CN115495863B - Heat dissipation topology optimization method for smooth boundary expression - Google Patents

Heat dissipation topology optimization method for smooth boundary expression Download PDF

Info

Publication number
CN115495863B
CN115495863B CN202211462679.7A CN202211462679A CN115495863B CN 115495863 B CN115495863 B CN 115495863B CN 202211462679 A CN202211462679 A CN 202211462679A CN 115495863 B CN115495863 B CN 115495863B
Authority
CN
China
Prior art keywords
node
unit
sensitivity
boundary
value
Prior art date
Legal status (The legal status 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 status listed.)
Active
Application number
CN202211462679.7A
Other languages
Chinese (zh)
Other versions
CN115495863A (en
Inventor
王琥
尹纪超
雷钧
蔡勇
王文伟
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenzhen Automotive Research Institute of Beijing University of Technology
Original Assignee
Shenzhen Automotive Research Institute of Beijing University of Technology
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 Shenzhen Automotive Research Institute of Beijing University of Technology filed Critical Shenzhen Automotive Research Institute of Beijing University of Technology
Priority to CN202211462679.7A priority Critical patent/CN115495863B/en
Publication of CN115495863A publication Critical patent/CN115495863A/en
Application granted granted Critical
Publication of CN115495863B publication Critical patent/CN115495863B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00Computer-aided design [CAD]
    • G06F30/10Geometric CAD
    • G06F30/18Network design, e.g. design based on topological or interconnect aspects of utility systems, piping, heating ventilation air conditioning [HVAC] or cabling
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00Computer-aided design [CAD]
    • G06F30/20Design optimisation, verification or simulation
    • G06F30/23Design optimisation, verification or simulation using finite element methods [FEM] or finite difference methods [FDM]

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Geometry (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Computer Hardware Design (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Computational Mathematics (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Pure & Applied Mathematics (AREA)
  • Complex Calculations (AREA)

Abstract

A heat dissipation topology optimization method of smooth boundary expression relates to the technical field of structure optimization. Dispersing a heat dissipation topological structure to be optimized into a finite element model, and performing optimization based on a finite element grid to obtain the optimal distribution of materials; converting the discrete optimization into continuous optimization based on an interpolation function, and analyzing by using a heat transfer coefficient obtained by the first interpolation function to obtain unit sensitivity; and realizing projection conversion from the unit sensitivity to the node sensitivity according to the preset corresponding relation between the unit and the node. And determining a key level value according to the node sensitivity, wherein when the node sensitivity part of the node corresponding to the unit is greater than the key level value, the unit is a boundary unit. And performing grid subdivision on the boundary unit, calculating the sensitivity of a new subdivision node by using a second interpolation function, determining the density value of the boundary unit according to the sensitivity of the new subdivision node, determining the volume fraction according to the density values of all units, updating the key level value by using the volume fraction, and finally finding out the key level value meeting the volume constraint.

Description

Heat dissipation topology optimization method for smooth boundary expression
Technical Field
The invention relates to the technical field of structural optimization, in particular to a heat dissipation topological optimization method for smooth boundary expression.
Background
The topological optimization as a structural optimization design method can fully excavate the potential of materials, save the material to the maximum extent on the premise of meeting design requirements and realize the lightweight design of the structure. The topological optimization design is different from the traditional design iteration 'design-realization' empirical design mode, and the conceptual design structure can be directly obtained, so that the product design period is shortened, the design cost is saved, and the method has very important application value for industrial design. Heat dissipation components are one of the important fields of application for topology optimization, and the topology optimization design can realize optimization of heat dissipation performance of components under given material properties and related constraints, including but not limited to minimizing thermal flexibility, minimizing maximum temperature, minimizing temperature field variance, and the like. It is clearly more attractive to meet the heat dissipation requirements by optimizing the structure, as opposed to using expensive heat dissipation materials.
The conventional heat transfer topology optimization method performs optimization and topology representation based on cells of finite element mesh division, as shown in fig. 1, the boundaries of the optimized structure are represented in a zigzag form. In addition, most of heat transfer topological optimization results are distributed in a tree shape, the small-size characteristics are complex, and gray level units are enriched, so that the manufacturing difficulty of the optimized structure of the tree distribution in the engineering practice is very high.
Disclosure of Invention
In order to solve the above problems, the present application provides a heat dissipation topology optimization method expressed by a smooth boundary.
According to a first aspect, an embodiment provides a method for optimizing a heat dissipation topology by smoothing boundary expression, including:
dispersing a heat dissipation topological structure to be optimized into a finite element model;
calculating a density field of the heat dissipation topological structure to be optimized based on a finite element model, wherein the density field comprises a plurality of units;
calculating a heat transfer coefficient by using a first interpolation function, updating the finite element analysis model according to the heat transfer coefficient, and determining the unit sensitivity of each unit according to the updated finite element analysis model; the finite element analysis model is used for solving the finite element model;
according to a preset mapping relation of a plurality of units corresponding to a single node, acquiring unit sensitivity of each unit of the plurality of units to determine node sensitivity of the corresponding single node;
determining a key level value according to the node sensitivity of each node, and when the node sensitivity of a node corresponding to a unit in the heat dissipation topological structure to be optimized is all larger than the key level value, the unit is a complete entity unit; when the node sensitivity of a node corresponding to one unit in the heat dissipation topological structure to be optimized is larger than the key level value, the unit is a boundary unit; when the node sensitivity of a node corresponding to one unit in the heat dissipation topological structure to be optimized is all smaller than the key level value, the unit is a cavity unit;
dividing each boundary unit to obtain a corresponding new division unit, and calculating the sensitivity of a new division node of the new division unit corresponding to each boundary unit by using a second interpolation function; for each boundary unit, the density value of the boundary unit is respectively determined according to the sensitivity of the newly-divided node of the boundary unit;
and determining volume fractions according to the density values of all the units, and updating the key level values according to the volume fractions so as to perform smooth boundary expression meeting volume constraints on the result to be optimized.
In an embodiment, the calculating the density field of the heat dissipation topology to be optimized based on the finite element model includes:
controlling the filtering radius of a finite element model based on a continuous density strategy, carrying out a plurality of iterations on the density field to obtain an iterative optimization target, and when the variance of the optimization target is smaller than a set tolerance, ending the iteration and outputting the density field; the filtering radius is calculated by the following formula:
Figure 729641DEST_PATH_IMAGE001
wherein r is the iteration radius, k is the number of iterations, r min To prepareSet minimum filtration radius r 0 And for a preset initial filtering radius, lp is the iteration number of the set initial filtering radius attenuated to the minimum filtering radius.
In one embodiment, said calculating a heat transfer coefficient using a first interpolation function, and updating said finite element analysis model based on said heat transfer coefficient, comprises:
calculating a heat transfer coefficient in the first interpolation function using a penalty factor greater than a set value, the first interpolation function calculated using the formula:
Figure DEST_PATH_IMAGE002
wherein k is 0 For a heat transfer coefficient, ε is 10 -3 Rho is unit density and the value range is [0,1 ]]P is a penalty factor, and k is an interpolation coefficient.
In an embodiment, the obtaining unit sensitivity of each unit of the plurality of units according to a preset mapping relationship between the plurality of units and a single node to determine node sensitivity of the corresponding single node includes:
calculating the node sensitivity of the single node by using the following mapping formula:
Figure 302574DEST_PATH_IMAGE003
wherein ns m Node acuity for node m; ne is the total number of units; i is the ith unit and has a value range of [1, ne ]];es i Cell sensitivity for the ith cell; w is a i,m Is a weighted function of the sensitivity of the cells in the set of cells.
In one embodiment, the weighting function of the cell sensitivity field of the number of cells is determined by the formula:
Figure 144628DEST_PATH_IMAGE005
wherein, w i,m A weighting function, r, of the cell sensitivity field of the number of cells dist For the mapping range of the node m,
Figure DEST_PATH_IMAGE006
is the distance from the ith cell to node m.
In an embodiment, the determining a key level value according to the node sensitivity of each node in the heat dissipation topology to be optimized includes:
acquiring a lower bound value of an interval of the minimum node sensitivity value in all node sensitivities and an upper bound value of an interval of the maximum node sensitivity value in all node sensitivities; and updating the lower bound value and the upper bound value by using a dichotomy, and calculating the average value of the updated lower bound value and the updated upper bound value to determine the key horizontal value.
In an embodiment, the dividing each boundary unit to obtain a corresponding new divided unit, and calculating the sensitivity of a new divided node of the new divided unit corresponding to each boundary unit by using a second interpolation function includes:
interpolating the node sensitivity of each boundary unit by using a second interpolation function to obtain the new subdivision node sensitivity corresponding to the new subdivision unit; the second interpolation function is expressed by the following formula:
Figure DEST_PATH_IMAGE008
wherein ns new (o) sensitivity of new subdivision node for new subdivision unit (xi) i ,η i ) Is the coordinate of the node corresponding to the boundary unit, i is the ith node corresponding to the boundary unit, ce is the total number of the nodes corresponding to the boundary unit, (xi) o ,η o ) For the coordinates of the new partitioning node corresponding to the new partitioning unit,
Figure 461209DEST_PATH_IMAGE009
the node acuity corresponding to the border cell.
In one embodiment, for each boundary cell, the determining, by the boundary cell, a density value of the boundary cell according to the sensitivity of the newly-divided node includes:
calculating the percentage of the nodes corresponding to the new subdivision units with the sensitivity of the new subdivision nodes higher than the key level value occupying the nodes of the corresponding boundary units, and determining the density values of the corresponding boundary units according to the percentage; the percentage is calculated using the following formula:
Figure DEST_PATH_IMAGE010
wherein b is the percentage of the corresponding node of the new subdivision unit occupying the node of the corresponding boundary unit, (ξ) o ,η o ) Coordinates of the nodes corresponding to the new subdivision unit, ns m Sensitivity, ns, of new subdivision for new subdivision unit key Is the key level value, N node Is the total number of nodes of the border cell.
In an embodiment, the updating the key level value according to the volume fraction to optimize the heat dissipation topology to be optimized includes:
updating the key level value according to the volume fraction, and when the volume fraction is larger than the preset volume fraction, updating the current lower bound value to be
Figure 446220DEST_PATH_IMAGE011
(ii) a When the volume fraction is smaller than the preset volume fraction, updating the upper bound value of the current time to be
Figure DEST_PATH_IMAGE012
(ii) a Wherein ns key As the key level value, B min For the last updated lower bound value, B max The upper bound value is updated last time; and updating the key level value according to the current lower bound value and the current upper bound value.
According to a second aspect, an embodiment provides a computer-readable storage medium having a program stored thereon, the program being executable by a processor to implement the above-mentioned method.
According to the heat dissipation topological optimization method expressed by the smooth boundary and the computer readable storage medium of the embodiment, the heat dissipation topological structure to be optimized is dispersed into a finite element model, and optimization is performed based on a finite element grid to obtain the optimal distribution of materials; converting the discrete optimization into continuous optimization based on an interpolation function, and analyzing by using a heat transfer coefficient obtained by the first interpolation function to obtain unit sensitivity; and realizing projection conversion from the unit sensitivity to the node sensitivity according to the preset corresponding relation between the unit and the node. And determining a key level value according to the node sensitivity, wherein when the node sensitivity part of the node corresponding to the unit is greater than the key level value, the unit is a boundary unit. And performing grid subdivision on the boundary unit, calculating the sensitivity of a new subdivision node by using a second interpolation function, determining the density value of the boundary unit according to the sensitivity of the new subdivision node, determining the volume fraction according to the density values of all units, updating the key level value by using the volume fraction, and finally finding out the key level value meeting the volume constraint. According to the traditional heat transfer topological optimization framework based on units, part of small-size features are eliminated, the boundary effect of an optimization result is weakened, and then the conversion from a jagged boundary to a smooth boundary expression is realized by utilizing boundary smoothing processing. The smooth structure topology deletes part of small-size detail features, important detail features are reserved, the performance requirements of the heat dissipation component are met, and meanwhile the manufacturability of an optimization result is improved.
Drawings
FIG. 1 is a diagram illustrating a result of topology boundary optimization in the background art;
FIG. 2 is a flowchart of an embodiment of a method for optimizing a heat dissipation topology using smooth boundary representation;
FIG. 3 is a graph of the optimization results after eliminating some small-sized detail features using a continuity density filtering strategy according to an embodiment;
FIG. 4 is a schematic diagram of a node sensitivity calculation according to an embodiment;
FIG. 5 is a schematic diagram illustrating the subdivision of a border cell according to one embodiment;
FIG. 6 is a diagram of an embodiment of an optimization result after subsequent processing of a density field of an iterative structure;
fig. 7 is a schematic diagram of an embodiment of a heat dissipation topology optimization apparatus with smooth boundary representation.
Detailed Description
The present invention will be described in further detail with reference to the following detailed description and accompanying drawings. Wherein like elements in different embodiments are numbered with like associated elements. In the following description, numerous details are set forth in order to provide a better understanding of the present application. However, one skilled in the art will readily recognize that some of the features may be omitted or replaced with other elements, materials, methods in different instances. In some instances, certain operations related to the present application have not been shown or described in detail in order to avoid obscuring the core of the present application from excessive description, and it is not necessary for those skilled in the art to describe these operations in detail, so that they may be fully understood from the description in the specification and the general knowledge in the art.
Furthermore, the features, operations, or characteristics described in the specification may be combined in any suitable manner to form various embodiments. Also, the various steps or actions in the method descriptions may be transposed or transposed in order, as will be apparent to one of ordinary skill in the art. Thus, the various sequences in the specification and drawings are for the purpose of describing certain embodiments only and are not intended to imply a required sequence unless otherwise indicated where such sequence must be followed.
The ordinal numbers used herein for the components, such as "first," "second," etc., are used merely to distinguish between the objects described, and do not have any sequential or technical meaning. The term "connected" and "coupled" when used in this application, unless otherwise indicated, includes both direct and indirect connections (couplings).
A level set method, which is one of common topology optimization methods, can directly implement smooth expression of a boundary, and the level set method uses a level set function with one dimension higher to describe a design space, and a response value of the level set function represents an attribute of the point: a solid, a void, or a boundary. Usually, the intersection of the zero horizontal plane and the level set function is defined as a structural outline, the set above the zero horizontal plane represents a solid material distribution region Ω, otherwise, the set represents a void distribution region D/Ω, and the mathematical model expression is as follows:
Figure 878469DEST_PATH_IMAGE013
wherein,
Figure DEST_PATH_IMAGE014
in the form of a function of the level set,
Figure 327860DEST_PATH_IMAGE015
representing the structural boundaries. Weak materials are also used to replace the material properties of the void region to avoid numerical singularities. In order to realize the flexibility minimization by evolving the structure contour, the normal speed of a level set function is determined by using sensitivity analysis, pseudo time is introduced to obtain a Hamiltonian-Jacobian equation, and the updated level set function is obtained by solving. However, solving the hamiltonian-jacquette equation results in significantly increased computational cost and memory burden, and thus the level set method is difficult to be generalized to a large-scale problem.
The moving deformable component method also enables smooth representation of the boundary and has few design variables to avoid additional computational cost. However, the motion variability component is sensitive to parameters, and tends to converge to a sub-optimal solution or even an infeasible solution, and therefore the computational cost of stealth due to iterative parameter tuning can also be expensive.
The results of heat transfer topology optimization are often expressed in terms of dendrites, and therefore the detailed features of the structure are complex. The level set method and the mobile deformable component method are inferior to the expression capability of detailed features, but the optimization result obtained by the classical topology optimization method based on density is too complex in detail, and is not beneficial to actual manufacturing.
The method is based on a classical SIMP (Solid Isotropic Material with Penalification model, which is a common density-stiffness interpolation model) interpolation model topology optimization framework, only uses a continuous density filtering strategy to control the radius size in the optimization process, and executes subsequent processing on a density field of an iterative structure. The continuous density filtering strategy is firstly used for eliminating the detail features with the small size in the optimization process, and then the subsequent processing is carried out on the density field of the iterative structure to further eliminate the detail features with the small size, however, the necessary feature structures are reserved and the connectivity of the structure is not changed. By using the method, the radius size is controlled only by using a continuous density filtering strategy in the optimization process, and the post-processing is performed on the density field after the iteration is finished, so that the high calculation cost and the huge memory burden are avoided.
Referring to fig. 2, an embodiment of the present application provides a method for optimizing a heat dissipation topology with smooth boundary expression, which includes the following steps.
Step S100: and dispersing the heat dissipation topological structure to be optimized into a finite element model.
In some embodiments, a topological optimization design space is established in consideration of actual needs of the heat dissipation component, a topological structure to be optimized is discretized to generate a finite element model, and an optimization target, boundary conditions, convergence conditions and initial values of relevant parameters are set.
Step S200: and calculating a density field of the heat dissipation topological structure to be optimized based on the finite element model.
In some embodiments, the filtering radius used at each iteration is controlled based on a continuity density strategy to control the bandwidth of the grayscale unit at the structure boundary. Wherein the filtering radius used by the iterative loop is obtained according to the following formula:
Figure 339810DEST_PATH_IMAGE001
wherein r is the iteration radius, k is the number of iterations, r min Is a predetermined minimum filter radius, r 0 And for a preset initial filtering radius, lp is the iteration number of the set initial filtering radius attenuated to the minimum filtering radius. Some embodimentsThe preset minimum filtering radius and the preset initial filtering radius are set with respect to the initial values of the parameters in step S100.
In some embodiments, the finite element model is used to calculate the temperature field of the heat dissipation topological structure to be optimized, the element sensitivity is calculated based on the temperature field of the heat dissipation topological structure to be optimized, and the density field is updated by using the calculated element sensitivity. When the density field is updated each time, the current filtering radius needs to be updated, a new density filtering mapping matrix is established until the optimization iteration meets the given convergence condition, namely the variance of the optimization target is smaller than the set tolerance, the optimization cycle is ended, and the final density field is output. Wherein, the density field comprises a plurality of units.
Referring to fig. 3, steps S100 to S200 are graphs of the optimization results after eliminating some small-sized detail features by using a continuous density filtering strategy in the optimization process.
Step S300: cell acuity is determined and node acuity for individual nodes is determined based on the cell acuity.
In some embodiments, a heat transfer coefficient is calculated using the first interpolation function, the finite element analysis model is updated based on the heat transfer coefficient, and the element sensitivity of each element is determined based on the updated finite element analysis model. And acquiring the unit sensitivity of each unit of the plurality of units according to the preset mapping relation of the plurality of units corresponding to the single node so as to determine the node sensitivity of the corresponding single node.
In some embodiments, a penalty factor greater than a set value is used in the first interpolation function to calculate the heat transfer coefficient, so that the penalty effect on the intermediate density is increased, and the islanding phenomenon of the result after the smoothing treatment is avoided. That is, in calculating the cell sensitivities of individual cells throughout the density field, a large penalty factor needs to be applied to calculate the heat transfer coefficient to prevent the presence of isolated cells. Wherein the first interpolation function is represented by the following formula:
Figure 805426DEST_PATH_IMAGE002
wherein k is 0 For a heat transfer coefficient, ε is 10 -3 Rho is unit density and the value range is [0,1 ]]P is a penalty factor, and k is an interpolation coefficient.
In some embodiments, after the unit sensitivity of each unit is calculated, each unit corresponds to a node according to a preset mapping relationship that a plurality of units correspond to one node, and some units correspond to the same node. A cell sensitivity field is acquired for each cell to determine a node sensitivity for a corresponding node of each cell. Referring to fig. 4, a plurality of cells covered in a circle all correspond to a node corresponding to the center of the circle, and then cell sensitivity fields of all the cells covered in the circle are acquired, and the node sensitivity of the node corresponding to the center of the circle is determined according to the cell sensitivity fields of the cells. In some embodiments, it is also possible that a single unit can map out a single node. Wherein the node sensitivity of a single node is calculated by using the following mapping formula:
Figure 110375DEST_PATH_IMAGE003
wherein ns m Node acuity for node m; ne is the total number of units; i is the ith unit and has a value range of [1,Ne];es i Cell sensitivity for the ith cell; w is a i,m A weighting function for the sensitivity of a cell in the set of cells.
In some embodiments, the weighting function for the cell sensitivity field for a number of cells is determined by the following equation:
Figure 723759DEST_PATH_IMAGE017
wherein, w i,m Is a weighted function of the cell sensitivity field of the plurality of cells, r dist For the mapping range of the node m,
Figure DEST_PATH_IMAGE018
is the distance from the ith cell to node m.
In some embodiments, a finite element analysis model, which is a general process of conventional topology optimization, is used to solve the finite element model. The method specifically comprises the following steps: based on the density field determined in step S200, the heat transfer coefficients of the units in the density field are calculated by using the first interpolation function, and after the heat transfer coefficients of the units are calculated, the structural response analysis is performed on the heat transfer coefficients by using the finite element model. After the structural response analysis is completed, the temperature field of the topological structure with the heat dissipation and the unit sensitivity of each unit are calculated, and the density field is updated by using the unit sensitivity at the moment. During the update of the density field, the sensitivity of the cells is calculated except for the individual cells.
Step S400: determining a complete entity cell, a boundary cell and a hole cell.
In some embodiments, a key level value is determined according to the node sensitivity of each node, and when the node sensitivity of a node corresponding to a unit in the heat dissipation topological structure to be optimized is all greater than the key level value, the unit is a complete entity unit; when the node sensitivity of a node corresponding to one unit in the heat dissipation topological structure to be optimized is larger than a key level value, the unit is a boundary unit; and when the node sensitivity of the corresponding node of one unit in the heat dissipation topological structure to be optimized is less than the key level value, the unit is a cavity unit.
In some embodiments, the lowest node sensitivity value among all node sensitivities is obtained as a lower bound value of the interval, and the highest node sensitivity value among all node sensitivities is obtained as an upper bound value of the interval; and updating the lower bound value and the upper bound value by using a dichotomy, and calculating the average value of the updated lower bound value and the updated upper bound value to determine a key horizontal value. And determining a key level value based on the node sensitivity value of each mapped node to ensure that the volume fractions before and after the smoothing treatment are consistent. Firstly initializing one-dimensional search maximum node sensitivity as an upper bound value B by utilizing a dichotomy max Minimum node acuity is a lower bound B min Continuously updating and iterating the upper bound value and the lower bound value by using a dichotomy, wherein the key level value is B max And B min Average value of (a).
In some embodiments, since one unit corresponds to four nodes, two adjacent units necessarily have two same nodes, and thus a single unit corresponds to multiple nodes in the entire heat dissipation topology to be optimized. Based on the above, if the node sensitivity values of all nodes corresponding to a certain cell are smaller than the key level value, the cell is a hole cell; if the node sensitivity values of all nodes corresponding to a certain unit are greater than the key level value, all the units are complete entity units; if the node sensitivity value of all nodes corresponding to a certain cell is partially greater than the key level value, the cell is a boundary cell.
Step S500: and circularly updating the key level value.
In some embodiments, each boundary unit is respectively subdivided to obtain a corresponding new subdivision unit, and the sensitivity of a new subdivision node of the new subdivision unit corresponding to each boundary unit is calculated by using a second interpolation function; for each boundary unit, the density value of the boundary unit is respectively determined according to the sensitivity of the newly-divided node of the boundary unit; and determining volume fractions according to the density values of all the units, and updating the key level value according to the volume fractions.
In some embodiments, referring to fig. 5, the boundary cells determined in step S400 are subdivided, that is, each boundary cell is subdivided to obtain new subdivided cells. The node sensitivities of the corresponding nodes (i.e., the original nodes in fig. 5) of each of the border cells are interpolated using a second interpolation function to determine the new split node sensitivities of the new split cells corresponding to the new split nodes (i.e., the new nodes in fig. 5). Wherein the second interpolation function is represented by the following formula:
Figure 116563DEST_PATH_IMAGE007
wherein ns new (o) sensitivity of new subdivision node for new subdivision unit (xi) i ,η i ) Is the coordinate of the node corresponding to the boundary unit, i is the ith node corresponding to the boundary unit, ce is the total number of the nodes corresponding to the boundary unit, (xi) o ,η o ) For the seat of the new subdivision node corresponding to the new subdivision unitThe mark is that,
Figure 702265DEST_PATH_IMAGE009
the node sensitivity corresponding to the border cell.
In some embodiments, after a plurality of new subdivision units are subdivided for each boundary unit, the density value of the boundary unit corresponding to the new subdivision unit is determined according to the node sensitivity of the new subdivision node corresponding to the new subdivision unit. After the boundary units are determined according to the key level values, the boundary units are kept unchanged all the time, but each boundary unit needs to be subdivided to determine the density value of each boundary unit. The method comprises the following specific steps: and acquiring new subdivision nodes of the new subdivision units with the sensitivity of the new subdivision nodes higher than the key level value, and determining the density values of the boundary units according to the percentages of the new subdivision nodes occupying the nodes of the corresponding boundary units. Wherein the percentage is calculated using the formula:
Figure DEST_PATH_IMAGE019
wherein b is the percentage of the corresponding node of the new subdivision unit occupying the node of the corresponding boundary unit, (ξ) o ,η o ) Coordinates of the nodes corresponding to the new subdivision unit, ns m Sensitivity, ns, of new subdivision for new subdivision unit key Is the key level value, N node Is the total number of nodes of the border cell.
In some embodiments, in calculating the density value of each boundary cell, the volume fraction is determined collectively from the density values of all cells.
In some embodiments, when a cell is determined to be a fully physical cell, the density value of the cell is determined accordingly, wherein the density value of the fully physical cell is 1. When the cell is determined to be a hole cell, the density value of the cell is determined accordingly, wherein the density value of the hole cell is 0. Therefore, the density values of all the units in the whole heat dissipation topological structure to be optimized can be determined only by calculating the density values of the boundary units.
Obtaining an updated finite element modelAnd when the optimization target does not meet the convergence condition, updating the key level value according to the volume fraction. When the volume fraction is larger than the preset volume fraction, updating the current lower bound value into the current lower bound value by utilizing the dichotomy
Figure 538634DEST_PATH_IMAGE011
(ii) a When the volume fraction is smaller than the preset volume fraction, updating the upper bound value of the current time into the upper bound value by using a dichotomy
Figure 91844DEST_PATH_IMAGE012
(ii) a Wherein ns key Is a key level value, B min For the last updated lower bound value, B max The upper bound value of the last update; when the volume fraction and the preset volume fraction meet the set tolerance, updating the key level value to be ns key I.e. keeping the key level value calculated in step S400 unchanged.
In some embodiments, the tolerance between the calculated volume fraction and the preset volume fraction is a minimum value:
Figure DEST_PATH_IMAGE020
wherein e represents a cell e, ne represents the total number of cells, ρ e Denotes the density, v, of the element e e Denotes the volume of the cell e, V 0 For a predetermined volume fraction, ε is a defined convergence threshold, typically 10 -9
In some embodiments, after calculating the key level value satisfying the set volume fraction, the cells above the key level value are completely solid materials, the cells below the key level value are voids, and the cells equal to the key level value are boundary cells.
Referring to fig. 6, steps S300 to S400 are diagrams illustrating an optimization result after performing subsequent processing on the density field of the iterative structure.
In the heat dissipation topology optimization method for smooth boundary expression, after the optimization iteration is finished, the boundary smooth expression post-processing technology for realizing the zigzag-smooth boundary conversion is applied, and the key point is to realize the conversion of the structural topology expression paradigm based on units and nodes. And moreover, by matching with a continuity density filtering strategy and a rear boundary smooth expression post-processing technology, the aim of eliminating partial small-size detail features in multiple stages can be achieved, and the manufacturability of an optimization result is further improved. The boundary smoothing post-processing technology is used after optimization iteration, the influence of the post-processing technology on the optimization process is avoided, the robustness of heat transfer topology optimization is ensured, and the high calculation cost caused by repeated execution in the optimization process is avoided in the mode of decoupling post-processing and optimization iteration, so that the method is easier to expand to a large-scale problem.
Referring to fig. 7, the present application further provides a heat dissipation topology optimization apparatus 700 with smooth boundary representation, which includes a memory 710 and a processor 720, which are described in detail below.
The memory 710 is used to store programs.
The processor 720 is configured to implement the heat dissipation topology optimization method for smoothing the boundary expression in the embodiment of the present application by executing the program of the memory.
In some embodiments, the processor 710 first discretizes the heat dissipation topology to be optimized into a finite element model, and then calculates a density field of the heat dissipation topology to be optimized based on the finite element model. After obtaining the density field, processor 710 calculates a heat transfer coefficient using the first interpolation function, updates the finite element analysis model based on the heat transfer coefficient, and determines a cell sensitivity for each cell based on the updated finite element analysis model. The processor 710 stores a preset mapping relationship of a plurality of units corresponding to a single node, and the processor 710 obtains a unit sensitivity field of each unit of the plurality of units to determine a node sensitivity of the corresponding single node.
In some embodiments, after determining the node sensitivity of each node, the processor 710 determines a key level value according to the node sensitivity of each node, and when the node sensitivity of a node corresponding to a unit in the to-be-optimized heat dissipation topology is greater than the key level value, the unit is a complete entity unit; when the node sensitivity of a node corresponding to one unit in the heat dissipation topological structure to be optimized is larger than the key level value, the unit is a boundary unit; and when the node sensitivity of the node corresponding to one unit in the heat dissipation topological structure to be optimized is less than the key level value, the unit is a cavity unit.
The processor 710 divides each boundary cell to obtain a corresponding new divided cell. Processor 710 calculates a new partition node acuity for the new partition unit corresponding to each of the boundary units using a second interpolation function. For each boundary unit, the density value of the boundary unit is respectively determined according to the sensitivity of the new subdivision node, the processor 710 determines the volume fraction according to the density values of all the units, and updates the key level value according to the volume fraction so as to perform smooth boundary expression meeting the volume constraint on the result to be optimized.
Those skilled in the art will appreciate that all or part of the functions of the various methods in the above embodiments may be implemented by hardware, or may be implemented by computer programs. When all or part of the functions of the above embodiments are implemented by a computer program, the program may be stored in a computer-readable storage medium, and the storage medium may include: a read only memory, a random access memory, a magnetic disk, an optical disk, a hard disk, etc., and the program is executed by a computer to realize the above functions. For example, the program may be stored in a memory of the device, and when the program in the memory is executed by the processor, all or part of the functions described above may be implemented. In addition, when all or part of the functions in the above embodiments are implemented by a computer program, the program may be stored in a storage medium such as a server, another computer, a magnetic disk, an optical disk, a flash disk, or a portable hard disk, and may be downloaded or copied to a memory of a local device, or may be version-updated in a system of the local device, and when the program in the memory is executed by a processor, all or part of the functions in the above embodiments may be implemented.
The present invention has been described in terms of specific examples, which are provided to aid in understanding the invention and are not intended to be limiting. Numerous simple deductions, modifications or substitutions may also be made by those skilled in the art in light of the present teachings.

Claims (10)

1. A heat dissipation topology optimization method for smooth boundary expression is characterized by comprising the following steps:
dispersing a heat dissipation topological structure to be optimized into a finite element model;
calculating a density field of the heat dissipation topological structure to be optimized based on a finite element model, wherein the density field comprises a plurality of units;
calculating a heat transfer coefficient by using a first interpolation function, updating the finite element analysis model according to the heat transfer coefficient, and determining the element sensitivity of each element according to the updated finite element analysis model; the finite element analysis model is used for solving the finite element model; according to a preset mapping relation of a plurality of units corresponding to a single node, acquiring unit sensitivity of each unit of the plurality of units to determine node sensitivity of the corresponding single node;
determining a key level value according to the node sensitivity of each node, and when the node sensitivity of a node corresponding to a unit in the heat dissipation topological structure to be optimized is all larger than the key level value, the unit is a complete entity unit; when the node sensitivity of a node corresponding to one unit in the heat dissipation topological structure to be optimized is larger than the key level value, the unit is a boundary unit; when the node sensitivity of a node corresponding to one unit in the heat dissipation topological structure to be optimized is all smaller than the key level value, the unit is a cavity unit;
dividing each boundary unit to obtain a corresponding new division unit, and calculating the sensitivity of a new division node of the new division unit corresponding to each boundary unit by using a second interpolation function; for each boundary unit, the density value of the boundary unit is respectively determined according to the sensitivity of the newly-divided node of the boundary unit;
and determining volume fractions according to the density values of all the units, and updating the key level values according to the volume fractions so as to perform smooth boundary expression meeting volume constraints on the result to be optimized.
2. The method for optimizing heat dissipation topology represented by smooth boundaries according to claim 1, wherein said calculating a density field of said heat dissipation topology to be optimized based on a finite element model comprises:
controlling the filtering radius of a finite element model based on a continuous density strategy, carrying out a plurality of iterations on the density field to obtain an iterative optimization target, and when the variance of the optimization target is smaller than a set tolerance, ending the iteration and outputting the density field; the filtering radius is calculated by the following formula:
Figure QLYQS_1
wherein r is the iteration radius, k is the number of iterations, r min Is a predetermined minimum filter radius r 0 And for a preset initial filtering radius, lp is the iteration number of the set initial filtering radius attenuated to the minimum filtering radius.
3. The method of claim 1, wherein the calculating a heat transfer coefficient using a first interpolation function and updating the finite element analysis model based on the heat transfer coefficient comprises:
calculating a heat transfer coefficient in the first interpolation function using a penalty factor greater than a set value, the first interpolation function being represented by the following equation:
Figure QLYQS_2
wherein k is 0 For a heat transfer coefficient, ε is 10 -3 Rho is unit density and the value range is [0,1 ]]P is a penalty factor, and k is an interpolation coefficient.
4. The method for optimizing heat dissipation topology expressed by smooth boundaries according to claim 1, wherein the obtaining unit sensitivity of each of a plurality of units according to a preset mapping relationship of the plurality of units corresponding to a single node to determine node sensitivity of the corresponding single node comprises:
calculating the node sensitivity of the single node by using the following mapping formula:
Figure QLYQS_3
wherein ns m Node acuity for node m; ne is the total number of units; i is the ith unit and has a value range of [1,Ne];es i Cell sensitivity for the ith cell; w is a i,m Is a weighted function of the cell sensitivities of the number of cells.
5. The method for smoothing the representation of a boundary for heat dissipation topology optimization of claim 4, wherein the weighting function of the cell sensitivities of the number of cells is determined by the formula:
Figure QLYQS_4
wherein, w i,m A weighting function, r, of the cell sensitivities of the number of cells dist For the mapping range of the node m,
Figure QLYQS_5
is the distance from the ith cell to node m.
6. The method for smoothing the boundary expression for optimizing heat dissipation topology of claim 1, wherein determining a key level value according to the node sensitivity of each node in the heat dissipation topology to be optimized comprises:
acquiring a lower bound value of an interval of the minimum node sensitivity value in all node sensitivities and an upper bound value of an interval of the maximum node sensitivity value in all node sensitivities; and updating the lower bound value and the upper bound value by using a dichotomy, and calculating the average value of the updated lower bound value and the updated upper bound value to determine the key horizontal value.
7. The method for optimizing heat dissipation topology with smooth boundary representation according to claim 1, wherein the dividing each boundary cell to obtain a corresponding new divided cell, and calculating a new division node sensitivity of the new divided cell corresponding to each boundary cell by using a second interpolation function comprises:
interpolating the node sensitivity of each boundary unit by using a second interpolation function to obtain the new subdivision node sensitivity corresponding to the new subdivision unit; the second interpolation function is expressed by the following formula:
Figure QLYQS_6
wherein ns new (o) sensitivity of new subdivision node for new subdivision unit (xi) i ,η i ) Is the coordinate of the node corresponding to the boundary unit, i is the ith node corresponding to the boundary unit, ce is the total number of the nodes corresponding to the boundary unit, (xi) o ,η o ) For the coordinates of the new partitioning node corresponding to the new partitioning unit,
Figure QLYQS_7
the node acuity corresponding to the border cell.
8. The method for optimizing heat dissipation topology by smoothing boundary expression of claim 1, wherein for each boundary cell, the boundary cell determines its density value according to its newly-divided node acuity, comprising:
calculating the percentage of the nodes corresponding to the new subdivision units with the sensitivity of the new subdivision nodes larger than the key level value occupying the nodes of the corresponding boundary units, and determining the density values of the corresponding boundary units according to the percentage; the percentage is calculated using the following formula:
Figure QLYQS_8
wherein b is the percentage of the corresponding node of the new subdivision unit occupying the node of the corresponding boundary unit, (ξ) o ,η o ) Coordinates of the nodes corresponding to the new subdivision unit, ns m Sensitivity, ns, of new subdivision for new subdivision unit key Is the key level value, N node Is the total number of nodes of the border cell.
9. The method according to claim 6, wherein the updating the key level value according to the volume fraction to perform the smooth boundary expression satisfying the volume constraint on the result to be optimized comprises:
updating the key level value according to the volume fraction, and when the volume fraction is larger than the preset volume fraction, updating the current lower bound value to be
Figure QLYQS_9
(ii) a When the volume fraction is smaller than the preset volume fraction, updating the upper bound value of the current time to be
Figure QLYQS_10
(ii) a Wherein ns key As the key level value, B min Lower bound value, B, for the last update max The upper bound value is updated last time; and updating the key level value according to the current lower bound value and the current upper bound value.
10. A computer-readable storage medium, characterized in that the medium has stored thereon a program which is executable by a processor to implement the method according to any one of claims 1-9.
CN202211462679.7A 2022-11-22 2022-11-22 Heat dissipation topology optimization method for smooth boundary expression Active CN115495863B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202211462679.7A CN115495863B (en) 2022-11-22 2022-11-22 Heat dissipation topology optimization method for smooth boundary expression

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202211462679.7A CN115495863B (en) 2022-11-22 2022-11-22 Heat dissipation topology optimization method for smooth boundary expression

Publications (2)

Publication Number Publication Date
CN115495863A CN115495863A (en) 2022-12-20
CN115495863B true CN115495863B (en) 2023-03-14

Family

ID=85115442

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202211462679.7A Active CN115495863B (en) 2022-11-22 2022-11-22 Heat dissipation topology optimization method for smooth boundary expression

Country Status (1)

Country Link
CN (1) CN115495863B (en)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110069800A (en) * 2018-11-17 2019-07-30 华中科技大学 Three-dimensional structure method of topological optimization design and equipment with smooth boundary expression
CN112100882A (en) * 2020-08-27 2020-12-18 华南理工大学 Continuum structure density evolution topological optimization method with smooth boundary
CN113094943A (en) * 2021-03-15 2021-07-09 华中科技大学 Level set topology optimization method, system, device and medium

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3647973A1 (en) * 2018-11-04 2020-05-06 Dassault Systèmes Designing a mechanical part with topology optimization
CN110110413B (en) * 2019-04-26 2022-11-18 大连理工大学 Structural topology optimization method based on material field reduction progression expansion

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110069800A (en) * 2018-11-17 2019-07-30 华中科技大学 Three-dimensional structure method of topological optimization design and equipment with smooth boundary expression
CN112100882A (en) * 2020-08-27 2020-12-18 华南理工大学 Continuum structure density evolution topological optimization method with smooth boundary
CN113094943A (en) * 2021-03-15 2021-07-09 华中科技大学 Level set topology optimization method, system, device and medium

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
An improved ordered SIMP approach for multiscale concurrent topology optimization with multiple microstructures;Xuechen Gu et al.;《Composite Structures》;20220218;第1-15页 *
基于Lanczos法的模态重分析法在拓扑优化中的应用;刘丹 等;《中国机械工程》;20150630;第26卷(第11期);第1516-1520、1526页 *

Also Published As

Publication number Publication date
CN115495863A (en) 2022-12-20

Similar Documents

Publication Publication Date Title
WO2020215533A1 (en) Structural topology optimization method based on material-field reduction series expansion
Harding et al. Meta-parametric design
CN111125942B (en) B-spline high definition unit level set method and computer storage medium for three-dimensional unit structure modeling and topology optimization
JP7287397B2 (en) Information processing method, information processing apparatus, and information processing program
CN107729648A (en) A kind of wavy fiber composite structural design Waterfall type multilevel optimization method based on Shepard interpolation
Panzer et al. A greedy rational Krylov method for ℋ 2-pseudooptimal model order reduction with preservation of stability
CN110991621A (en) Method for searching convolutional neural network based on channel number
WO2022083527A1 (en) Method for determining logical core arrangement, model training method, electronic device and medium
CN106529044B (en) method for extracting 0-1 variable-configuration topological graph of compliant mechanism
CN104851133A (en) Image self-adaptive grid generation variational method
CN113821983A (en) Engineering design optimization method and device based on proxy model and electronic equipment
CN113935235A (en) Engineering design optimization method and device based on genetic algorithm and agent model
Chen et al. Tetrahedral mesh improvement by shell transformation
WO2019152597A1 (en) Methods for combinatorial constraint in topology optimization using shape transformation
CN116362194A (en) Wiring resource pre-allocation method, device, computing equipment and storage medium
CN113515824B (en) Topology optimization design method for cooperation of rib layout and substrate shape
CN115358136A (en) Structural rigidity optimization design method based on neural network
CN115495863B (en) Heat dissipation topology optimization method for smooth boundary expression
CN107480096B (en) High-speed parallel computing method in large-scale group simulation
CN113792458A (en) Method and device for optimizing finite element triangular mesh
CN113128617A (en) Spark and ASPSO based parallelization K-means optimization method
CN113610711A (en) Single-image-guided three-dimensional surface reconstruction method and device
Sinaei et al. Novel heuristic mapping algorithms for design space exploration of multiprocessor embedded architectures
Moukalled et al. The discretization process
CN116185377A (en) Optimization method and device for calculation graph and related product

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant