CN112733275B - Satellite component thermal layout temperature field prediction method based on semi-supervised learning - Google Patents
Satellite component thermal layout temperature field prediction method based on semi-supervised learning Download PDFInfo
- Publication number
- CN112733275B CN112733275B CN202110068649.7A CN202110068649A CN112733275B CN 112733275 B CN112733275 B CN 112733275B CN 202110068649 A CN202110068649 A CN 202110068649A CN 112733275 B CN112733275 B CN 112733275B
- Authority
- CN
- China
- Prior art keywords
- satellite component
- satellite
- temperature field
- layout
- loss function
- 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
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/15—Vehicle, aircraft or watercraft design
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/23—Design optimisation, verification or simulation using finite element methods [FEM] or finite difference methods [FDM]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/088—Non-supervised learning, e.g. competitive learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/08—Thermal analysis or thermal optimisation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- Computer Hardware Design (AREA)
- Biophysics (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Pure & Applied Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Mathematical Optimization (AREA)
- Medical Informatics (AREA)
- Mathematical Analysis (AREA)
- Computational Mathematics (AREA)
- Aviation & Aerospace Engineering (AREA)
- Automation & Control Theory (AREA)
- Investigating Or Analyzing Materials Using Thermal Means (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a satellite component thermal layout temperature field prediction method based on semi-supervised learning, which comprises the following steps: building a structural model of satellite component layout; generating a plurality of training data sets comprising satellite component layouts and temperature field distributions corresponding to the satellite component layouts, and generating a plurality of test data sets comprising satellite component layouts and temperature field distributions corresponding to the satellite component layouts; constructing a supervised learning loss function of embedded physical knowledge; training a deep learning model; and inputting the satellite component layout of the temperature field distribution to be calculated into a trained deep learning model, and obtaining the temperature field distribution corresponding to the satellite component layout. The invention uses a large number of training samples without labels, a small number of training samples with labels and a supervised learning loss function with embedded physical knowledge to train the deep learning model, can realize the rapid calculation of the temperature field distribution of the satellite assembly, and has the advantages of less calculation resources and workload, and high precision and efficiency.
Description
Technical Field
The invention relates to the technical field of satellite layout design, in particular to a satellite component thermal layout temperature field prediction method based on semi-supervised learning.
Background
From mobile phones, televisions, automobiles in daily life, satellites, tanks, unmanned aerial vehicles and the like in the national defense industry, electronic equipment is quite common, and in the working process of the electronic equipment, heat dissipation inevitably occurs due to power loss of the electronic equipment, so that the temperature of the electronic equipment is increased. With the improvement of the functional requirements of modern electronic devices and the progress of the modern industrial manufacturing process, the internal integration level of the electronic devices is higher and higher, and the internal structure of the corresponding satellite assembly is more and more complex. Because the satellite works in space, when the satellite component on the satellite has a problem, the maintenance and the replacement cannot be performed in time; in order to ensure that the satellite can operate efficiently for a long period of time to perform tasks, satellite components are required to have better service performance and longer service lives. However, the high integration and complexity of the satellite assembly easily lead the internal temperature to rise sharply during normal operation, which has important influence on the service life, safety, reliability and the like of electronic components in the satellite assembly, thereby influencing the service performance and service life of the whole satellite assembly. Therefore, the electronic components inside the satellite assembly need to be reasonably arranged to control the temperature field distribution of the satellite assembly, so that the thermal load of the satellite assembly is reduced, and the service performance and the service life of the satellite assembly are improved.
