CN113255181A - Heat transfer inverse problem identification method and device based on deep learning - Google Patents
Heat transfer inverse problem identification method and device based on deep learning Download PDFInfo
- Publication number
- CN113255181A CN113255181A CN202110459377.3A CN202110459377A CN113255181A CN 113255181 A CN113255181 A CN 113255181A CN 202110459377 A CN202110459377 A CN 202110459377A CN 113255181 A CN113255181 A CN 113255181A
- Authority
- CN
- China
- Prior art keywords
- wall surface
- parameter
- surface temperature
- corrected
- deep learning
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 68
- 238000013135 deep learning Methods 0.000 title claims abstract description 54
- 238000012546 transfer Methods 0.000 title claims abstract description 53
- 238000003062 neural network model Methods 0.000 claims abstract description 29
- 238000013528 artificial neural network Methods 0.000 claims abstract description 15
- 238000003384 imaging method Methods 0.000 claims abstract description 8
- 238000012549 training Methods 0.000 claims description 50
- 238000004422 calculation algorithm Methods 0.000 claims description 20
- 238000012360 testing method Methods 0.000 claims description 19
- 238000004590 computer program Methods 0.000 claims description 11
- 238000013527 convolutional neural network Methods 0.000 claims description 11
- 238000009826 distribution Methods 0.000 claims description 9
- 238000013507 mapping Methods 0.000 claims description 7
- 238000012937 correction Methods 0.000 claims description 6
- 238000003860 storage Methods 0.000 claims description 4
- 230000006870 function Effects 0.000 abstract description 10
- 230000007547 defect Effects 0.000 abstract description 4
- 238000004364 calculation method Methods 0.000 description 9
- 238000010586 diagram Methods 0.000 description 7
- 238000004891 communication Methods 0.000 description 6
- 230000008569 process Effects 0.000 description 6
- 238000000605 extraction Methods 0.000 description 5
- 238000010606 normalization Methods 0.000 description 5
- 238000005457 optimization Methods 0.000 description 5
- 239000000463 material Substances 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 238000004088 simulation Methods 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 230000008901 benefit Effects 0.000 description 3
- 239000012530 fluid Substances 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 238000007796 conventional method Methods 0.000 description 2
- 238000005520 cutting process Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000012897 Levenberg–Marquardt algorithm Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 238000002939 conjugate gradient method Methods 0.000 description 1
- 238000013075 data extraction Methods 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 206010020718 hyperplasia Diseases 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 235000001968 nicotinic acid Nutrition 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 230000000704 physical effect Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000006641 stabilisation Effects 0.000 description 1
- 238000011105 stabilization Methods 0.000 description 1
- 238000002945 steepest descent method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- 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
- G06F30/28—Design optimisation, verification or simulation using fluid dynamics, e.g. using Navier-Stokes equations or computational fluid dynamics [CFD]
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2113/00—Details relating to the application field
- G06F2113/08—Fluids
-
- 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
-
- 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
- G06F2119/14—Force analysis or force optimisation, e.g. static or dynamic forces
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Software Systems (AREA)
- Data Mining & Analysis (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- Algebra (AREA)
- Fluid Mechanics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Investigating Or Analyzing Materials Using Thermal Means (AREA)
Abstract
The embodiment of the invention provides a method and a device for identifying inverse problems of heat transfer science based on deep learning, wherein the method comprises the following steps: acquiring an inner wall surface temperature cloud picture; inputting the inner wall surface temperature cloud picture into a preset neural network model to obtain a parameter to be corrected corresponding to the inner wall surface temperature cloud picture; wherein the parameter to be corrected comprises a convection heat exchange coefficient, and/or an incoming flow temperature, and/or a heat source intensity; the embodiment of the invention simplifies the flow of the inverse problem of heat transfer, inputs the imaging temperature data into the trained deep neural network, and the output result is the parameter to be corrected, thereby overcoming the defects of difficulty, discomfort, initial value dependence, easy falling into local convergence and the like of target function selection in the traditional method, and reducing errors caused by calculated amount and operation.
Description
Technical Field
The invention relates to the technical field of inverse heat transfer problems, in particular to a method and a device for identifying inverse heat transfer problems based on deep learning.
Background
The inverse problem of heat transfer is one of the branches of all inverse problems, as opposed to the positive problem of heat transfer. For the positive thermal conductivity problem, if the physical parameters of the control equation, the geometry of the solution domain, the initial conditions and the boundary conditions are known, the temperature field can be directly solved by numerical methods. However, in many practical engineering problems, many conditions cannot be directly obtained due to the limitation of measurement technology, so that a counter-problem research method is proposed. The inverse problem of heat transfer is to invert unknown solution parameters by measuring temperature values (or changes thereof with time) of one or more points inside (or on the boundary) of the study object as complementary conditions according to thermodynamic equilibrium equations and other known solution conditions in the absence of one or more solution conditions, and the solving process is as shown in fig. 1.
