拟装置,以及一种计算机可读储存介质。
根据本发明的第一个方面,提供一种用于堆芯模型的特征线追踪方法,所述堆芯模型用于将堆芯的空间范围划分成分别具有多个子模型的多个层级,其中,所述多个层级的同一层级中的至少一个所述子模型属于上一层级中的一个所述子模型,最下层级的每个所述子模型仅包括一种所述堆芯的材料,各上层级的所有所述子模型经过并运算及与上层级空间范围的交运算构成上层级子模型,最上层级为所述堆芯模型,所述特征线追踪方法包括:按层级由上到下依次判断每个所述层级中符合第一预设条件的所述子模型与所述特征线的交集是否符合第二预设条件;获得最下层级符合所述第一预设条件的所述子模型与所述特征线的符合所述第二预设条件的交集,根据获得的所述交集得到所述特征线与所述堆芯的材料的分段对应关系。
根据本发明的第二个方面,提供一种堆芯中子物理计算方法,包括:建模步骤:构建堆芯模型,所述堆芯模型用于将堆芯的空间范围划分成分别具有多个子模型的多个层级,其中,所述多个层级的同一层级中的至少一个所述子模型属于上一层级中的一个所述子模型,最下层级的每个所述子模型仅包括一种所述堆芯的材料,各上层级的所有所述子模型经过并运算及与上层级空间范围的交运算构成上层级子模型,最上层级为所述堆芯模型,所述特征线追踪方法包括;追踪步骤:定义特征线并根据上述任一所述的特征线追踪方法对所述特征线进行追踪;迭代计算步骤:基于所述特征线的追踪结果进行迭代计算,完成所述堆芯的中子输送方程的求解。
根据本发明的第三个方面,提供一种反应堆模型,所述反应堆模型使用如本发明第二个方面所述的堆芯中子物理计算方法构建。
根据本发明的第四个方面,提供一种堆芯中子物理计算装置,包括:建模模块,构建堆芯模型,所述堆芯模型用于将堆芯的空间范围划分成分别具有多个子模型的多个层级,其中,所述多个层级的同一层级中的至少一个所述子模型属于上一层级中的一个所述子模型,最下层级的每个所述子模型仅包括一种所述堆芯的材料,各上层级的所 有所述子模型经过并运算及与上层级空间范围的交运算构成上层级子模型,最上层级为所述堆芯模型;特征线追踪模块,定义特征线并根据上述任一所述的特征线追踪方法对所述特征线进行追踪;迭代计算模块,基于所述特征线的追踪结果进行迭代计算,完成所述堆芯的中子输送方程的求解。
根据本发明的第五个方面,提供一种反应堆模拟装置,包括如本发明的第四个方面所述的堆芯中子物理计算装置。
根据本发明的第六个方面,提供一种计算机可读存储介质,其中,所述计算机可读存储介质存储有指令,所述指令用于实现根据上述所述的堆芯中子物理计算方法。
根据本发明实施例的特征线追踪方法较为简单,减少了计算量,提高了计算效率,另外,从按层级的追踪方式也避免了对堆芯模型进行几何预处理。
附图说明
图1为根据本发明的实施例的一种堆芯模型的示意图;
图2为根据本发明的实施例的追踪方法示意图;
图3为根据本发明的实施例的基本单元计算树示意图;
图4为根据本发明的实施例的辅助函数图像;
图5为根据本发明的实施例的子解集交或并运算方法示意图;
图6为现有技术中Takeda问题的示意图;
图7为现有技术中C5G7问题的示意图。
具体实施方式
下面结合附图,对本发明的实施例进行详细地说明。另外,在下面的详细描述中,为便于解释,阐述了许多具体的细节以提供对本披露实施例的全面理解。然而明显地,一个或多个实施例在没有这些具体细节的情况下也可以被实施。
本实施例首先提供一种用于堆芯模型的特征线追踪方法,所述堆芯模型用于将堆芯的空间范围划分成分别具有多个子模型的多个层级,即,堆芯模型中存在有多个层级,每个层级中都有多个子模型。
在多个层级中,同一层级的至少一个子模型属于上一层级中的一个子模型,即,在一个层级中,一个子模型或多个子模型经过并运算后,在于上层级空间做交运算,将会得到上一层级的子模型。位于最下层级的每个子模型仅包括一种所述堆芯的材料,可以理解,通过上述运算,随着层级的不断上升,每个子模型内所包括的堆芯材料将会越来越多,从而最上层级的所有子模型经过运算后,能够对应堆芯的空间范围并且包括堆芯的所有材料。
下面对堆芯模型的构建方法进行具体的解释。
在反应堆物理中,建模主要是描述一个三维空间范围中每一位置的材料信息。空间范围可以表示为
Figure PCTCN2021098806-appb-000001
材料可以用标号形式写成集合materials={1,2,…,M}。那么堆芯模型中的一个子模型就可以表示为空间点集U到材料集合materials的映射关系,记为G=(U,f:U→materials)。
由于最下层级的子模型U中只有一种材料,即materials中只有一个象m,则该子模型可以写为(U,f:U→{m}),即:
f(r)=m,r∈U
这种子模型也可以称为图元,简写为(U,m),它是最简单的模型单位。
为了使用子模型构造更复杂的堆芯模型,需要对子模型进行逻辑操作:
如果若干个子模型G 1=(U 1,f 1)、G 2=(U 2,f 2)、...、G n=(U n,f n)满足
Figure PCTCN2021098806-appb-000002
即,其表示的空间范围两两互斥,那么联立f 1,f 2,...,f n就能构成一个新的映射f:
Figure PCTCN2021098806-appb-000003
所以
Figure PCTCN2021098806-appb-000004
构成一个新的几何模型,即,上一层级的子模型,记为:
Figure PCTCN2021098806-appb-000005
这种运算称为并运算,
Figure PCTCN2021098806-appb-000006
称为G 1,G 2,...,G n的并集,它相当将若干个下一层级的子模型装配到一起。
这种并运算具备交换律和结合律。