With the development of computer technology, a computer simulation technology is generally adopted to complete the thermal design of a satellite component so as to improve the thermal design efficiency of the satellite component. And by using a computer simulation technology, continuously adjusting the positions of components in the satellite assembly through repeated tests, and finally finding out the satellite assembly layout with reasonable temperature field distribution. In the thermal analysis of satellite components by using a computer simulation technology, the calculation of a temperature field needs to be completed by a mathematical method, and the calculation method commonly used at present mainly comprises an analysis method and a numerical solution method. The method is limited by the difficulty of solving a high-order differential equation, and an accurate temperature field model is difficult to obtain. Numerical solutions include finite difference, finite volume and finite element methods, which theoretically can achieve any desired accuracy, however, higher accuracy is accompanied by a large number of matrix calculations, and satellite component layout changes once, requiring a re-calculation of the temperature field distribution, resulting in low test efficiency and long test period.
In order to solve the technical problem existing when the thermal analysis of the satellite component is performed by using the computer simulation technology to complete the calculation of the temperature field, the prediction of the layout temperature field of the satellite component is currently realized by constructing a depth proxy model by using a deep learning method in view of the universal approximation property of the deep learning method. However, in the training process of the depth proxy model, a large number of training samples with labels are required, that is, the training samples of different satellite component layouts should have temperature field distributions corresponding to the training samples as labels, but true value samples of the satellite component layouts and the temperature field distributions are difficult to obtain, at this time, the satellite component layout temperature fields still need to be calculated by mathematical methods such as a finite difference method, so that a large amount of calculation resources need to be consumed, the design cost is high, and the design period is long.
Disclosure of Invention
In order to solve part or all of the technical problems in the prior art, the invention provides a satellite component thermal layout temperature field prediction method based on semi-supervised learning.
Therefore, the invention discloses a satellite component thermal layout temperature field prediction method based on semi-supervised learning, which comprises the following steps:
approximately describing the structure of the satellite component, and establishing a structural model of the satellite component layout;
generating a plurality of training data sets comprising satellite component layouts and temperature field distributions corresponding to the satellite component layouts, and generating a plurality of test data sets comprising satellite component layouts and temperature field distributions corresponding to the satellite component layouts;
constructing a supervised learning loss function of embedded physical knowledge;
training the deep learning model by using a training data set and a test data set based on the supervised learning loss function of the embedded physical knowledge to fit the mapping relation between the satellite component layout and the temperature field distribution;
and inputting the satellite component layout of the temperature field distribution to be calculated into a trained deep learning model, and obtaining the temperature field distribution corresponding to the satellite component layout.
In some alternative embodiments, the structural model of the satellite component layout is built in the following manner:
setting the satellite component layout as a square area, dividing the square area into n 1 ×n 1 A grid, a small hole with a set length is arranged on one of four sides of the square area to be used as a radiating hole, and the temperature of the radiating hole area is constant to be T 0 ；
The electronic components are set to be square structures, one electronic component can be placed on one grid position, and different electronic components are placed on different grid positions according to the specific position of each electronic component in the satellite assembly layout.
In some alternative embodiments, generating the training data set and the test data set includes:
randomly selecting a plurality of grids in a square area, placing electronic components on the selected grids to obtain a satellite component layout, and repeating the random generation process for a plurality of times to obtain N satellite component layouts;
randomly extracting M satellite component layouts from N satellite component layouts, determining temperature field distribution corresponding to the M satellite component layouts, and dividing the M satellite component layouts into M 1 First group of satellite component layouts includes M 2 A second set of satellite component layouts, generating N-M including satellite component layouts and M 1 Generating M by training data set comprising satellite component layout and temperature field distribution corresponding to the satellite component layout 2 A test data set comprising a satellite component layout and a temperature field distribution corresponding to the satellite component layout N, M, M 1 And M 2 Are all positive integers, m=m 1 +M 2 。
In some alternative embodiments, the supervised learning penalty function to build embedded physical knowledge is:
L＝w 1 L laplace +w 2 L outside +w 3 L label
wherein w is 1 、w 2 And w 3 Respectively represent a heat conduction loss function based on a heat conduction equation, a boundary loss function based on boundary conditions, and an L1 normWeight parameters of the loss function, L laplace Representing a heat transfer loss function based on a heat transfer equation, L outside Representing a boundary loss function based on boundary conditions, L label Representing the L1 norm loss function.