Generally, the inverse problem of heat transfer is divided into five types: the first category is the inverse problem of estimating thermophysical parameters. When a new material or a plurality of materials are used as a heat transfer medium and thermal physical property parameters are lacked, estimating heat conductivity coefficient, specific heat and the like; the second category is the inverse problem of estimating boundary conditions; temperature values at a given boundary, heat flow density values at a given boundary, convective heat transfer coefficients between objects and surrounding fluid at a given boundary, and temperature of the surrounding fluid are the main three boundary conditions for the problem of thermal conduction. When the temperature value, heat flow density, or convective heat transfer coefficient at the boundary is unknown, the value is estimated. The two types of heat conduction inverse problems are the most common and most urgent problems to be solved in engineering practice; the third category is the estimation of heat source terms. The method comprises two types of heat source intensity identification and heat source position identification; the fourth category is the evaluation of initial conditions. An initial temperature field is necessary, if the temperature field in the solution domain at the initial time is unknown, the initial temperature field needs to be determined; the fifth inverse problem is the estimation of the geometry of the object. If the geometric boundaries of the domain are not known, an estimate of the object's geometry is needed.
At present, the method and thought of the inverse problem of heat transfer science mainly focuses on regularization and an improvement method thereof, a gradient optimization method, an intelligent optimization algorithm and the like. Regularization and an improvement method thereof comprise a Tikhonov regularization method and an iterative regularization method with weight, wherein the method adopts a group of solutions of suitable problems which are adjacent to the original unsuitable problems to approach the solution of the original problems, and the stability of the solution is improved by introducing regularization parameters and a stabilization functional. And selecting proper parameters to directly determine whether the Tikhonov regularization method can quickly and effectively obtain inversion results. If the regularization parameter value is too large, the obtained proper problem is not similar to the original problem, and the solution of the original problem cannot be obtained; similarly, if the regularization parameter value is too small, many original problems are inherited to be ill-defined, and an accurate solution of the original problems cannot be obtained. However, in the actual solution, the selection of the regularization parameter has no specific standard, and mostly depends on prior experience, so that the method has certain limitation; the gradient optimization method includes a conjugate gradient method, a damped least square method and a steepest descent method, and such methods search in a space near an original model according to a certain rule and find a correction amount of the model, and then correct the original model according to the correction amount. And repeating the iteration until the solution of the inverse problem reaches the expected precision. The variance between the calculated temperature and the true temperature difference at the measured point is often used as an objective function to iteratively find an optimal solution. The method has high convergence speed, is easy to fall into a local extreme value, has certain requirements on the accuracy of an initial guess value, and has huge calculation amount and obvious consumed time when an objective function is complex (such as inverse problem calculation of a three-dimensional model); the intelligent optimization algorithm is an evolutionary algorithm with global searching performance obtained based on bionics, such as a genetic algorithm, an ant colony algorithm, a particle swarm optimization algorithm and the like. By setting specific constraint conditions and searching in a solution space consisting of solution coefficients, an optimal solution is expected to be found, so that the solution of the inverse problem can reach the expected accuracy. The method has strong global property, but the calculation amount of the algorithm is very large and the convergence speed is slow because the positive problem needs to be continuously calculated when the search is carried out in the solution space.
Disclosure of Invention
Aiming at the problems in the prior art, the embodiment of the invention provides a method and a device for identifying the inverse problem of heat transfer science based on deep learning.
In a first aspect, an embodiment of the present invention provides a method for identifying inverse problems of heat transfer science based on deep learning, including:
acquiring an inner wall surface temperature cloud picture;
inputting the inner wall surface temperature cloud picture into a preset neural network model to obtain a parameter to be corrected corresponding to the inner wall surface temperature cloud picture; wherein the parameter to be corrected comprises a convection heat exchange coefficient, and/or an incoming flow temperature, and/or a heat source intensity; the preset neural network model is as follows: and training the model based on a deep learning algorithm by using the sample inner wall surface temperature cloud picture as input data and using the parameter to be corrected corresponding to the sample inner wall surface temperature cloud picture as output data.
Further, the inner wall surface temperature cloud picture is a multi-channel image.
Further, still include:
correcting the parameter to be corrected by utilizing a convolutional neural network, and establishing mapping between temperature data and the parameter to be corrected; the temperature data is imaging temperature data represented by an inner wall surface temperature cloud chart.
Further, the preset neural network model is: the method comprises the following steps of training a model based on a deep learning algorithm by using a sample inner wall surface temperature cloud picture as input data and a parameter to be corrected corresponding to the sample inner wall surface temperature cloud picture as output data, and specifically comprises the following steps:
determining training samples and testing samples required by deep learning based on the internal wall surface temperature cloud image samples;
determining the generation number of samples based on the training samples and the test samples, selecting parameters to be corrected, giving a parameter sample distribution range, and determining the sample values of the parameters to be corrected; the parameter sample is a training set label of the convolutional neural network;
taking the parameter sample value to be corrected as an input parameter of a preset neural network model;
and performing regression learning on the training set by the deep neural network based on ResNet, and determining parameters to be corrected corresponding to the inner wall surface temperature cloud image samples.