假设存在几何模型G=(U,f)和空间范围U′,构建映射f′:U′∩U→materials:
f′(r)=f(r),r∈U′∩U
则(U∩U′,f′)构成一个几何模型,记:
U′∩G=(U′∩U,f′)
称这种运算为交运算,U′∩G称为空间范围U′与几何体G的交集。交运算本质是用一个空间范围去切割图形,而真正发生作用的实际是U′∩U的空间范围。特别地,如果空间范围
Figure PCTCN2021098806-appb-000007
有:
Figure PCTCN2021098806-appb-000008
下面将给出使用上述方法进行堆芯模型构建的具体步骤。
具体实现时和蒙卡程序一样,采用Cell和Universe两种数据结构进行实现。其中Cell表示一个几何模型,需要用户给出一个空间范围U和材料索引m,即,需要用户设定仅包括一种材料的最下层级子模型:
Cell base=(U,m)
Universe中需要用户给出若干个几何模型的列表{Cell down,i},表示这些模型的并集,即,表示多个子模型经过的并运算:
Figure PCTCN2021098806-appb-000009
Cell的另一种格式需要用户给出一个Universe和一个空间范围U,表示U与Universe的交集:
Cell up=U∩Universe也就是说,Cell中一定有一个空间范围U,另一个可以填充某个材料索引构成图元Cell,也可以填充一个Universe构成一个组合体Cell,即,相对上方层级的子模型将表示更大的空间范围并且包含更多的材料。
图1示出了1个简单的堆芯模型构建示意图。图中平行四边形框表示材料,圆角框表示空间范围U,方框表示子模型,六边形框表示Universe;实线表示并集,虚线表示交集。最下层的子模型燃料、包壳、慢化剂中仅包括一种材料;这些最下层级的子模型生成并集栅元3,与栅元3的空间范围做交集,构成上一层级的子模型栅元3,即,构成了栅元层级的一个子模型。若干栅元层级的子模型再通过并和交运算构成组件2,即组件层级中的一个子模型,也就是该模型中的最上层级的子模型,所最上层级的子模型最终构成了整个堆芯模型,可以理解,在实际建模过程中,由于堆芯包括多种不同的结构,表示每种结构所需要的层级数目可以不同,但最下层级的子模型一定只包含 一种堆芯材料,最上层级的所有子模型经过运算一定构成了堆芯的空间范围并且包括了堆芯的所有材料。
在特征线追踪方法中,需要求解特征线被整个几何模型切割产生的各个线段的长度以及每段对应的Cell索引,根据本发明的特征线追踪方法包括:按层级由上到下依次判断每个所述层级中符合第一预设条件的所述子模型与所述特征线的交集是否符合第二预设条件;获得最下层级符合所述第一预设条件的所述子模型与所述特征线的符合所述第二预设条件的交集,根据获得的所述交集得到所述特征线与所述堆芯的材料的分段对应关系。
根据本发明的特征线追踪方法,利用特征线的直线性质,对于堆芯模型的遍历是只往下层进行的,也就是说,在整个特征线追踪中,避免了现有技术的逐段追踪算法中返回上层级节点进行遍历的步骤,从而节省了计算的次数,提高了效率。同时,现有技术中的逐段追踪算法,在进行实时追踪时需要对分段信息进行储存,本发明的特征线追踪方法即使进行实时追踪也仅需要增加一倍左右的计算时间,避免了进行了分段信息的储存,节省了大量储存空间。
进一步,最上层级对应的所述第一预设条件为:位于最上层级;非最上层级对应的所述第一预设条件为:与非最上层级的上一层级中与所述特征线的交集符合所述第二预设条件的所述子模型对应。本领域技术人员可以理解地,当判断最上层级时,判断的是最上层级的所有子模型,当判断非最上层级时,判断的是非最上层级中的部分子模型,该部分子模型是与非最上层级的上一层级的某些空间范围对应,该某些空间范围指的是非最上层级的上一层级中与特征线的交集符合第二预设条件的子模型。
即,在进行特征线追踪的时候,如图2所示,首先判断最上层级中的多个子模型与特征线的交集是否符合第二预设条件,对于符合第二预设条件的子模型,继续判断其对应的下一层级的子模型与特征线的交集是否符合第二预设条件,可以理解,对应的下一层级的子模型是指,该子模型是由这些下一层级的子模型经过并运算得到的,相当于图2的组件2与栅元1、2、3之间的关系。
经过多次这样的判断后,将会得到所有与特征线的交集符合第二预设条件的最下层级子模型,由于每个最下层级的子模型中仅包括一种材料,因此可以获得特征线与堆芯材料的分段对应关系。
进一步,第二预设条件为交集不是空集,即,在每一层级的判断过程中,将会排除与特征线没有交集的子模型,例如,在对图2中的组件层级进行计算时,如果子模型组件1与特征线的交集为空集,就不再对组件1在下一层级对应的子模型进行计算。
具体而言,在对例如图2的堆芯模型进行追踪的时候,首先计算最上层级中符合了第一预设条件的子模型,对于最上层级而言,所有的子模型都符合第一预设条件,特征线与子模型组件1、子模型组件2、子模型组件3…的交集,并判断交集是否符合第二预设条件,即,筛选出该层级中与特征线交集不是空集的子模型。进一步,完成组件这一层级的计算后,例如结果为子模型组件2符合第二预设条件,进行下一层级的计算,在该层级中,符合第一预设条件的子模型是指其对应的上一层级的子模型符合第二预设条件,即,与子模型组件2对应的子模型栅元1、栅元2等符合了第一预设条件,在这一层级中只需要对这些栅元进行判断。同样的,如果只有栅元3与特征线的交集不是空集,那么对子模型栅元3下一层的子模型燃料、反射层和包壳进行计算,由于燃料、反射层核包壳是最低层级的子模型,在计算完成后,可以得到特征线与材料的对应关系。