In some alternative embodiments, the heat transfer loss function based on the heat transfer equation is:
wherein, the liquid crystal display device comprises a liquid crystal display device,represents the temperature of the heat source output with position coordinates x and y in the square grid area, l represents the side length of the square grid area, n represents the grid number, +.>Representing normalization operation, heat source intensity corresponding to heat sources with Input (x, y) position coordinates of x and y>And (3) performing normalization processing to obtain a system input, wherein m represents a normalization constant.
In some alternative embodiments, the boundary loss function based on the boundary condition is:
L outside ＝|T output | Ω |
wherein T is output | Ω Indicating the temperature output at the heat sink area.
In some alternative embodiments, the L1 norm loss function is:
x label representing labeled training samples, T output Indicating the temperature of the output of the heat source,representing the true temperature corresponding to the sample.
In some alternative embodiments, the deep learning model employs a neural network model with a body structure that is a feature pyramid network.
The technical scheme of the invention has the main advantages that:
according to the satellite component thermal layout temperature field prediction method based on semi-supervised learning, the calculation and determination of the satellite component temperature field distribution are realized by adopting a deep learning model, so that the problems of large workload and low efficiency caused by adopting a traditional numerical solution can be avoided; meanwhile, on the basis of adopting a deep learning model, the supervised learning loss function based on embedded physical knowledge utilizes a large number of training samples without labels and a small number of training samples with labels to complete the training of the deep learning model, so that the problem that the existing deep learning model needs a large number of samples with labels during training can be solved, the calculation resources and the calculation workload are effectively saved, the working efficiency is improved, and the calculation precision of the satellite component thermal distribution temperature field distribution can be ensured.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flowchart of a satellite component thermal layout temperature field prediction method based on semi-supervised learning according to an embodiment of the present invention;
fig. 2 is a schematic diagram of a satellite component layout according to an embodiment of the invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be clearly and completely described below with reference to specific embodiments of the present invention and corresponding drawings. It will be apparent that the described embodiments are only some, but not all, embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The following describes in detail the technical scheme provided by the embodiment of the invention with reference to the accompanying drawings.
As shown in fig. 1, an embodiment of the present invention provides a satellite component thermal layout temperature field prediction method based on semi-supervised learning, which includes the following steps:
s1, approximately describing the structure of a satellite component, and establishing a structural model of satellite component layout;
s2, generating a plurality of training data sets comprising satellite component layouts and temperature field distributions corresponding to the satellite component layouts, and generating a plurality of test data sets comprising the satellite component layouts and the temperature field distributions corresponding to the satellite component layouts;
s3, constructing a supervised learning loss function embedded with physical knowledge;
s4, training the deep learning model by utilizing a training data set and a testing data set based on the supervised learning loss function of the embedded physical knowledge so as to fit the mapping relation between the satellite component layout and the temperature field distribution;
s5, inputting the satellite component layout of the temperature field distribution to be calculated into the trained deep learning model, and obtaining the temperature field distribution corresponding to the satellite component layout.
The following describes each step and principle of the satellite component thermal layout temperature field prediction method based on semi-supervised learning according to an embodiment of the present invention.
S1, approximately describing the structure of a satellite component, and establishing a structural model of satellite component layout;
specifically, based on the structural features and layout features of the satellite components, the structure of the satellite components is approximately described, and a structural model of the satellite component layout is established.
In an embodiment of the present invention, the structural features and layout features of the satellite assembly are: the satellite component layout is a set area layout, a set number of electronic components with different sizes are distributed in the set area, and the electronic components continuously generate heat dissipation during normal operation, namely are regarded as heat sources; the satellite assembly radiates heat in a natural radiating mode, radiating holes with set sizes are formed in the upper side of the satellite assembly layout, the temperature at the radiating holes is constant, and the periphery of the satellite assembly layout is insulated except the radiating holes.