In a second aspect, an embodiment of the present invention provides a device for identifying inverse problems of heat transfer science based on deep learning, including:
the acquisition module is used for acquiring an inner wall surface temperature cloud picture;
the parameter result output module is used for inputting the inner wall surface temperature cloud picture into a preset neural network model to obtain a parameter to be corrected corresponding to the inner wall surface temperature cloud picture; wherein the parameter to be corrected comprises a convection heat exchange coefficient, and/or an incoming flow temperature, and/or a heat source intensity; the preset neural network model is as follows: and training the model based on a deep learning algorithm by using the sample inner wall surface temperature cloud picture as input data and using the parameter to be corrected corresponding to the sample inner wall surface temperature cloud picture as output data.
Further, the internal wall surface temperature cloud picture in the acquisition module is a multi-channel image.
Further, still include:
the correction module is used for correcting the parameter to be corrected by utilizing the convolutional neural network and establishing mapping between the temperature data and the parameter to be corrected; the temperature data is imaging temperature data represented by an inner wall surface temperature cloud chart.
Further, the preset neural network model is: the method comprises the following steps of training a model based on a deep learning algorithm by using a sample inner wall surface temperature cloud picture as input data and a parameter to be corrected corresponding to the sample inner wall surface temperature cloud picture as output data, and specifically comprises the following steps:
determining training samples and testing samples required by deep learning based on the internal wall surface temperature cloud image samples;
determining the generation number of samples based on the training samples and the test samples, selecting parameters to be corrected, giving a parameter sample distribution range, and determining the sample values of the parameters to be corrected; the parameter sample is a training set label of the convolutional neural network;
taking the parameter sample value to be corrected as an input parameter of a preset neural network model;
and performing regression learning on the training set by the deep neural network based on ResNet, and determining parameters to be corrected corresponding to the inner wall surface temperature cloud image samples.
In a third aspect, an embodiment of the present invention further provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and when the processor executes the program, the processor implements the steps of the deep learning-based inverse problem identification method for heat transfer science as described in the first aspect.
In a fourth aspect, the embodiments of the present invention further provide a non-transitory computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the steps of the deep learning-based inverse problem identification method for heat transfer science as described in the first aspect.
According to the technical scheme, the method and the device for identifying the inverse problem of the heat transfer science based on the deep learning provided by the embodiment of the invention have the advantages that the temperature cloud picture of the inner wall surface is obtained; inputting the inner wall surface temperature cloud picture into a preset neural network model to obtain a parameter to be corrected corresponding to the inner wall surface temperature cloud picture; wherein the parameter to be corrected comprises a convection heat exchange coefficient, and/or an incoming flow temperature, and/or a heat source intensity; the embodiment of the invention simplifies the flow of the inverse problem of heat transfer, inputs the imaging temperature data into the trained deep neural network, and the output result is the parameter to be corrected, thereby overcoming the defects of difficulty, discomfort, initial value dependence, easy falling into local convergence and the like of target function selection in the traditional method, and reducing errors caused by calculated amount and operation.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.
FIG. 1 is a schematic diagram of a solution flow provided by the prior art;
FIG. 2 is a schematic flow chart illustrating a method for identifying inverse problems of heat transfer science based on deep learning according to an embodiment of the present invention;
FIG. 3 is a schematic flow chart illustrating a method for identifying inverse problems of heat transfer science based on deep learning according to another embodiment of the present invention;
fig. 4 is a schematic structural diagram of a closed square cavity according to an embodiment of the present invention;
FIG. 5 is a schematic diagram of a finite element model according to an embodiment of the present invention;
FIG. 6 is a temperature cloud provided in accordance with an embodiment of the present invention;
FIG. 7 is a schematic structural diagram of an apparatus for recognizing inverse problem of heat transfer science based on deep learning according to an embodiment of the present invention;
fig. 8 is a schematic physical structure diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some embodiments, but not all embodiments, of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention. The method for recognizing inverse problems of heat transfer science based on deep learning provided by the present invention will be explained and illustrated in detail by specific examples.
FIG. 2 is a schematic flow chart illustrating a method for identifying inverse problems of heat transfer science based on deep learning according to an embodiment of the present invention; as shown in fig. 2, the method includes:
step 101: and obtaining a temperature cloud picture of the inner wall surface.