进一步,为了保证追踪的有序性,在一个层级的子模型判断完成后,可以根据预定规则对获得的不为空集的交集,以及这些交集的子模型进行排序。具体而言,这个预定规则可以根据这些交集中的元素在特征线上的位置来设定。
可以理解,在进行上述特征线与子模型的交集计算的时候,本领域技术人员可以使用多种方法来进行。
在一些实施方式中,使用了一种较为优选的计算特征线与子模型交集的方法,所述计算方法包括:构建辅助函数,所述辅助函数用于:计算对应的所述子模型在所述特征线上的多个边界点;计算每个所述边界点在所述辅助函数上的左极限;在所述左极限满足第三预设条件 时,判定所述边界点与相邻的与所述左极限对应的另一所述边界点之间的线段在所述子模型内。下面将说明该辅助函数的推导过程。
首先来说明对子模型进行描述的方法。
子模型通常较为复杂,采用构造实体几何法可以通过以下简单的基本单元之间的集合操作来描述复杂的空间范围。所谓基本单元是指一些能够通过数学公式来表达的基本几何体,例如:圆柱体、球体、立方体等,这些几何体可以用来描述堆芯中的相关结构,基本单元可以表示成{r|±f(r)>0}。在堆芯模型中,子模型使用的基本单元的方程一般小于等于二次,其统一形式可以写为:
f(r)=a 11x 2+a 22y 2+a 33z 2+a 12xy+a 23yz+a 13xz+a 1x+a 2y+a 3z+a 0
构造实体几何法将若干个基本单元通过正则交并补运算得到更复杂的空间范围。在本发明中,交并运算与集合论保持一致,即:
A∩B={r|r∈A且r∈B}
A∪B={r|r∈A或r∈B}
而对于补运算,使用空间范围闭包的补集,记为
Figure PCTCN2021098806-appb-000010
其中,cl为闭包(closure)的符号,采用上述形式是为了保证计算结果仍然是开区间,比如
Figure PCTCN2021098806-appb-000011
Figure PCTCN2021098806-appb-000012
更进一步,这里补运算满足:
Figure PCTCN2021098806-appb-000013
从形式上这一运算更类似于负运算。容易证明,补运算满足:
Figure PCTCN2021098806-appb-000014
Figure PCTCN2021098806-appb-000015
需要指出,这里的定义方式只是用于描述,不能直接反映点的性质。实际使用中还需要进行正则化处理,比如(x>1)∪(x<1)正则化处理为R3,即x=1的点也应当被认为是内点。
由于交并运算是二元运算,而补运算是一元运算,如图3所示,根据运算的先后顺序可以将计算过程画成一个树状图,图3中的树状图表示的运算为
Figure PCTCN2021098806-appb-000016
可以理解,图3中所有的叶子节点都是一个基本单元,其他的节点都是一种运算符。补运算有一个子树,而交并运算有两个子树,即,需要由两个基本单元来进行。
接下来说明辅助函数的构建方法。
假设特征线方程Track={r=Ωt+r 0|t∈(-∞,∞)},使用segment i标识特征线上的一个线段,则segment i={r=Ωt+r 0,i|t∈(0,l i)},可以简记为segment(r 0,i,Ω,l i)。其中,l为特征线线段的长度,r 0为特征线线段的起始点,Ω为特征线线段的方向向量
进一步的,引进Rvachev函数来对子模型进行进一步的描述,将基于构造实体几何描述的子模型描述成隐含数的不等式形式,具体而言,对于单个基本单元,可以表示为
R(r)=±f(r)
对于交并补运算,定义为:
Figure PCTCN2021098806-appb-000017
从而,对空间范围U,R(r)>0表示在U内,R(r)<0表示在
Figure PCTCN2021098806-appb-000018
内,R(r)=0表示在
Figure PCTCN2021098806-appb-000019
内,因此所有边界点必然满足 R(r)=0。但由于需要进行正则化处理,满足R(r)=0的点未必一定是边界点,也可能是孤立点或者聚点。
通过Rvachev函数,边界点必须满足的条件是:
R U(r 0+Ωt)=0,0<t<l
为保证求出t确实是边界点,还需要检测:
Figure PCTCN2021098806-appb-000020
该式中sgn为符号函数,即t的左右两端必须异号。这样必然一端在体内,一段在体外,从而构成边界点。为此,定义一个辅助函数:
Figure PCTCN2021098806-appb-000021
则在该辅助函数中,r U(t)=1的在空间范围内,r U(t)=0的在边界,r U(t)=-1的在空间范围外,即,令r U(t)=0,所得到的解集将是特征线与子模型的交界点的有序合集。因为所有令r U(t)=0点左右两端必然异号,所以(1)式中的辅助函数图像一定是锯齿状,如图4所示。
该辅助函数可以由一个有序零点数列和第一段值唯一确定:设零点为有序数列{t 1,...,t n},记t 0=0,t n+1=l。因为交点两端的符号不同,如果令
Figure PCTCN2021098806-appb-000022
即,边界点在该辅助函数上的左极限值,来表示线段(t 0,t 1)是否在体内,那么辅助函数可以写为:
Figure PCTCN2021098806-appb-000023
特征线上的线段与空间范围的交集segment∩U可以写为:
Figure PCTCN2021098806-appb-000024
本式中,对应的是位于该空间范围内的线段的有序集合,即,辅助函数值满足的第三预设条件为δ=1,在一些实施方式中,可以仅取δ=1部分,即,将根据该辅助函数计算特征线与对应的子模型的交集获得的解集写成特征线落在对应的所述子模型内的多条线段的集合。