As shown in fig. 2, based on the structural features and layout features of the satellite assembly, the following manner is adopted to build a structural model of the satellite assembly layout:
setting the satellite component layout as a square area, and meshing the square area to divide the square area into n 1 ×n 1 A grid, a small hole with a set length is arranged on one of four sides of the square layout area to be used as a radiating hole, and the temperature of the radiating hole area is constant to be T 0 The method comprises the steps of carrying out a first treatment on the surface of the The electronic components are set to be square structures, one electronic component can be placed on one grid position, and different electronic components are placed on different grid positions according to the specific position of each electronic component in the satellite assembly layout.
In an embodiment of the present invention, for different satellite component layouts, the above manner is used to obtain structural models corresponding to different satellite component layouts.
S2, generating a plurality of training data sets comprising satellite component layouts and temperature field distributions corresponding to the satellite component layouts, and generating a plurality of test data sets comprising the satellite component layouts and the temperature field distributions corresponding to the satellite component layouts;
specifically, based on the above-established structural model of satellite component layout, a plurality of grids in a square area are randomly selected, electronic components are placed on the selected grids to obtain a satellite component layout, the random generation process is repeated for a plurality of times to obtain N satellite component layouts, and the N satellite component layouts are obtained and then the N satellite component layouts are mixed with each otherExtracting M satellite component layouts by a machine, calculating temperature field distribution corresponding to the M satellite component layouts by using a finite difference method or a finite element method or an feics decomposer, and dividing the M satellite component layouts into M components 1 First group of satellite component layouts includes M 2 A second set of satellite component layouts, generating N-M including satellite component layouts and M 1 Generating M by training data set comprising satellite component layout and temperature field distribution corresponding to the satellite component layout 2 A test data set comprising a satellite component layout and a temperature field distribution corresponding to the satellite component layout N, M, M 1 And M 2 Are all positive integers, m=m 1 +M 2 。
In one embodiment of the invention, NM; alternatively, n=50m.
The more the number of training data sets and test data sets, the higher the prediction accuracy of the trained deep learning model, but the more the number of training data sets and test data sets, the larger the corresponding calculation workload. To this end, in an embodiment of the present invention, 50000 satellite component layouts are randomly acquired, 1000 satellite component layouts are randomly extracted from the 50000 satellite component layouts, temperature field distributions corresponding to the 1000 satellite component layouts are determined, the 1000 satellite component layouts are divided into a first group including 10 satellite component layouts and a second group including 990 satellite component layouts, 49000 training data sets including satellite component layouts and 10 temperature field distributions corresponding to the satellite component layouts are generated, and 990 test data sets including satellite component layouts and temperature field distributions corresponding to the satellite component layouts are generated.
S3, constructing a supervised learning loss function embedded with physical knowledge;
the satellite component temperature field distribution is set to be unchanged with time, and can be solved through a heat conduction equation shown in a formula 1;
in the formula, x and y are as followsShowing the coordinates of a point in a two-dimensional plane, T representing the temperature at that point, k representing the thermal conductivity,representing the intensity of the heat source at that point; for a square grid area corresponding to the satellite component layout, x and y represent the position coordinates of the grid.
Because the satellite component temperature field can be solved through the heat conduction equation, in one embodiment of the invention, the heat conduction loss function of the deep learning model is constructed based on the heat conduction equation, and the supervised learning loss function of the embedded physical knowledge is obtained to conduct targeted guidance on the training process of the deep learning model.
Specifically, for the heat conduction equation shown in equation 1, set is:
the problem of solving the heat conduction equation can be converted into minimizing S 0 Is a problem of (a). In one embodiment of the present invention, the thermal conductivity k=1, for the intensity of the heat sourceDividing m by m as the normalization process yields the system Input (x, y), then equation 2 can be further derived as:
in the formula, T x,y Representing the temperature at points of position coordinates x and y, T x+1,y Temperature, T, at points representing position coordinates x+1 and y x-1,y Temperature, T, at points representing the position coordinates x-1 and y x,y+1 Representing the temperature at points with position coordinates x and y+1, T x,y-1 Representing the temperature at points with position coordinates x and y-1, and m represents the normalization constant for normalizing the neural network input.
Due to the sanitationThe star component layout is set as a square grid area, the side length of the square grid area is l, the grid number is n, and n=n 1 ×n 1 Δx=Δy=l/x.
Further, equation 3 may be derived as:
normalization operation of equation 4 yields:
further, the heat transfer loss function can be expressed as:
in the formula (i),representing normalization operation, T 0 Indicating the temperature at the heat sink.
Setting the output temperature of the heat source asThe heat transfer loss function can be rewritten as:
according to the above-described derivation process, a heat conduction loss function based on a heat conduction equation can be constructed.
Further, because a certain boundary condition is also required for solving the satellite component temperature field distribution, a boundary loss function based on the boundary condition is also required to be constructed. In one embodiment of the present invention, the temperature at the heat dissipation hole of the satellite component layout is constant to be T 0 Namely, the following equation 8 needs to be satisfied at the heat dissipation hole region Ω:
T output | Ω indicating the temperature output at the heat sink area.
For this purpose, a boundary loss function based on boundary conditions is constructed as follows:
L outside ＝|T output | Ω equation 9
Other boundary conditions, such as a first class boundary condition (Dirichlet boundary condition), a second class boundary condition (Neumann boundary condition), and a third class boundary condition (Robin boundary condition), may also be satisfied due to the solution of the temperature field distribution; in an embodiment of the present invention, except for heat dissipation at the heat dissipation holes, the rest boundary conditions are realized by copying the boundary, that is, the internal and external temperatures at the boundary of the satellite component layout are the same.
Further, in an embodiment of the present invention, prior physical knowledge is embedded into a deep learning model loss function, so as to conduct targeted guidance on a training process of the deep learning model, and meanwhile, a small amount of labeled samples are introduced, wherein the labeled samples comprise satellite component layouts and training samples of temperature field distribution corresponding to the satellite component layouts, and the labeled supervised loss function is used as a regularization term to be combined with the loss function of the embedded physical knowledge.
For the supervised learning section, L1 loss is employed as a corresponding loss function to assist in the unsupervised training process.
Specifically, the L1 norm loss function is constructed as:
in the formula (i),representing the true temperature, x, of the sample label Representing a labeled training sample.
Further, on the basis of the constructed heat conduction loss function based on the heat conduction equation, the boundary loss function based on the boundary condition and the L1 norm loss function, the supervised learning loss function with embedded physical knowledge is as follows:
L＝w 1 L laplace +w 2 L outside +w 3 L label equation 11
In the formula, w 1 、w 2 And w 3 The weight parameters of the heat conduction loss function based on the heat conduction equation, the boundary loss function based on the boundary condition, and the L1 norm loss function are represented, respectively.
S4, training the deep learning model by utilizing a training data set and a testing data set based on the supervised learning loss function of the embedded physical knowledge so as to fit the mapping relation between the satellite component layout and the temperature field distribution;
specifically, in an embodiment of the present invention, the deep learning model adopts a neural network model with a main structure as a feature pyramid network (Feature Pyramid Networks, FPN), and the feature pyramid network uses a residual network (res net) as a skeleton, and is divided into a bottom-up path, a top-down path and a middle connection, so that feature maps with different dimensions output by each layer are fused and mapped, and distribution output of a satellite component layout temperature field can be obtained.
In one embodiment of the present invention, the input of the deep learning model is n 1 ×n 1 The satellite components are laid out at the grid scale, and the value of each grid indicates whether or not there is an electronic component that can generate heat at the grid location.
Because the deep learning model based on supervised learning requires a large number of labeled training data sets during training, the temperature field distribution predicted by the deep learning model based on unsupervised learning usually changes to a certain extent, so that the maximum temperature predicted is usually different from the actual temperature to a certain extent, and the prediction error is larger.
In an embodiment of the invention, a large number of samples including satellite component layout and a small number of samples including satellite component layout and temperature field distribution corresponding to the satellite component layout are randomly generated as training data sets, a small number of samples including satellite component layout and temperature field distribution corresponding to the satellite component layout are simultaneously generated as test data sets, a deep learning model is trained by adopting error back propagation based on a supervised learning loss function of built-in physical knowledge, and the deep learning model is stored after the training reaches a set training frequency.
In an embodiment of the present invention, when the total number of training data sets and test data sets is 50000, the number of training iterations may be set to 50.
S5, inputting the satellite component layout of the temperature field distribution to be calculated into a trained deep learning model, and obtaining the temperature field distribution corresponding to the satellite component layout;
further, after training and storing of the deep learning model are completed, the stored deep learning model is loaded, and the satellite component layout of the temperature field distribution to be calculated is input into the deep learning model to obtain the temperature field distribution corresponding to the satellite component layout, so that the optimization design of the satellite component layout is assisted.
Therefore, according to the satellite component thermal layout temperature field prediction method based on semi-supervised learning, which is provided by the embodiment of the invention, the calculation and the determination of the satellite component temperature field distribution are realized by adopting a deep learning model, so that the problems of large workload and low efficiency caused by adopting a traditional numerical solution can be avoided; meanwhile, on the basis of adopting a deep learning model, the supervised learning loss function based on embedded physical knowledge utilizes a large number of training samples without labels and a small number of training samples with labels to complete the training of the deep learning model, so that the problem that the existing deep learning model needs a large number of samples with labels during training can be solved, the calculation resources and the calculation workload are effectively saved, the working efficiency is improved, and the calculation precision of the satellite component thermal distribution temperature field distribution can be ensured.
It should be noted that in this document, relational terms such as "first" and "second" and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. In this context, "front", "rear", "left", "right", "upper" and "lower" are referred to with respect to the placement state shown in the drawings.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting thereof; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (4)
1. The satellite component thermal layout temperature field prediction method based on semi-supervised learning is characterized by comprising the following steps of:
approximately describing the structure of the satellite component, and establishing a structural model of the satellite component layout;
generating a plurality of training data sets comprising satellite component layouts and temperature field distributions corresponding to the satellite component layouts, and generating a plurality of test data sets comprising satellite component layouts and temperature field distributions corresponding to the satellite component layouts;
constructing a supervised learning loss function of embedded physical knowledge;
training the deep learning model by using a training data set and a test data set based on the supervised learning loss function of the embedded physical knowledge to fit the mapping relation between the satellite component layout and the temperature field distribution;
inputting the satellite component layout of the temperature field distribution to be calculated into a trained deep learning model, and obtaining the temperature field distribution corresponding to the satellite component layout;
wherein, the supervised learning loss function of the embedded physical knowledge is constructed as follows:
L＝w 1 L laplace +w 2 L outside +w 3 L label
w 1 、w 2 and w 3 Weight parameters respectively representing a heat conduction loss function based on a heat conduction equation, a boundary loss function based on boundary conditions, and an L1 norm loss function, L laplace Representing a heat transfer loss function based on a heat transfer equation, L outside Representing a boundary loss function based on boundary conditions, L label Representing an L1 norm loss function;
the heat conduction loss function based on the heat conduction equation is:
represents the temperature of the heat source output with position coordinates x and y in the square grid area, l represents the side length of the square grid area, n represents the grid number, +.>Representing normalization operation, heat source intensity corresponding to heat sources with Input (x, y) position coordinates of x and y>Performing normalization processing to obtain a system input, wherein m represents a normalization constant;
the boundary loss function based on the boundary condition is:
L outside ＝|T output | Ω |
T output | Ω representing the temperature output at the heat sink region;
the L1 norm loss function is:
x label representing labeled training samples, T output Indicating the temperature of the output of the heat source,representing the true temperature corresponding to the sample.
2. The method for predicting the thermal layout temperature field of a satellite assembly based on semi-supervised learning as recited in claim 1, wherein the structural model of the satellite assembly layout is created by:
setting the satellite component layout as a square area, dividing the square area into n 1 ×n 1 A grid, a small hole with a set length is arranged on one of four sides of the square area to be used as a radiating hole, and the temperature of the radiating hole area is constant to be T 0 ；
The electronic components are set to be square structures, one electronic component can be placed on one grid position, and different electronic components are placed on different grid positions according to the specific position of each electronic component in the satellite assembly layout.
3. The semi-supervised learning based satellite component thermal layout temperature field prediction method of claim 2, wherein generating training and test data sets comprises:
randomly selecting a plurality of grids in a square area, placing electronic components on the selected grids to obtain a satellite component layout, and repeating the random generation process for a plurality of times to obtain N satellite component layouts;
random extraction from N satellite component layoutsM satellite component layouts are taken, temperature field distribution corresponding to the M satellite component layouts is determined, and the M satellite component layouts are divided into M components 1 First group of satellite component layouts includes M 2 A second set of satellite component layouts, generating N-M including satellite component layouts and M 1 Generating M by training data set comprising satellite component layout and temperature field distribution corresponding to the satellite component layout 2 A test data set comprising a satellite component layout and a temperature field distribution corresponding to the satellite component layout N, M, M 1 And M 2 Are all positive integers, m=m 1 +M 2 。
4. A satellite component thermal layout temperature field prediction method based on semi-supervised learning as recited in any of claims 1-3, wherein the deep learning model employs a neural network model with a body structure as a feature pyramid network.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110068649.7A CN112733275B (en) | 2021-01-19 | 2021-01-19 | Satellite component thermal layout temperature field prediction method based on semi-supervised learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110068649.7A CN112733275B (en) | 2021-01-19 | 2021-01-19 | Satellite component thermal layout temperature field prediction method based on semi-supervised learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112733275A CN112733275A (en) | 2021-04-30 |
| CN112733275B true CN112733275B (en) | 2023-07-25 |
Family
ID=75592400
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110068649.7A Active CN112733275B (en) | 2021-01-19 | 2021-01-19 | Satellite component thermal layout temperature field prediction method based on semi-supervised learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112733275B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113609763B (en) * | 2021-07-27 | 2022-11-11 | 中国人民解放军军事科学院国防科技创新研究院 | Uncertainty-based satellite component layout temperature field prediction method |
| CN114781207B (en) * | 2022-03-29 | 2024-04-12 | 中国人民解放军军事科学院国防科技创新研究院 | Heat source layout temperature field prediction method based on uncertainty and semi-supervised learning |
| CN115630565B (en) * | 2022-09-28 | 2023-12-15 | 中国人民解放军军事科学院国防科技创新研究院 | Temperature field reconstruction method based on embedded physical knowledge neural network |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111222648A (en) * | 2020-01-15 | 2020-06-02 | 深圳前海微众银行股份有限公司 | Semi-supervised machine learning optimization method, device, equipment and storage medium |
| CN111507032A (en) * | 2020-03-31 | 2020-08-07 | 中国人民解放军军事科学院国防科技创新研究院 | Deep learning technology-based component layout optimization design method for predicting temperature distribution |
| CN111782459A (en) * | 2020-06-03 | 2020-10-16 | 东南大学 | Traveling wave tube internal temperature prediction method based on tube external temperature image and CNN |
| CN111783337A (en) * | 2020-06-28 | 2020-10-16 | 中国人民解放军军事科学院国防科技创新研究院 | Hyper-parameter optimization method of satellite thermal layout agent model |
| CN111783331A (en) * | 2020-06-04 | 2020-10-16 | 中国人民解放军军事科学院国防科技创新研究院 | Satellite component temperature field determination method based on transfer learning |
-
2021
- 2021-01-19 CN CN202110068649.7A patent/CN112733275B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111222648A (en) * | 2020-01-15 | 2020-06-02 | 深圳前海微众银行股份有限公司 | Semi-supervised machine learning optimization method, device, equipment and storage medium |
| CN111507032A (en) * | 2020-03-31 | 2020-08-07 | 中国人民解放军军事科学院国防科技创新研究院 | Deep learning technology-based component layout optimization design method for predicting temperature distribution |
| CN111782459A (en) * | 2020-06-03 | 2020-10-16 | 东南大学 | Traveling wave tube internal temperature prediction method based on tube external temperature image and CNN |
| CN111783331A (en) * | 2020-06-04 | 2020-10-16 | 中国人民解放军军事科学院国防科技创新研究院 | Satellite component temperature field determination method based on transfer learning |
| CN111783337A (en) * | 2020-06-28 | 2020-10-16 | 中国人民解放军军事科学院国防科技创新研究院 | Hyper-parameter optimization method of satellite thermal layout agent model |
Non-Patent Citations (1)
| Title |
|---|
| AI技术与卫星资料应用研究现状分析;匡秋明;于廷照;;气象科技进展(03);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112733275A (en) | 2021-04-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112733275B (en) | Satellite component thermal layout temperature field prediction method based on semi-supervised learning | |
| CN112257321B (en) | Unsupervised learning-based satellite component temperature field determination method | |
| Chen et al. | The heat source layout optimization using deep learning surrogate modeling | |
| Nautiyal et al. | Improved salp swarm algorithm with mutation schemes for solving global optimization and engineering problems | |
| CN111507032B (en) | Deep learning technology-based component layout optimization design method for predicting temperature distribution | |
| CN113177290B (en) | Satellite component temperature field prediction method based on depth agent model normalization | |
| Szpiech et al. | On the size distribution of private microsatellite alleles | |
| CN111783331B (en) | Satellite component temperature field determination method based on transfer learning | |
| Rao et al. | A time-parallel approach to strong-constraint four-dimensional variational data assimilation | |
| Otaki et al. | Thermal design optimization of electronic circuit board layout with transient heating chips by using Bayesian optimization and thermal network model | |
| CN114036787A (en) | Satellite component thermal layout temperature field prediction based on physical prior self-supervision learning | |
| CN114781610A (en) | Data processing method, neural network training method and related equipment | |
| US20210295167A1 (en) | Generative networks for physics based simulations | |
| CN113221326B (en) | Satellite component temperature field prediction method based on teacher-student agent model | |
| Wang | Evidence-theory-based uncertain parameter identification method for mechanical systems with imprecise information | |
| WO2022247092A1 (en) | Methods and systems for congestion prediction in logic synthesis using graph neural networks | |
| Bai et al. | DEIM reduced order model constructed by hybrid snapshot simulation | |
| CN114444379A (en) | Temperature field reconstruction method based on data-driven deep learning | |
| CN114781207B (en) | Heat source layout temperature field prediction method based on uncertainty and semi-supervised learning | |
| CN112347638B (en) | Dual-unit method based three-dimensional integrated micro-system electric-thermal coupling analysis method | |
| Raghavan et al. | Implicit constraint handling for shape optimisation with pod-morphing | |
| Deng et al. | Data-driven physics-constrained recurrent neural networks for multiscale damage modeling of metallic alloys with process-induced porosity | |
| Fourmigue et al. | Explicit transient thermal simulation of liquid-cooled 3D ICs | |
| Fabian et al. | Estimating the execution time of the coupled stage in multiscale numerical simulations | |
| US20210173983A1 (en) | Systems and methods for machine learning based fast static thermal solver |
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 |