Step 102: inputting the inner wall surface temperature cloud picture into a preset neural network model to obtain a parameter to be corrected corresponding to the inner wall surface temperature cloud picture; wherein the parameter to be corrected comprises a convection heat exchange coefficient, and/or an incoming flow temperature, and/or a heat source intensity; the preset neural network model is as follows: and training the model based on a deep learning algorithm by using the sample inner wall surface temperature cloud picture as input data and using the parameter to be corrected corresponding to the sample inner wall surface temperature cloud picture as output data.
In this embodiment, it should be noted that the method for identifying an inverse heat transfer problem based on deep learning according to the embodiment of the present invention is directed to research the second and third inverse heat transfer problems, that is, the inversion of the boundary convective heat transfer coefficient, the incoming flow temperature, and the heat source intensity. Taking a three-dimensional closed square cavity structure as an example, a selected temperature cloud chart of the inner wall surface of the square cavity is used as an initial condition, and a deep learning network is used for simultaneously inverting and outputting a convective heat transfer coefficient, an incoming flow temperature and a heat source intensity.
In this embodiment, it should be noted that the method for identifying the inverse problem of heat transfer science based on deep learning provided by the embodiment of the present invention can overcome the defects that in the conventional method for identifying the inverse problem of heat transfer, the difficulty, discomfort and initial dependency of target function selection during multi-parameter inversion and the like are easily caused to fall into local convergence, and extract and process multi-point data, and convert the experimentally measured inner wall temperature data into a multi-channel image. And (3) performing feature extraction on the image by using a deep learning network, establishing rapid mapping of the inner wall temperature data and three parameters to be corrected, namely the convective heat transfer coefficient, the incoming flow temperature and the heat source intensity, and solving the inverse problem.
In order to solve the above-mentioned problems, see fig. 3, for example:
step 1: training samples and testing samples required by deep learning are generated, and the data source is generally an existing experimental database or obtained through a simulation method. Step 2 is performed if the experimental data originates from a simulation experiment, and step 4 is performed directly if the temperature data originates from an existing database.
Step 2: determining the generation number N of samples, selecting parameters to be corrected and giving a distribution range of the parameter samples. The parameter sample is a training set label of the deep neural network, and normalization processing is carried out on the training set label.
And step 3: and (3) taking the sample value of the parameter to be corrected generated in the step (2) as an input parameter of the heat transfer model, and carrying out a heat transfer test on the model.
And 4, step 4: and (4) carrying out pretreatment of extraction of experimental data or simulation result data, selecting the same inner wall surface of the structure, and estimating the maximum value and the minimum value of the temperature of the inner wall surface in all results so as to determine the temperature calibration range. Setting the same temperature marking range and resolution.
And 5: and (4) extracting the temperature cloud picture of the inner wall surface selected in the step (4), wherein the multichannel image is the original input of the deep neural network.
Step 6: and establishing a deep neural network based on ResNet for performing regression learning on the training set. And initializing a network structure. And initializing network parameters. An initial convolution kernel is set to convert the input image obtained in step 5 to a size suitable for computer processing. And setting parameters such as the initial layer number of the network, the number of feature maps of each layer, the model precision threshold value and the like.
And 7: and (5) performing network training by using the training samples obtained in the step (5), forwarding the network to obtain a parameter predicted value, and inputting the parameter predicted value and the parameter label into a loss function until the loss function is converged, so as to finish training.
And 8: inputting the test set into the trained network, verifying the accuracy of the network, and carrying out the next step when the accuracy meets the requirement.
And step 9: and verifying the network training result. And (4) preprocessing the temperature distribution cloud picture measured in the experiment according to the same method in the step 4, and inputting the preprocessed temperature distribution cloud picture into the network trained in the step 8 to obtain an estimated value of the parameter.
Step 10: and further verifying the corrected model by using the temperature data outside the training data range, and finishing correction after meeting the precision requirement, namely outputting a parameter correction result.
According to the technical scheme, the heat transfer science anti-problem identification method based on deep learning provided by the embodiment of the invention comprises the steps of obtaining a temperature cloud picture of the inner wall surface; inputting the inner wall surface temperature cloud picture into a preset neural network model to obtain a parameter to be corrected corresponding to the inner wall surface temperature cloud picture; wherein the parameter to be corrected comprises a convection heat exchange coefficient, and/or an incoming flow temperature, and/or a heat source intensity; the embodiment of the invention simplifies the flow of the inverse problem of heat transfer, inputs the imaging temperature data into the trained deep neural network, and the output result is the parameter to be corrected, thereby overcoming the defects of difficulty, discomfort, initial value dependence, easy falling into local convergence and the like of target function selection in the traditional method, and reducing errors caused by calculated amount and operation.
In addition to the above embodiments, in this embodiment, the inner wall surface temperature cloud map is a multi-channel image.
In this embodiment, it should be noted that, by directly extracting the structural temperature image for processing, incompleteness and errors caused by manually setting and extracting temperature data of multiple measurement points are avoided, so that the data result utilization rate of the structural temperature characteristic is higher, and the capability of representing a real structure is strong.
On the basis of the above embodiment, in this embodiment, the method further includes:
correcting the parameter to be corrected by utilizing a convolutional neural network, and establishing mapping between temperature data and the parameter to be corrected; the temperature data is imaging temperature data represented by an inner wall surface temperature cloud chart.
In this embodiment, it should be noted that, in the embodiment of the present invention, a convolutional neural network model is applied to modify a parameter to be modified in combination with the advantage of deep learning in the field of image recognition, so as to establish a fast mapping between temperature data and the parameter to be modified.
On the basis of the foregoing embodiment, in this embodiment, the preset neural network model is: the method comprises the following steps of training a model based on a deep learning algorithm by using a sample inner wall surface temperature cloud picture as input data and a parameter to be corrected corresponding to the sample inner wall surface temperature cloud picture as output data, and specifically comprises the following steps:
determining training samples and testing samples required by deep learning based on the internal wall surface temperature cloud image samples;
determining the generation number of samples based on the training samples and the test samples, selecting parameters to be corrected, giving a parameter sample distribution range, and determining the sample values of the parameters to be corrected; the parameter sample is a training set label of the convolutional neural network;
taking the parameter sample value to be corrected as an input parameter of a preset neural network model;
and performing regression learning on the training set by the deep neural network based on ResNet, and determining parameters to be corrected corresponding to the inner wall surface temperature cloud image samples.
In this embodiment, it should be noted that the embodiment of the present invention can perform high-precision solution on the multi-parameter inversion problem, and avoid precision loss caused by a calculation process of building a solution model for many times and simplifying calculation in the conventional method.
In order to better understand the present invention, the following examples are further provided to illustrate the content of the present invention, but the present invention is not limited to the following examples.
The invention adopts a certain closed square cavity structure as shown in figure 4, and a finite element model is shown in figure 5. The square cavity is provided with a square heat source in the center at the bottom, the square cavity radiates heat to the inside in a body heat source mode, the inside fluid is air, and the outside of the square cavity is influenced by convection and radiates heat to the outside. In the numerical calculation example of the invention, 3 parameters to be corrected are selected, namely the convective heat transfer coefficient, the incoming flow temperature and the heat source intensity, and the actual values are shown in table 1.
TABLE 1 parameters to be corrected and truth values
Step 1: setting the number of samples to be 5000, generating an initial distribution range of the parameters to be corrected according to the actual working conditions, extracting sample parameters from the range, and carrying out data recording on each group of parameters. Establishing a heat transfer model in finite element software ANSYS, properly dividing grids, inputting other known boundary conditions, inputting each group of parameters, initializing and solving the heat transfer model.
Step 2: and (4) carrying out pretreatment of extraction of experimental data or simulation result data, selecting the same inner wall surface of the structure, and estimating the maximum value and the minimum value of the temperature of the inner wall surface in all results so as to determine the temperature calibration range. Setting the same temperature marking range and resolution.
And step 3: and carrying out data extraction on the calculation result. The calculation example of the invention can directly extract the temperature cloud picture of the inner wall surface of the model, and the temperature cloud picture is stored as an image (see figure 6) with 3 channels 960 x 960 pixels, namely a training sample picture of the deep neural network, and 5000 groups of corrected parameter values are label values of the training sample.
And 4, step 4: due to the presence of unnecessary data in the training set input images, the raw images need to be processed. The specific method is that PythonoOpenCV programming is used for cutting an input image, cutting temperature calibration marks and a model visual field direction schematic diagram, only an inner wall surface temperature cloud picture is reserved, and meanwhile, normalization processing is carried out on labels of a training set.
And 5: and (5) initializing the network. After the network acquires data, the deep neural network is trained under a ResNet deep learning framework. The number of the network layers is 11, wherein the sub-structure in the network comprises a convolution layer, a normalization layer and an activation layer, and the number of the regression parameters is 3. The convolution layer realizes feature extraction, the normalization layer is used for data normalization, and the activation layer adds the nonlinearity of the network.
Step 6: in training, in order to avoid the problems of gradient degradation, detail information loss and the like caused by hyperplasia increase, the upper-layer detail information is subjected to quick connection operation after a plurality of layers of convolution operation, and is superposed with the feature map after the plurality of layers of convolution. If the sizes of the front and rear layers of images are not matched, one-step convolution operation is firstly carried out to unify the sizes of the images and then the images are superposed.
And 7: after feature extraction, the data are converted into 3 output parameters through a full connection layer, and the output parameters are regression results of the deep neural network.
And 8: and (4) building a network according to the network structure in the step (7), and calculating a loss function and a root mean square error (rmse) according to a formula 1. Wherein the loss function is the Euclidean distance between the network regression data and the training label, and the rmse calculation method is shown inEquation (1). Wherein y isiIn order to be a true sample,
for prediction values, the network is trained using a stochastic gradient descent method until convergence. The convergence errors of the samples on the training data set and the test data set are calculated separately.
And step 9: and testing the trained network on a test set outside the training samples. If overfitting occurs, sample data is adjusted or network parameters are modified until the accuracy requirement is met.
Step 10: and taking the temperature cloud pictures (three-channel images with the same size as the sample images) of the results of the true values of the three parameters to be corrected as input images of the test set, wherein the corresponding true values of the parameters to be corrected are the labels of the test samples. And comparing the network output result with the test sample label to obtain a corrected result set error.
Step 11: and storing the deep neural network and outputting a result.
Fig. 7 is a schematic structural diagram of an apparatus for identifying inverse problem of heat transfer science based on deep learning according to an embodiment of the present invention, as shown in fig. 7, the apparatus includes: an obtaining module 201 and a parameter result output module 202, wherein:
the acquiring module 201 is used for acquiring an inner wall surface temperature cloud chart;
a parameter result output module 202, configured to input the inner wall surface temperature cloud map into a preset neural network model, so as to obtain a parameter to be corrected corresponding to the inner wall surface temperature cloud map; wherein the parameter to be corrected comprises a convection heat exchange coefficient, and/or an incoming flow temperature, and/or a heat source intensity; the preset neural network model is as follows: and training the model based on a deep learning algorithm by using the sample inner wall surface temperature cloud picture as input data and using the parameter to be corrected corresponding to the sample inner wall surface temperature cloud picture as output data.
The device for identifying the inverse problem of heat transfer science based on deep learning provided by the embodiment of the invention can be specifically used for executing the method for identifying the inverse problem of heat transfer science based on deep learning of the embodiment, the technical principle and the beneficial effect are similar, and specific reference can be made to the embodiment, and the detailed description is omitted here.
Based on the same inventive concept, an embodiment of the present invention provides an electronic device, and referring to fig. 8, the electronic device specifically includes the following contents: a processor 301, a communication interface 303, a memory 302, and a communication bus 304;
the processor 301, the communication interface 303 and the memory 302 complete mutual communication through the communication bus 304; the communication interface 303 is used for realizing information transmission between related devices such as modeling software, an intelligent manufacturing equipment module library and the like; the processor 301 is used for calling the computer program in the memory 302, and the processor executes the computer program to implement the method provided by the above method embodiments, for example, the processor executes the computer program to implement the following steps: acquiring an inner wall surface temperature cloud picture; inputting the inner wall surface temperature cloud picture into a preset neural network model to obtain a parameter to be corrected corresponding to the inner wall surface temperature cloud picture; wherein the parameter to be corrected comprises a convection heat exchange coefficient, and/or an incoming flow temperature, and/or a heat source intensity; the preset neural network model is as follows: and training the model based on a deep learning algorithm by using the sample inner wall surface temperature cloud picture as input data and using the parameter to be corrected corresponding to the sample inner wall surface temperature cloud picture as output data.
Based on the same inventive concept, another embodiment of the present invention further provides a non-transitory computer-readable storage medium, on which a computer program is stored, the computer program being implemented by a processor to perform the methods provided by the above method embodiments, for example, obtaining a temperature cloud map of an inner wall surface; inputting the inner wall surface temperature cloud picture into a preset neural network model to obtain a parameter to be corrected corresponding to the inner wall surface temperature cloud picture; wherein the parameter to be corrected comprises a convection heat exchange coefficient, and/or an incoming flow temperature, and/or a heat source intensity; the preset neural network model is as follows: and training the model based on a deep learning algorithm by using the sample inner wall surface temperature cloud picture as input data and using the parameter to be corrected corresponding to the sample inner wall surface temperature cloud picture as output data.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and the parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium, such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods of the various embodiments or some parts of the embodiments.
In addition, in the present invention, terms such as "first" and "second" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of the present invention, "a plurality" means at least two, e.g., two, three, etc., unless specifically limited otherwise.
Moreover, in the present invention, relational terms such as first and second, and the like may be 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. Also, 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. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
Furthermore, in the description herein, reference to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above are not necessarily intended to refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, various embodiments or examples and features of different embodiments or examples described in this specification can be combined and combined by one skilled in the art without contradiction.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present 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 solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. A heat transfer inverse problem identification method based on deep learning is characterized by comprising the following steps:
acquiring an inner wall surface temperature cloud picture;
inputting the inner wall surface temperature cloud picture into a preset neural network model to obtain a parameter to be corrected corresponding to the inner wall surface temperature cloud picture; wherein the parameter to be corrected comprises a convection heat exchange coefficient, and/or an incoming flow temperature, and/or a heat source intensity; the preset neural network model is as follows: and training the model based on a deep learning algorithm by using the sample inner wall surface temperature cloud picture as input data and using the parameter to be corrected corresponding to the sample inner wall surface temperature cloud picture as output data.
2. The deep learning-based inverse problem identification method of heat transfer science according to claim 1, wherein the inner wall surface temperature cloud map is a multichannel image.
3. The deep learning-based inverse problem identification method of heat transfer science according to claim 1, further comprising:
correcting the parameter to be corrected by utilizing a convolutional neural network, and establishing mapping between temperature data and the parameter to be corrected; the temperature data is imaging temperature data represented by an inner wall surface temperature cloud chart.
4. The deep learning-based inverse problem identification method of heat transfer science according to claim 3, wherein the preset neural network model is: the method comprises the following steps of training a model based on a deep learning algorithm by using a sample inner wall surface temperature cloud picture as input data and a parameter to be corrected corresponding to the sample inner wall surface temperature cloud picture as output data, and specifically comprises the following steps:
determining training samples and testing samples required by deep learning based on the internal wall surface temperature cloud image samples;
determining the generation number of samples based on the training samples and the test samples, selecting parameters to be corrected, giving a parameter sample distribution range, and determining the sample values of the parameters to be corrected; the parameter sample is a training set label of the convolutional neural network;
taking the parameter sample value to be corrected as an input parameter of a preset neural network model;
and performing regression learning on the training set by the deep neural network based on ResNet, and determining parameters to be corrected corresponding to the inner wall surface temperature cloud image samples.
5. A device for recognizing inverse problems of heat transfer science based on deep learning, comprising:
the acquisition module is used for acquiring an inner wall surface temperature cloud picture;
the parameter result output module is used for inputting the inner wall surface temperature cloud picture into a preset neural network model to obtain a parameter to be corrected corresponding to the inner wall surface temperature cloud picture; wherein the parameter to be corrected comprises a convection heat exchange coefficient, and/or an incoming flow temperature, and/or a heat source intensity; the preset neural network model is as follows: and training the model based on a deep learning algorithm by using the sample inner wall surface temperature cloud picture as input data and using the parameter to be corrected corresponding to the sample inner wall surface temperature cloud picture as output data.
6. The device for identifying inverse problems of heat transfer science based on deep learning of claim 5 is characterized in that the temperature cloud map of the inner wall surface in the acquisition module is a multi-channel image.
7. The deep learning-based inverse problem identification device for heat transfer science according to claim 5, further comprising:
the correction module is used for correcting the parameter to be corrected by utilizing the convolutional neural network and establishing mapping between the temperature data and the parameter to be corrected; the temperature data is imaging temperature data represented by an inner wall surface temperature cloud chart.
8. The deep learning-based inverse problem identification device of heat transfer science according to claim 7, wherein the preset neural network model is: the method comprises the following steps of training a model based on a deep learning algorithm by using a sample inner wall surface temperature cloud picture as input data and a parameter to be corrected corresponding to the sample inner wall surface temperature cloud picture as output data, and specifically comprises the following steps:
determining training samples and testing samples required by deep learning based on the internal wall surface temperature cloud image samples;
determining the generation number of samples based on the training samples and the test samples, selecting parameters to be corrected, giving a parameter sample distribution range, and determining the sample values of the parameters to be corrected; the parameter sample is a training set label of the convolutional neural network;
taking the parameter sample value to be corrected as an input parameter of a preset neural network model;
and performing regression learning on the training set by the deep neural network based on ResNet, and determining parameters to be corrected corresponding to the inner wall surface temperature cloud image samples.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the deep learning based inverse problem identification method of heat transfer science as claimed in any one of claims 1 to 4 when executing the program.
10. A non-transitory computer readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the deep learning based inverse problem identification method of heat transfer science according to any one of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110459377.3A CN113255181A (en) | 2021-04-27 | 2021-04-27 | Heat transfer inverse problem identification method and device based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110459377.3A CN113255181A (en) | 2021-04-27 | 2021-04-27 | Heat transfer inverse problem identification method and device based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113255181A true CN113255181A (en) | 2021-08-13 |
Family
ID=77221776
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110459377.3A Pending CN113255181A (en) | 2021-04-27 | 2021-04-27 | Heat transfer inverse problem identification method and device based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113255181A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113947002A (en) * | 2021-12-17 | 2022-01-18 | 广东粤港澳大湾区硬科技创新研究院 | Method for calculating convection heat transfer coefficient of spectrum beam combination semiconductor laser |
| CN114841023A (en) * | 2022-07-04 | 2022-08-02 | 中山大学 | Thermal data determination method, device and equipment |
| GB2623404A (en) * | 2022-07-04 | 2024-04-17 | Univ Sun Yat Sen | Thermal data determination method, apparatus and device |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060145680A1 (en) * | 2004-12-31 | 2006-07-06 | Tatung Co., Ltd. | Method for analyzing the inner temperature field and flow field in power transformers |
| CN108959794A (en) * | 2018-07-13 | 2018-12-07 | 北京航空航天大学 | A kind of structural frequency response modification methodology of dynamics model based on deep learning |
| CN110083895A (en) * | 2019-04-12 | 2019-08-02 | 浙江大学 | A kind of surface heat flux three-dismensional effect modification method neural network based |
| CN112092676A (en) * | 2020-09-25 | 2020-12-18 | 吉林大学 | Estimation and correction method for battery pack temperature field by using virtual temperature sensor |
-
2021
- 2021-04-27 CN CN202110459377.3A patent/CN113255181A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060145680A1 (en) * | 2004-12-31 | 2006-07-06 | Tatung Co., Ltd. | Method for analyzing the inner temperature field and flow field in power transformers |
| CN108959794A (en) * | 2018-07-13 | 2018-12-07 | 北京航空航天大学 | A kind of structural frequency response modification methodology of dynamics model based on deep learning |
| CN110083895A (en) * | 2019-04-12 | 2019-08-02 | 浙江大学 | A kind of surface heat flux three-dismensional effect modification method neural network based |
| CN112092676A (en) * | 2020-09-25 | 2020-12-18 | 吉林大学 | Estimation and correction method for battery pack temperature field by using virtual temperature sensor |
Non-Patent Citations (3)
| Title |
|---|
| 寇蔚等: "基于随机有限元与神经网络的传热参数的智能辨识", 《红外技术》 * |
| 张璐等: "混杂纤维混凝土热工参数反分析研究", 《混凝土》 * |
| 苏超等: "基于BP神经网络的船闸闸首热学参数反分析", 《南水北调与水利科技》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113947002A (en) * | 2021-12-17 | 2022-01-18 | 广东粤港澳大湾区硬科技创新研究院 | Method for calculating convection heat transfer coefficient of spectrum beam combination semiconductor laser |
| CN113947002B (en) * | 2021-12-17 | 2022-03-29 | 广东粤港澳大湾区硬科技创新研究院 | Method for calculating convection heat transfer coefficient of spectrum beam combination semiconductor laser |
| CN114841023A (en) * | 2022-07-04 | 2022-08-02 | 中山大学 | Thermal data determination method, device and equipment |
| CN114841023B (en) * | 2022-07-04 | 2022-09-09 | 中山大学 | Thermal data determination method, device and equipment |
| WO2024007474A1 (en) * | 2022-07-04 | 2024-01-11 | 中山大学 | Thermal data determination method, apparatus and device |
| GB2623404A (en) * | 2022-07-04 | 2024-04-17 | Univ Sun Yat Sen | Thermal data determination method, apparatus and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113255181A (en) | Heat transfer inverse problem identification method and device based on deep learning | |
| Shi et al. | Improved Iterative Closest Point (ICP) 3D point cloud registration algorithm based on point cloud filtering and adaptive fireworks for coarse registration | |
| WO2018227800A1 (en) | Neural network training method and device | |
| CN111310846A (en) | Method, device, storage medium and server for selecting sample image | |
| BR112015029806A2 (en) | systems and methods for performing Bayesian optimization | |
| CN110276363A (en) | A kind of birds small target detecting method based on density map estimation | |
| CN113218537B (en) | Training method, training device, training equipment and training storage medium for temperature anomaly detection model | |
| CN113065525A (en) | Age recognition model training method, face age recognition method and related device | |
| CN111488498A (en) | Node-graph cross-layer graph matching method and system based on graph neural network | |
| CN110503148B (en) | Point cloud object identification method with scale invariance | |
| CN109902641B (en) | Semantic alignment-based face key point detection method, system and device | |
| Ding et al. | Resolution of unknown heat source inverse heat conduction problems using particle swarm optimization | |
| CN111354076A (en) | Single-image three-dimensional part combined modeling method based on embedding space | |
| Belyaev et al. | Building data fusion surrogate models for spacecraft aerodynamic problems with incomplete factorial design of experiments | |
| CN111062406B (en) | Heterogeneous domain adaptation-oriented semi-supervised optimal transmission method | |
| Patel et al. | Deep Learning-Based Plant Organ Segmentation and Phenotyping of Sorghum Plants Using LiDAR Point Cloud | |
| CN115690546B (en) | Shrimp length measuring method, device, electronic equipment and storage medium | |
| CN109829939B (en) | Method for narrowing search range of multi-view image matching same-name image points | |
| CN117036781A (en) | Image classification method based on tree comprehensive diversity depth forests | |
| CN116402828A (en) | Battery piece defect detection method and system based on saliency map | |
| CN114781207B (en) | Heat source layout temperature field prediction method based on uncertainty and semi-supervised learning | |
| CN113553708B (en) | Method and device for tracing key influence factors of simulation model | |
| CN108205805B (en) | Automatic establishment method for dense correspondence of cone beam CT image intermediate pixels | |
| Yu et al. | Defect identification in heat transfer problems using boundary data | |
| CN115062551A (en) | Wet physical process parameterization method based on time sequence neural network |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210813 |