上式也可以写成统一形式:
Figure PCTCN2021098806-appb-000025
其中,
Figure PCTCN2021098806-appb-000026
为向下取整。
该辅助函数的意义在于,完成了三维到一维的降维。空间范围的所有面都被转换为了点,体积都被转换为了线段,并且这些线段因为交替关系可以只用一个量表示。利用该辅助函数大大简化交并补操作的难度。
本领域技术人员可以使用多种方法来对该辅助函数中的δ和{t 1,...,t n}进行求解,在一些实施方式中,使用了一种较优的计算方法。
如果子模型仅有一个基本单元组成,在这种情况下可以直接等价于求解:
根据前述对基本单元的定义,这是一个一元二次方程,很容易进行求解并表示成δ和{t 1,...,t n}。
Figure PCTCN2021098806-appb-000027
对于经过了由基本单元经过交并运算得到的子模型,因为:
segment∩(U 1∩U 2)=(segment∩U 1)∩(segment∩U 2)
segment∩(U 1∪U 2)=(segment∩U 1)∪(segment∩U 2)
则利用上述(1)式,可得:
Figure PCTCN2021098806-appb-000028
Figure PCTCN2021098806-appb-000029
Figure PCTCN2021098806-appb-000030
所以,对于补运算,若r U(t)可表示为[{t i},δ],则有:
Figure PCTCN2021098806-appb-000031
对于交并运算,则需要进行判断,即,先利用上述辅助函数求解特征线与多个基本单元的交集,获得多个子解集,然后根据这些基本单元之间的交并运算关系,来对这些子解集进行相应的交或并运算,直到获得最终的解集。
根据前述对于子模型的构建方法,在一些情况下子模型可能是多个基本单元进行了多次运算得到的,在这些情况下,需要按照基本单元进行运算的顺序,来对相应的子解集进行运算,例如,参照图3中的计算树,要获得子模型Y对应的解集,需要先对A、B进行运算获得I对应的子解集,对C、D进行运算获得Y对应的子解集,依此类推,直到完成计算树中的所有计算获得子模型Y对应的解集。
在对两个子解集进行交并运算的时候,将
Figure PCTCN2021098806-appb-000032
表示为[{t 1,i} ,δ 1],
Figure PCTCN2021098806-appb-000033
表示为[{t 2,i},δ 2],则有:
Figure PCTCN2021098806-appb-000034
其中,(2)式中第一项的第三行是为了保证
Figure PCTCN2021098806-appb-000035
同理可知,
Figure PCTCN2021098806-appb-000036
的结果有:
Figure PCTCN2021098806-appb-000037
可知,在进行交、并运算时,需要判断一个子解集中的元素在另一个子解集中的函数值,而直接进行判断只能使用搜索算法,效率较低,在一些实施方式中提供了一种一次性完成所有点的判断方法,包括:设定第一指针和第二指针,分别指向所述两个子解集中位于数轴最左侧的元素;解集步骤:在位于数轴上相对右侧的第一指针或第二指针指向的元素在对应辅助函数的左极限满足第四预设条件时,将位于数轴上相对左侧的第一指针或第二指针指向的元素纳入交或并运算的解集,并将位于相对左侧的第一指针或第二指针向右侧移动指向下一个元素;当第一指针和第二指针指向的元素时,分别计算该元素在两个子解集对应的辅助函数的左极限,在所述左极限满足第五预设条件时,将该元素纳入交或并运算的解集,并将所述第一指针或所述第二指针分别向数轴右侧移动指向下一个所述元素;重复上述解集获取步骤,直到获得两个子解集进行交或并运算后的解集。
具体而言,该判断方法利用了{t 1,i}和{t 2,j}的有序性,如图5所示,设定第一指针P1和第二指针P2,初始时它们分别指向两个有序数列的第一个元素,若序列为空,则指向无穷大;用
Figure PCTCN2021098806-appb-000038
Figure PCTCN2021098806-appb-000039
表示指针P1和P2指向的元素在相应的辅助函数上的左极限,初始值即为δ 1和δ 2
由于初始时它们分别指向了两个子交集中的位于数轴最左侧的元素,在移动过程中每次只移动位于相对左侧的指针或同时移动,因此可以保证,如果P1在P2左侧,那么P2指向的元素一定是{t i,2}中 第一个比P1指向的元素大的数,即P1指向的元素一定位于P2之前的那一段,即
Figure PCTCN2021098806-appb-000040
反之,若P2在P1之前,则
Figure PCTCN2021098806-appb-000041
由此可以直接完成(2)、(3)式中前两项的判断,即,判断位于相对右侧的第一指针或第二指针指向的元素的在对应辅助函数上的左极限是否满足第四预设条件,由(2)、(3)式可知,在进行交运算时,第四预设条件为左极限为1,在进行并运算时,第四预设条件为左极限为-1,在判断完成后,将在前的指针向右侧移动一位,再次进行判断。
当指针P1和P2重合,即,P1和P2指向的元素相等的时候,则只需要判断
Figure PCTCN2021098806-appb-000042
Figure PCTCN2021098806-appb-000043
Figure PCTCN2021098806-appb-000044
时,将P1和P2指向的元素纳入解集,即,第五预设条件为左极限相等,在判断完成后,将P1和P2同时向右侧移动一位,再次进行判断。
上述的解集获取步骤,在P1和P2移动到无穷大后结束,即,完成所有元素的判断,利用这样的方法使得单次交并运算的时间复杂度被降低为O(n)。整个交并补算法中,补运算计算时间可以忽略,交并运算的总次数与基本单元的计算树有关,假设交并运算一共有b层,则最大计算次数就是满二叉树且无删去点时的计算次数O((b-1)·2 b-1)。
整个计算过程中,只有与基本单元的交集需要方程求解,后续过程都只是判断,效率较高,主要计算时间集中于与基本单元交集计算。而该计算代价很低,可以认为与单次E指数计算时间相当,也就是现有技术中长特征线法中一段的单次计算时间。
在利用上述的特征线与子模型的交集计算方法进行特征线追踪时,首先将特征线中的原点沿负方向移动到一个非常远的值r -∞,取一个非常大的长度l ,则也可近似将特征线写为segment(r -∞,Ω,l ),则根据上述堆芯模型的构建过程,必有:
Track∩U top=segment top
显然
Figure PCTCN2021098806-appb-000045
接下来考虑求解任意空间范围为U的层级Cell和一个在其范围内线段
Figure PCTCN2021098806-appb-000046
的交集segment∩Cell。
层级Cell为最下层级Cell(U,m),即,层级Cell中的每个子模型仅包括一种材料,则有:
segment∩Cell=(segment,m)      (4)
即,获得了特征线与材料的对应关系。
若层级Cell为非最下层级Cell(U,Universe),其中:
Figure PCTCN2021098806-appb-000047
则有:
Figure PCTCN2021098806-appb-000048
进一步,结合前述内容,特征线线段segment与该层级Cell中的一个符合第一预设条件的子模型Cell down,i的空间范围U down,i进行交集运算可求得:
Figure PCTCN2021098806-appb-000049
在上式中,J i为特征线线段segment被子模型的空间范围U down,i切割后形成的线段的数量。
因此:
Figure PCTCN2021098806-appb-000050
代入可以得到:
Figure PCTCN2021098806-appb-000051
其中,Cell为一个层级,I为该层级Cell中符合第一预设条件的子模型的数量。
由于为
Figure PCTCN2021098806-appb-000052
所以segment i,j∩Cell down,i也可以按照(4)或(5)继续往下写。这样上层的子模型就被逐渐向下分解,直至全部变成最下层级的子模型,根据(5)最终可以得到:
segment∩Cell=U(segment i,m i)
由此完成了整个追踪过程,在整个追踪过程中,主要计算时间集中在特征线的线段与子模型的计算上,该计算的次数等于特征线经过的所有子模型列表数量之和,如果整个计算模型可以用二叉树表示,那么计算次数可以估计为段数的2倍。如果认为平均每个最下层级子模型有4个基本单元,那么也就相当于段数×8次左右的E指数计算时间。
本实施例还提供一种堆芯中子物理计算方法,包括:
建模步骤:构建堆芯模型,所述堆芯模型用于将堆芯的空间范围划分成分别具有多个子模型的多个层级,其中,所述多个层级的同一层级中的至少一个所述子模型组成上一层级中的一个所述子模型,最下层级的每个所述子模型仅包括一种所述堆芯的材料,最上层级的所有所述子模型经过并运算组成所述堆芯的空间范围;
追踪步骤:定义特征线并根据上述任一所述的特征线追踪方法对所述特征线进行追踪;
迭代计算步骤:基于所述特征线的追踪结果进行迭代计算,完成所述堆芯的中子输送方程的求解。
本实施例还提供一种堆芯中子物理计算装置,包括:
建模模块,构件堆芯模型,所述堆芯模型用于将堆芯的空间范围划分成分别具有多个子模型的多个层级,其中,所述多个层级的同一层级中的至少一个所述子模型组成上一层级中的一个所述子模型,最下层级的每个所述子模型仅包括一种所述堆芯的材料,最上层级的所有所述子模型经过并运算组成所述堆芯的空间范围;
特征线追踪模块,定义特征线并根据上述任一所述的特征线追踪方法对所述特征线进行追踪;
迭代计算模块,基于所述特征线的追踪结果进行迭代计算,完成所述堆芯的中子输送方程的求解。
本发明的堆芯中子物理计算方法和装置,使用了上述任一所述的特征线追踪方法,从而节省了计算步骤,提高了计算的效率。
本实施例还提供一种反应堆模型,反应堆模型可以用于计算和描述反应堆中各物理量的状态,其中中子物理计算方法主要用于计算和描述中子学相关的中子通量密度、反应性和反应率分布等物理量。反应堆模型的构建一般包括建模和数值处理。建模即通过给定方式将反应堆模型输入给计算机,数值处理是将输入后的模型转换为计算所需的形式。本发明实施例的反应堆模型使用上述任一所述的堆芯中子物理计算方法来完成上述建模后的数值处理步骤,从而能够节省建模过程中的计算量,并提高计算效率。
本实施例还提供一种反应堆模拟装置。反应堆模拟装置可以用于对真实的反应堆运行情况进行模拟,这其中较为重要的是对堆芯的中子物理进行模拟和计算,因此,本实施提供了一种包括上述任一所述的堆芯中子物理计算装置的反应堆模拟装置,从而使用该反应堆模拟装置进行反应堆模拟时能够节省计算步骤,提高计算效率。
本实施例还提供一种计算机可读存储介质,其中,所述计算机可读存储介质存储有指令,所述指令用于实现上述的堆芯中子物理计算方法。
下面将结合几个具体的实施例来对上述实施方式中涉及到的技术方案进行进一步的说明。
首先给出一种求解特征线与基本单元的交集的计算方法,以及交并补运算的计算方法。
设直线方程为(x 0xt,y 0yt,z 0zt),根据上述基本单元的方程:
f(r)=a 11x 2+a 22y 2+a 33z 2+a 12xy+a 23yz+a 13xz+a 1x+a 2y+a 3z+a 0则可得:
At 2+Bt+C=0
Figure PCTCN2021098806-appb-000053
对上述方程进行求解时要进行如下判断:
●若A≠0
■若B 2-4AC≤0,因为单解也是可去点,所以
Figure PCTCN2021098806-appb-000054
δ与A同号;
■若B 2-4AC>0,存在两解t 1和t 2
◆t 1<t 2≤0<l,
Figure PCTCN2021098806-appb-000055
δ与A同号;
◆t 1≤0<t 2<l,{t i}={t 2},δ与A异号;
◆t 1≤0<l≤t 2
Figure PCTCN2021098806-appb-000056
δ与A异号;
◆0<t 1<t 2<l,{t i}={t 1,t 2},δ与A同号;
◆0<t 1<l≤t 2,{t i}={t 1},δ与A同号;
◆0<l≤t 1<t 2
Figure PCTCN2021098806-appb-000057
δ与A同号;
●若A=0,则必有B≠0,存在单解t
■若t≤0,
Figure PCTCN2021098806-appb-000058
δ与B同号;
■若0<t<l,{t i}={t},δ与B同号;
■若l≤t,
Figure PCTCN2021098806-appb-000059
δ与B异号;
其中,若
Figure PCTCN2021098806-appb-000060
δ表示(0,l)是否在体内;若不为空,则表示(0,t 1)是否在体内。
对于补运算,输入{t i|i=0,1,...,n}和δ,输出{t i|i=0,1,...,n}不变和-δ。
对于交并算法,给出了如下计算方法。
Figure PCTCN2021098806-appb-000061
Figure PCTCN2021098806-appb-000062
具体的实施例还给出了一种追踪方法。
对于各级的子模型Cell:
Figure PCTCN2021098806-appb-000063
对于子模型的并集Universe:
Figure PCTCN2021098806-appb-000064
具体的实施例还提供了根据本发明的追踪方法在基准题Takeda和C5G7进行了测试计算。
Takeda问题是一个两群三维简化堆芯模型,几何布置如图6所示。
图中x=0、y=0和z=0平面为反射边界条件,其余为真空边界条件;能群数为2。case1中红色部分填充材料为空,case2中红色部分填充材料为控制棒。计算时的参数选择为:网格尺寸取为0.5cm,射线密度取为0.05cm×0.05cm,求积组选择勒让德-切比雪夫求积组,8极角和16方位角。
计算结果如下表所示:
表1 Takeda问题计算结果
Figure PCTCN2021098806-appb-000065
表明了追踪结果的正确性。但这里更关心的是追踪效率:
表2 Takeda问题追踪效率
Figure PCTCN2021098806-appb-000066
其中时间比是指特征线追踪时间除以内迭代时间。在我们的测试中,如果Takeda问题在不进行加速,需要370次内迭代,此时特征线追踪时间仅需要占1.66%的时间。如果采用实时生成特征线的方法,那么只需要多付出原先61.58%的计算时间,就可以避免所有的分段信息存储。为了更好的说明追踪效果,提出了等效能群数,也就是时间比乘以能群数。因为几何追踪与能群数无关,即使能群变多,等效能群数也不发生改变。所以如果Takeda问题的能群数增加到8个,等效能群数仍然为1.23,此时时间比可以进一步减少到15.38%。
同样,在C5G7问题上进行了上述测试。为了提高追踪难度,这里采用了八分之一的处理,如图7所示,因为C5G7问题具备对称性,为此在对角线上一分为二。计算中采用的是C5G7-2D算例,通过补充轴向上的反射边界条件构成三维问题,共有四盒组件,每盒组件由17×17个栅元组成,每个栅元被划分成3环和8个扇区。
计算结果如下表所示,表示追踪正确:
表3 C5G7问题计算结果
Figure PCTCN2021098806-appb-000067
追踪效率如下:
表4 C5G7问题追踪效率
Figure PCTCN2021098806-appb-000068
相比于Takeda问题,追踪效率有所下降,主要因为结构复杂很多,且存在二次面。但总体而言,满足原先追踪时间与单次内迭代计算时间相当的估计,可以实现快速的特征线追踪,追踪效率较高,且支持以可接受的时间代价完成特征线实时计算。
本说明书的描述中,参考术语“一个实施方式”、“一些实施方式”、“示意性实施方式”、“示例”、“具体示例”、或“一些示例”等的描述意指结合所述实施方式或示例描述的具体特征、结构、材料或者特点包含于本发明的至少一个实施方式或示例中。在本说明书中,对上述术语的示意性表述不一定指的是相同的实施方式或示例。而且,描述的具体特征、结构、材料或者特点可以在任何的一个或多个实施方式或示例中以合适的方式结合。
此外,术语“第一”、“第二”仅用于描述目的,而不能理解为指示或暗示相对重要性或者隐含指明所指示的技术特征的数量。由此,限定有“第一”、“第二”的特征可以明示或者隐含地包括至少一个该特征。在本发明的描述中,“多个”的含义是至少两个,例如两个,三个等,除非另有明确具体的限定。
尽管上面已经示出和描述了本发明的实施方式,可以理解的是,上述实施方式是示例性的,不能理解为对本发明的限制,本领域的普通技术人员在本发明的范围内可以对上述实施方式进行变化、修改、替换和变型。

Claims (23)

  1. 一种用于堆芯模型的特征线追踪方法,其中,所述堆芯模型用于将堆芯的空间范围划分成分别具有多个子模型的多个层级,其中,所述多个层级的同一层级中的至少一个所述子模型属于上一层级中的一个所述子模型,最下层级的每个所述子模型仅包括一种所述堆芯的材料,各上层级的所有所述子模型经过并运算及与上层级空间范围的交运算构成上层级子模型,最上层级为所述堆芯模型;所述特征线追踪方法包括:
    按层级由上到下依次判断每个所述层级中符合第一预设条件的所述子模型与所述特征线线段的交集是否符合第二预设条件;
    获得最下层级符合所述第一预设条件的所述子模型与所述特征线的符合所述第二预设条件的交集,根据获得的所述交集得到所述特征线与所述堆芯的材料的分段对应关系。
  2. 根据权利要求1所述的特征线追踪方法,其中,
    最上层级对应的所述第一预设条件为:位于最上层级;
    非最上层级对应的所述第一预设条件为:与所述非最上层级的上一层级中与所述特征线线段的交集符合所述第二预设条件的所述子模型对应。
  3. 根据权利要求1所述的特征线追踪方法,其中,
    所述第二预设条件为不是空集。
  4. 根据权利要求1所述的特征线追踪方法,其中,还包括:
    在每个所述层级判断完成后,根据预定规则对符合所述第二预设条件的所述交集对应的所述子模型排序。
  5. 根据权利要求4所述的特征线追踪方法,其中,
    所述预定规则根据符合所述第二预设条件的所述交集在所述特征线上的位置设定。
  6. 根据权利要求1所述的特征线追踪方法,其中,判断每个所述层级中符合第一预设条件的所述子模型与所述特征线线段的交集是否符合第二预设条件前还包括计算所述特征线线段与对应的所述子模型的交集,所述计算所述特征线与对应的所述子模型的交集包括:
    构建辅助函数,所述辅助函数用于:
    计算对应的所述子模型在所述特征线线段上的多个边界点;
    计算每个所述边界点在所述辅助函数上的左极限;
    在所述左极限满足第三预设条件时,判定所述边界点与相邻的与所述左极限对应的另一所述边界点之间的线段在所述子模型内。
  7. 根据权利要求6所述的特征线追踪方法,其中,所述辅助函数为:
    Figure PCTCN2021098806-appb-100001
    其中,t为特征线线段上的点,当所述点t满足r U(t)=0,0<t<l时,判定所述点t为边界点,sgn为符号判断函数,l为特征线线段的长度,r 0为特征线线段的起始点,Ω为特征线线段的方向向量;
    当满足r U(t)=1时,判定所述点t与相邻的与所述左极限对应的另一所述边界点之间的线段在所述子模型内。
  8. 根据权利要求6或7所述的特征线追踪方法,其中,
    根据所述辅助函数计算所述特征线线段与对应的所述子模型的交集获得的解集包括所述特征线落在对应的所述子模型内的多条线段的有序集合。
  9. 根据权利要求8所述的特征线追踪方法,其中,所述解集的计算方法包括:
    当对应的所述子模型包括经过逻辑运算的多个基本单元时,分别使用所述辅助函数求解所述特征线线段与所述多个基本单元的交集,获得多个子解集,所述基本单元为基本几何体的数学模型;
    根据所述多个基本单元之间的交、并或补运算关系,分别对所述多个子解集进行相应的交、并或补运算,直到获得所述解集。
  10. 根据权利要求9所述的特征线追踪方法,其中,对于每个基本单元U,辅助函数r U(t)表示为[{t i},δ],其中,t i为边界点,δ为边界点的左极限。
  11. 根据权利要求9所述的特征线追踪方法,其中,所述分别对多个所述子解集进行相应交、并或补运算包括:
    每次从所述多个子解集中选择两个所述子解集进行相应的交或并运算,直到获得最终的解集。
  12. 根据权利要求11所述的特征线追踪方法,其中,将一个子解集
    Figure PCTCN2021098806-appb-100002
    表示为[{t 1,i},δ 1],另一个子解集
    Figure PCTCN2021098806-appb-100003
    表示为[{t 2,i},δ 2],则进行交运算时
    Figure PCTCN2021098806-appb-100004
    的结果为:
    Figure PCTCN2021098806-appb-100005
  13. 根据权利要求11所述的特征线追踪方法,其中,将一个子解集
    Figure PCTCN2021098806-appb-100006
    表示为[{t 1,i},δ 1],另一个子解集
    Figure PCTCN2021098806-appb-100007
    表示为[{t 2,i},δ 2],则进行交运算时
    Figure PCTCN2021098806-appb-100008
    的结果为:
    Figure PCTCN2021098806-appb-100009
  14. 根据权利要求9-13中任一项所述的特征线追踪方法,其中,所述分别对多个所述子解集进行相应交、并或补运算包括还包括:
    每次从所述多个子解集中选择一个所述子解集进行补运算。
  15. 根据权利要求14所述的特征线追踪方法,其中,将子解集r U(t)表示为[{t i},δ],则在进行补运算时
    Figure PCTCN2021098806-appb-100010
    的结果为:
    Figure PCTCN2021098806-appb-100011
  16. 根据权利要求15所述的特征线追踪方法,其中,对所述两个所述子解集进行相应的交或并运算包括:
    设定第一指针和第二指针,初始时分别指向所述两个子解集中位于数轴的最左侧的元素;
    解集获取步骤:在位于数轴相对右侧的所述第一指针或所述第二指针指向的所述元素的在对应的所述辅助函数上的左极限满足第四预设条件时,将位于数轴相对左侧的所述第一指针或所述第二指针指向的所述元素纳入交或并运算的解集;并将所述位于数轴相对左侧的所述第一指针或所述第二指针向数轴右侧移动指向下一个所述元素;
    当所述第一指针和所述第二指针指向的所述元素相等时,分别判断所述元素在所述两个子解集对应的所述辅助函数的左极限,在所述 左极限满足第五预设条件时,将所述元素纳入交或并运算的解集,并将所述第一指针或所述第二指针分别向数轴右侧移动指向下一个所述元素;
    重复所述解集获取步骤,直到获得所述两个子解集进行交或并运算后的解集。
  17. 根据权利要求1-5中任一项所述的特征线追踪方法,其中,所述子模型与所述特征线线段的交集表示为:
    Figure PCTCN2021098806-appb-100012
    其中,segment为特征线上的线段,J i为特征线线段segment被子模型的空间范围U down,i切割后形成的线段的数量。
  18. 根据权利要求17所述的特征线追踪方法,其中,每个所述层级中所有符合第一预设条件的所述子模型与所述特征线线段的交集写为:
    Figure PCTCN2021098806-appb-100013
    其中,Cell为一个所述层级,I为Cell中符合第一预设条件的所述子模型的数量。
  19. 一种堆芯中子物理计算方法,包括:
    建模步骤:构建堆芯模型,所述堆芯模型用于将堆芯的空间范围划分成分别具有多个子模型的多个层级,其中,所述多个层级的同一层级中的至少一个所述子模型属于上一层级中的一个所述子模型,最下层级的每个所述子模型仅包括一种所述堆芯的材料,各上层级的所有所述子模型经过并运算及与上层级空间范围的交运算构成上层级子模型,最上层级为所述堆芯模型;
    追踪步骤:定义特征线并根据权利要求1至18中任意一项所述的特征线追踪方法对所述特征线进行追踪;
    迭代计算步骤:基于所述特征线的追踪结果进行迭代计算,完成所述堆芯的中子输送方程的求解。
  20. 一种反应堆模型,所述反应堆模型使用如权利要求19所述的堆芯中子物理计算方法构建。
  21. 一种堆芯中子物理计算装置,包括:
    建模模块,用于构建堆芯模型,所述堆芯模型用于将堆芯的空间范围划分成分别具有多个子模型的多个层级,其中,所述多个层级的同一层级中的至少一个所述子模型属于上一层级中的一个所述子模型,最下层级的每个所述子模型仅包括一种所述堆芯的材料,各上层级的所有所述子模型经过并运算及与上层级空间范围的交运算构成上层级子模型,最上层级为所述堆芯模型;
    特征线追踪模块,用于定义特征线并根据权利要求1至10中任意一项所述的特征线追踪方法对所述特征线进行追踪;
    迭代计算模块,基于所述特征线的追踪结果进行迭代计算,完成所述堆芯的中子输送方程的求解。
  22. 一种反应堆模拟装置,包括:
    如权利要求21所述的堆芯中子物理计算装置。
  23. 一种计算机可读存储介质,其中,
    所述计算机可读存储介质存储有指令,所述指令用于实现根据权利要求19所述的堆芯中子物理计算方法。
PCT/CN2021/098806 2020-06-09 2021-06-08 特征线追踪方法、堆芯中子物理计算方法和装置 WO2021249374A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
EP21755885.7A EP3951634A4 (en) 2020-06-09 2021-06-08 Feature line tracing method, and core neutron physics calculation method and device
US17/435,273 US20220336116A1 (en) 2020-06-09 2021-06-08 Method of tracking characteristic line, and method and apparatus of calculating core neutron physics

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010519286.XA CN111666689B (zh) 2020-06-09 2020-06-09 特征线追踪方法、堆芯中子物理计算方法和装置
CN202010519286.X 2020-06-09

Publications (1)

Publication Number Publication Date
WO2021249374A1 true WO2021249374A1 (zh) 2021-12-16

Family

ID=72386178

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/098806 WO2021249374A1 (zh) 2020-06-09 2021-06-08 特征线追踪方法、堆芯中子物理计算方法和装置

Country Status (4)

Country Link
US (1) US20220336116A1 (zh)
EP (1) EP3951634A4 (zh)
CN (1) CN111666689B (zh)
WO (1) WO2021249374A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111666689B (zh) * 2020-06-09 2023-09-29 中国原子能科学研究院 特征线追踪方法、堆芯中子物理计算方法和装置

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150142411A1 (en) * 2013-11-21 2015-05-21 Terrapower, Llc Method and System for Generating a Nuclear Reactor Core Loading Distribution
CN106528916A (zh) * 2016-09-22 2017-03-22 中国科学院合肥物质科学研究院 一种应用于聚变反应堆核分析的自适应子空间迭代分割的协同系统
CN108734730A (zh) * 2018-04-27 2018-11-02 西安交通大学 一种基于模块化特征线求解任意csg几何体体积的方法
CN110427707A (zh) * 2019-08-06 2019-11-08 中国原子能科学研究院 适应于多堆型几何及特征线法的构造实体几何建模方法
CN111666689A (zh) * 2020-06-09 2020-09-15 中国原子能科学研究院 特征线追踪方法、堆芯中子物理计算方法和装置

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9424376B2 (en) * 2011-11-18 2016-08-23 Terrapower, Llc Enhanced neutronics systems
CN106528942B (zh) * 2016-10-13 2019-07-16 哈尔滨工程大学 一种基于中子输运特征线理论的数据可视化方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150142411A1 (en) * 2013-11-21 2015-05-21 Terrapower, Llc Method and System for Generating a Nuclear Reactor Core Loading Distribution
CN106528916A (zh) * 2016-09-22 2017-03-22 中国科学院合肥物质科学研究院 一种应用于聚变反应堆核分析的自适应子空间迭代分割的协同系统
CN108734730A (zh) * 2018-04-27 2018-11-02 西安交通大学 一种基于模块化特征线求解任意csg几何体体积的方法
CN110427707A (zh) * 2019-08-06 2019-11-08 中国原子能科学研究院 适应于多堆型几何及特征线法的构造实体几何建模方法
CN111666689A (zh) * 2020-06-09 2020-09-15 中国原子能科学研究院 特征线追踪方法、堆芯中子物理计算方法和装置

Also Published As

Publication number Publication date
US20220336116A1 (en) 2022-10-20
CN111666689B (zh) 2023-09-29
CN111666689A (zh) 2020-09-15
EP3951634A4 (en) 2023-06-28
EP3951634A1 (en) 2022-02-09

Similar Documents

Publication Publication Date Title
CN111708531B (zh) 数据处理方法和装置
Uy et al. Joint learning of 3d shape retrieval and deformation
Anderson et al. Smooth, volume-accurate material interface reconstruction
CN103049555B (zh) 保证语义正确性的动态分层集成数据访问方法
Guo et al. A hybrid 3D feature recognition method based on rule and graph
CN110427707A (zh) 适应于多堆型几何及特征线法的构造实体几何建模方法
WO2021249374A1 (zh) 特征线追踪方法、堆芯中子物理计算方法和装置
Lee et al. Partial entity structure: a compact boundary representation for non-manifold geometric modeling
CN114048204A (zh) 基于数据库倒排索引的北斗网格空间索引方法和装置
Cao et al. A knowledge-based prototype system to support product conceptual design
US11755554B2 (en) Chart engine
US11017027B2 (en) Index machine
CN111709601B (zh) 数据处理方法和装置
Marinić-Kragić et al. Superimposed RBF and B-spline parametric surface for reverse engineering applications
Ding et al. 3D CAD Model Representation and Retrieval based on Hierarchical Graph.
Shang et al. Effective re-parameterization and GA based knot structure optimization for high quality T-spline surface fitting
Andersen et al. Building an ontology of CAD model information
Zhuang et al. Integration of subdivision method into boundary element analysis
Li et al. Towards sparse matrix operations: graph database approach for power grid computation
WO2024108580A1 (zh) 多维参数化城市信息模型构建方法、系统及计算机设备
Sun A robust simplex cut-cell method for adaptive high-order discretizations of aerodynamics and multi-physics problems
CN117494509B (zh) 一种基于多物理场的块结构自适应网格细化方法
Lu et al. Visualization of reactor core based on triangular mesh method
CN115017372A (zh) 一种支持体参数化模型表达的翼面数据结构构建方法
Li et al. FreeCAD based Monte Carlo modeling approach for fusion reactor facilities

Legal Events

Date Code Title Description
ENP Entry into the national phase

Ref document number: 2021755885

Country of ref document: EP

Effective date: 20210827

121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21755885

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE