CN114970240A - Method and equipment for rapidly evaluating load state of multi-phase composite structure image - Google Patents
Method and equipment for rapidly evaluating load state of multi-phase composite structure image Download PDFInfo
- Publication number
- CN114970240A CN114970240A CN202210471459.4A CN202210471459A CN114970240A CN 114970240 A CN114970240 A CN 114970240A CN 202210471459 A CN202210471459 A CN 202210471459A CN 114970240 A CN114970240 A CN 114970240A
- Authority
- CN
- China
- Prior art keywords
- strain
- image
- distribution
- composite structure
- stress
- 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
- 239000002131 composite material Substances 0.000 title claims abstract description 68
- 238000000034 method Methods 0.000 title claims abstract description 59
- 238000009826 distribution Methods 0.000 claims abstract description 97
- 238000012549 training Methods 0.000 claims abstract description 38
- 238000004458 analytical method Methods 0.000 claims abstract description 16
- 230000011218 segmentation Effects 0.000 claims description 20
- 238000005457 optimization Methods 0.000 claims description 18
- 239000000463 material Substances 0.000 claims description 17
- 239000011159 matrix material Substances 0.000 claims description 16
- 238000004364 calculation method Methods 0.000 claims description 15
- 230000008569 process Effects 0.000 claims description 12
- 230000005483 Hooke's law Effects 0.000 claims description 10
- 238000011068 loading method Methods 0.000 claims description 9
- 238000013528 artificial neural network Methods 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 8
- 230000006870 function Effects 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 8
- 238000001816 cooling Methods 0.000 claims description 7
- 238000007781 pre-processing Methods 0.000 claims description 5
- 238000012935 Averaging Methods 0.000 claims description 4
- 239000003086 colorant Substances 0.000 claims description 4
- 238000001914 filtration Methods 0.000 claims description 4
- 238000003491 array Methods 0.000 claims description 3
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 238000012937 correction Methods 0.000 claims description 3
- 230000002708 enhancing effect Effects 0.000 claims description 3
- 230000002194 synthesizing effect Effects 0.000 claims description 3
- 238000010200 validation analysis Methods 0.000 claims description 3
- 238000011156 evaluation Methods 0.000 description 14
- 238000004088 simulation Methods 0.000 description 13
- 238000000576 coating method Methods 0.000 description 10
- 239000011248 coating agent Substances 0.000 description 9
- 239000000919 ceramic Substances 0.000 description 7
- 238000003860 storage Methods 0.000 description 7
- 238000001878 scanning electron micrograph Methods 0.000 description 5
- 238000004891 communication Methods 0.000 description 4
- 239000012720 thermal barrier coating Substances 0.000 description 4
- 230000003287 optical effect Effects 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 229910004298 SiO 2 Inorganic materials 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000005520 cutting process Methods 0.000 description 2
- 238000007405 data analysis Methods 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 239000011148 porous material Substances 0.000 description 2
- 238000002360 preparation method Methods 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 230000003042 antagnostic effect Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000013075 data extraction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000009413 insulation Methods 0.000 description 1
- 238000001000 micrograph Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000013433 optimization analysis Methods 0.000 description 1
- 238000007750 plasma spraying Methods 0.000 description 1
- 238000012805 post-processing Methods 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 230000003252 repetitive effect Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000005507 spraying Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 239000000758 substrate Substances 0.000 description 1
- 238000010998 test method Methods 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
- G06F30/27—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
- G06N3/04—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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/70—Labelling scene content, e.g. deriving syntactic or semantic representations
-
- 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)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Medical Informatics (AREA)
- Computational Linguistics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Molecular Biology (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Investigating Strength Of Materials By Application Of Mechanical Stress (AREA)
Abstract
The invention relates to a method and equipment for rapidly evaluating the load state of a multiphase composite structure image, wherein the method comprises the following steps: acquiring a section morphology image of the multiphase composite structure, establishing an image restoration finite element analysis model, and calculating load distribution under different boundary conditions and geometric characteristics; constructing a generation countermeasure network constrained by an elastic mechanical model, and training based on data acquired by a finite element analysis model; and evaluating the load state of the multiphase composite structure to be detected based on the trained generated countermeasure network. Compared with the prior art, the method has the advantages of good prediction precision, high efficiency and the like.
Description
Technical Field
The invention relates to the field of load state evaluation, in particular to a method and equipment for rapidly evaluating a load state of a multi-phase composite structure image.
Background
Multiphase composite structures are widely used in engineering applications, and the difference in performance of each phase generally causes structural failure when boundary conditions are drastically changed. For example, thermal barrier coatings, which are widely used in aerospace high temperature components, are typically prepared by plasma spray coating (APS), resulting in a complex and varied pore structure in the ceramic top layer (TC). The TC internal pore structure is easy to be infiltrated by molten CaO-MgO-Al 2O 3-SiO 2(CMAS) during high-temperature service while improving the heat insulation performance to form a TC-CMAS two-phase structure. The coefficient of thermal expansion of the CMAS is greatly different from that of the ceramic matrix, and during cooling, the solidified CMAS can induce the TC layer to generate higher thermal mismatch stress, so that the coating fails in advance. Therefore, a large number of scholars develop numerical simulation aiming at TC-CMAS thermal cycle load, and effective progress is made. However, these methods have a great challenge in that when the boundary conditions or microstructure characteristics change, modeling needs to be performed again and calculation needs to be performed, which consumes a great deal of computing resources. The optimization of the characteristics of the composite structure is an important technical approach for improving the performance of the multi-phase material, and the low efficiency of the traditional numerical calculation method becomes an obstacle for restricting the rapid optimization design of the composite structure. Along with the development of artificial intelligence technology, a data-driven neural network model has higher calculation speed. Therefore, the load rapid evaluation method based on the neural network architecture is developed, and has important significance for the process and composite structure optimization of the multiphase material.
At present, the evaluation techniques aiming at the load states of multi-phase structure stress, strain and the like are mainly divided into two types:
the first is a test based on a composite structure sample, and generally, a strain field of the sample is measured by using a DIC method under a simulation boundary condition. However, the device resolution is limited, it is difficult to capture the detail change of the strain distribution of the composite microstructure, and the sample preparation and test period is long, and the test cost is high. Therefore, the test method cannot meet the requirements of multi-phase composite structure optimization and load state online monitoring.
The second method is numerical simulation by means of relevant finite element software, and the method obtains the thermal strain state of the multiphase composite structure through numerical simulation analysis by means of computer simulation software. However, numerical simulation of complex structures places extremely high demands on computer hardware; the composite structure is generally complex and changeable, a geometric model needs to be re-established every time a new structure is faced in numerical calculation, a large amount of work is repeated, a large amount of time and calculation resources are consumed, and the requirement of optimization analysis of the composite structure is difficult to support; in addition, part of work is also carried out to simplify a composite structure into a regular shape, and a simplified geometric model simulating inclusion rate and other structural parameters is established for numerical simulation.
As shown in fig. 1, taking a conventional stress and strain state evaluation method based on numerical simulation as an example, it is first necessary to photograph or simplify the structural morphology from a multiphase composite structure sample. After semantic segmentation is carried out on the structure, a numerical simulation model of the structure is established by combining service parameters such as boundary conditions, material attributes and the like. The numerical simulation process includes the steps of establishing an image restoration geometric model, setting material properties, dividing grids, setting boundary conditions, submitting calculation, post-processing results and the like, and generally consumes a large amount of time and calculation resources. For the task of obtaining a data set of a certain scale by solving a plurality of groups of data, all the processes need to be repeated for a new composite structure to be tested, and particularly the time-consuming numerical simulation process is included. Therefore, the traditional numerical simulation method is difficult to acquire more data in a short time to support the work of the composite structure strength, the technological parameter optimization and the like.
Disclosure of Invention
The present invention is directed to a method, an apparatus, and a medium for fast load state estimation of a multi-phase composite structure image, which overcome the above-mentioned drawbacks of the prior art.
The purpose of the invention can be realized by the following technical scheme:
according to a first aspect of the present invention, there is provided a method for rapidly evaluating a load state of a multi-phase composite structure image, comprising the steps of:
acquiring a profile image of the multiphase composite structure, establishing an image restoration finite element analysis model, and calculating load distribution under different boundary conditions and geometric characteristics;
constructing a generation countermeasure network constrained by an elastic mechanical model, and training based on data acquired by a finite element analysis model;
and evaluating the load state of the multiphase composite structure to be detected based on the trained generated countermeasure network.
Preferably, the method specifically comprises the following steps:
s1: acquiring a section morphology image of a multiphase composite structure, and constructing an image library;
s2: performing semantic segmentation on images in an image library according to different phases to obtain a semantic segmentation structure chart;
s3: performing boundary identification and anti-aliasing optimization on the semantically segmented image;
s4: establishing an image restoration finite element model based on the optimized structure boundary coordinates, and calculating the load distribution of the structure under the preset working condition parameters;
s5: establishing an elastic mechanical constraint generation confrontation network model for fitting a structural image and load distribution relation;
s6: constructing a training sample set by taking the semantic segmentation structure chart obtained in the step S2 as model input and the load distribution obtained in the step S4 as model output;
s7: training the generated confrontation network model based on the training sample set;
s8: and acquiring a multi-phase structural profile image to be detected, preprocessing the multi-phase structural profile image based on S2, and sending the multi-phase structural profile image to a trained generated confrontation network model to acquire a load state.
Further preferably, the step S2 specifically includes:
s21: processing the profile image by Gaussian filtering;
s22: adjusting the contrast of the image and enhancing the gray difference of each phase;
s23: the different phase components are marked with different colors.
Further preferably, the step S3 specifically includes:
s31: identifying each phase component area in the semantic segmentation structure chart in the step S2 based on a connected component analysis method;
s32: and performing antialiasing optimization on the identified structure boundary by using a coordinate averaging method.
Further preferably, the step S4 specifically includes:
carrying out finite element analysis on the structure boundary optimized in the step S3, and establishing a geometric model of the two-dimensional multiphase structure; setting corresponding material properties for each phase component, a base material and the like; dividing the geometric model by adopting quadrilateral meshes, wherein the sizes of the meshes of different groups of structures are the same; and setting boundary conditions according to preset working conditions, and submitting and calculating the load distribution of the multiple groups of structures under the nonlinear condition.
Further preferably, the step S5 specifically includes:
establishing an initial generation countermeasure network for pixel-to-pixel conversion based on a pix2pix architecture: a multi-layer U-shaped convolution neural network is established as a generator G for generating a countermeasure network, and a downsampling convolution neural network is established as a discriminator D for generating the countermeasure network.
Further preferably, the load distribution is a stress distribution and a strain field of the multiphase composite structure, the generation of the antagonistic network is trained by using the stress and the strain of the multiphase composite structure as outputs, and the step S5 specifically includes:
initially the loss function to generate the countermeasure network is noted as:
wherein G is trained to maximize logD (r, G (r)) loss and D is trained to minimize logD (r, s), E e Representing the mathematical expectation, λ is the L1 norm loss L 1 R is the input of the network, s is the true reference of the output, L 1 (G) Is recorded as:
L 1 (G)＝E r,s [||s-G(r)|| 1 ]
when the predicted target is a stress distribution, s is the finite element stress field σ t (ii) a When the predicted target is a strain distribution, s is the finite element strain field ε t ，
A generalized hooke's law is introduced that considers thermal expansion, written as:
wherein E ═ E/(1-v) 2 ) ν/(1- ν), α ═ α (1+ ν), G ═ E/2(1+ ν), E is the elastic modulus, and ∈ is 11 Is an elastic strain component in the x-axis direction, ε 22 Is the strain component in the y-axis direction, ε 12 Is the shear strain in the x-y plane, v represents the Poisson's ratio, α is the coefficient of thermal expansion, σ represents the stress component in each axial direction, σ 11 As a stress component in the direction of the x-axis, σ 22 Stress component in the y-axis direction, σ 12 Is the shear stress in the x-y plane, T r Represents room temperature, T 0 Represents the initial temperature of cooling;
each stress component is expressed as a strain component, and is expressed as:
the deformation coordination conditions for introducing the assumption of elastomechanics continuity are recorded as:
let the multi-phase structure distribution after semantic segmentation in step S2 be x c The corresponding strain distribution calculated based on the finite element model in step S4 is set to ε t Combining the three strain components into a three-dimensional array, which is recorded as:
wherein, # denotes the switching layer and denotes ε t Is composed of epsilon t,11 ,ε t,22 ,
Three-dimensional arrays stacked in sequence;
generator generated strain array
Is recorded as:
ε m ＝[ε m,11 #ε m,22 #ε m,12 ]
ε t and epsilon m All can be equivalent to a multi-channel picture A, a convolution kernel is set as a multi-dimensional array C, and the two-dimensional convolution calculation of three channels can be expressed as:
where ζ and ξ are the receptor field positions, k is the number of channels,. indicates the convolution,
calculating the strain coordination error of the strain image by adopting a difference method, wherein any pixel point P ζ,ξ The second order partial derivatives in each direction are noted as:
the deformation coordination condition for the elasto-mechanical continuity assumption can be expressed as:
introduction of custom fixed convolution kernel C p And is recorded as:
using convolutionStrain coordination error L of strain field of calculation of convenient operation C And is recorded as:
wherein f is a pixel side length correction coefficient,
when the predicted target is strain distribution, the stress load distribution calculated by finite element is set as a real reference, and the stress-strain relation error L is calculated based on the stress component H,12 And L H,3 And is recorded as:
wherein L is H,12 And L H,3 For constraining the axial and shear strain components, L, respectively C ，L H,12 And L H,3 While the weight optimization for the constraint generator,
discriminator D loss L of true strain load distribution S,D And is recorded as:
L S,D ＝logD(r c ,ε t ,L C,t )
wherein r is c For semantically segmented structural images,. epsilon t For true strain distribution corresponding to the structure, L C,t For strain coordination errors of the true strain distribution,
the strain distribution generated by the generator G, the discriminator loss L thereof SG,D And is recorded as:
L SG,D ＝log(1-D(r c ,G(r c )),L C,m )
wherein L is C,m Is the strain coordination error of the composite strain distribution;
discriminator loss L D Can be written as:
where b is the sample batch size,
when the predicted target is stress distribution, the strain distribution calculated by finite element is set as a real reference, and the stress-strain relation error L is calculated based on the generalized Hooke's law considering thermal expansion H,12 And L H,3 And is recorded as:
wherein L is H,12 And L H,3 For constraining the axial and shear stress components, L, respectively C ，L H,12 And L H,3 While the weight optimization for the constraint generator,
discriminator D loss L of true strain load distribution S,D And is recorded as:
L S,D ＝log D(r c ,σ t ,L C,t )
stress distribution generated by generator G, and discriminator loss L thereof SG,D And is recorded as:
L SG,D ＝log(1-D(r c ,G(r c )),L C,m )
wherein L is C,m Synthesizing strain coordination errors of strain distribution, calculating the generated stress distribution into the strain distribution according to a generalized Hooke's law considering thermal expansion, and then performing convolution calculation based on the strain coordination errors; discriminator loss L D Can be written as:
generator loss L G Including L1 norm L 1 And elastomechanically constrained L C ，L S,12 And L S,3 And is recorded as:
therefore, the generation of the elastic-mechanical constraint opposes the net total loss function, noted as:
wherein,
is a loss of elastodynamics L E Coefficient matrix of, L E Is recorded as:
L E ＝[L C,m ,L S,12 ,L S,3 ] T 。
further preferably, the step S6 specifically includes:
the semantic segmentation structure chart and the corresponding load distribution are segmented into a plurality of groups of local pictures meeting the size condition of the generated confrontation network input picture established in the step S5, and a training sample set is established; establishing a corresponding material attribute and boundary condition matrix according to the structural picture for calculating the elastic mechanical loss in the training process; approximately 70% of the samples were selected for training of the network, and the remaining samples were used for validation.
Further preferably, the step S8 specifically includes:
preprocessing the high-resolution multiphase composite structure picture to be detected according to the step S2, and adjusting the image scale to be the same as the size of the image in the training set; dividing the preprocessed structure into a plurality of groups of local pictures meeting the size condition of the generated confrontation network input picture established in the step S5, and sequentially predicting the corresponding strain distribution by adopting the network model trained in the step S7; and combining the predicted multiple groups of local strain distributions, and removing the overlapped area to obtain complete strain distribution.
According to a second aspect of the present invention, there is provided an electronic device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the method for fast estimating the loading state of the multi-phase composite structure image when executing the program.
According to a third aspect of the present invention, there is provided a computer-readable storage medium, on which a computer program is stored, which when executed by a processor, implements the method for fast load state estimation of a multiphase composite structure image.
Compared with the prior art, the invention has the following advantages:
1. according to the method, a stress, strain and other load evaluation method of the multi-phase composite structure is established on the basis of the elastic mechanics equation constraint generation countermeasure network, the evaluation speed of the trained network fitting model on load distribution is much higher than that of a numerical simulation method, the prediction precision meets the engineering requirements, a rapid data analysis method can be provided for strength evaluation of the multi-phase composite structure, and the design period is shortened;
2. the method combines physical constraints to generate the confrontation network and the picture segmentation method, can calculate sample data based on small-scale numerical values to establish a training data set, reduces the data dependence of the neural network, enhances the generalization capability of a network model, and is convenient for engineering application.
Drawings
FIG. 1 is a flow chart of a multi-phase composite structure strain state evaluation based on numerical simulation.
Fig. 2 is an overall architecture diagram of the method for rapidly estimating the load state of the multiphase composite structure image according to the invention.
FIG. 3 is a flow chart of the rapid evaluation of the thermal strain state of the thermal barrier coating based on the microstructure image according to the present invention.
Detailed Description
The invention is described in detail below with reference to the figures and specific embodiments. Note that the following description of the embodiments is merely a substantial example, and the present invention is not intended to be limited to the application or the use thereof, and is not limited to the following embodiments.
Examples
The invention aims to provide a load state rapid evaluation method of a multiphase composite structure image, which is used for stress or strain load state evaluation of a complex multiphase composite structure, and comprises the following steps: acquiring a profile image of the multiphase composite structure, establishing an image restoration finite element analysis model, and calculating load distribution under different boundary conditions and geometric characteristics; constructing a generation countermeasure network constrained by an elastic mechanical model, and training based on data acquired by a finite element analysis model; and evaluating the load state of the multiphase composite structure to be detected based on the trained generated countermeasure network. The invention is realized by the following technical steps:
step 1, shooting a section morphology image of a multiphase composite structure: preparing a sample of the multiphase material to be detected or sampling, shooting a microstructure SEM image of the sample by using an electron microscope, determining the scale of the image, and constructing an image library;
step 2, carrying out semantic segmentation on the structural image acquired in the step 1 according to different phases: adopting Gaussian filtering to process the SEM image shot in the step 1, and reducing the noise level of the SEM image; properly adjusting the contrast and enhancing the gray difference of each phase; and marking each phase component by adopting different colors to obtain a semantic segmentation structure chart.
And 3, carrying out boundary identification on the composite structure segmented in the step 2, and carrying out anti-aliasing optimization on the coordinates by adopting a coordinate averaging method: identifying the region of each phase component in the semantically segmented picture in the step 2 based on a connected domain analysis method, and recording the region of each phase as:
Q area ＝{(Q 1 ,Q 2 ,L,Q i )|1≤i≤Num(Q area )}
wherein Q is area For a collection of phase regions, Q i For one of the phase regions, i denotes a region number, Num (Q) area ) Representing the number of component regions;
the boundary coordinates of the region are denoted as B C ：
j represents the number of the boundary pixel point, B i Is a region boundary, b p,ij Is a boundary B i (x) of (c) ij ,y ij ) Is b is p,ij The coordinates of (a);
and (3) carrying out anti-aliasing optimization on the identified composite structure boundary by adopting a Coordinate Averaging (CA) method, wherein an expression of a CA algorithm is recorded as:
P x and P y Respectively representing the optimized x and y coordinates, wherein n is the recursion times; wherein, there is b ij To b i(j+n) A total of n +1 points participate in CA optimization.
Step 4, establishing an image recovery finite element model based on the boundary coordinates of the composite structure optimized in the step 3, and calculating the stress and strain distribution of the structure under the preset working condition parameters: importing the structure boundary optimized in the step 3 into finite element analysis software, and establishing a geometric model of the two-dimensional multi-phase structure; setting corresponding material properties for each phase component, a base material and the like; dividing the geometric model by adopting quadrilateral meshes, wherein the sizes of the meshes of different groups of structures are the same; and setting boundary conditions according to preset working conditions, and submitting and calculating the load distribution of the multiple groups of structures under the nonlinear condition. In this embodiment, the load distribution is stress distribution or strain distribution of the multiphase composite structure.
Step 5, establishing an elastic mechanical constraint generation confrontation network model for fitting the structural image and strain distribution relation: establishing an initial generation countermeasure network for pixel-to-pixel conversion based on a pix2pix architecture: establishing a multilayer U-shaped convolution neural network as a generator G for generating a countermeasure network, and establishing a down-sampling convolution neural network as a discriminator D for generating the countermeasure network; initially generating a loss function against the network, noted:
wherein G is trained to maximize logD (r, G (r)) loss and D is trained to minimize logD (r, s), E e Representing the mathematical expectation, λ is the L1 norm loss L 1 R is the input of the network, s is the true reference of the output, L 1 (G) Is recorded as:
L 1 (G)＝E r,s [||s-G(r)|| 1 ] (2)
when the predicted target is a stress distribution, y is a finite element stress distribution σ t (ii) a When the predicted target is a strain field, y is the finite element strain field ε t 。
The stress and the strain of the multi-phase composite structure meet the generalized Hooke's law considering thermal expansion; in addition, the structure is kept continuous all the time in the cooling process, and each strain component of the structure meets the strain coordination condition.
Therefore, a generalized hooke's law is introduced that considers thermal expansion, written as:
wherein E ═ E/(1-v) 2 ) ν '═ ν/(1- ν), α' ═ α (1+ ν), G ═ E/2(1+ ν), E is the elastic modulus, epsilon 11 Is the elastic strain component in the x-axis direction, ε 22 Is the strain component in the y-axis direction, ε 12 Is the shear strain in the x-y plane, v represents the Poisson's ratio, α is the coefficient of thermal expansion, σ represents the stress component in each axial direction, σ 11 As a stress component in the direction of the x-axis, σ 22 Stress component in the y-axis direction, σ 12 Is the shear stress in the x-y plane, T r Represents room temperature, T 0 Represents the initial temperature of cooling;
the stress component is expressed as a strain component, and is noted as:
the deformation coordination conditions for introducing the assumption of elastomechanics continuity are recorded as:
setting the multi-phase structure distribution after semantic segmentation in the step 2 as x c Setting the corresponding strain distribution calculated based on the finite element model in step 4 as epsilon t Combining the three strain componentsForming a three-dimensional array, and recording as:
wherein, # denotes the switching layer and denotes ε t Is composed of epsilon t,11 ,ε t,22 ,
Three-dimensional arrays stacked in sequence;
generator generated strain array
Is recorded as:
ε m ＝[ε m,11 #ε m,22 #ε m,12 ] (7)
ε t and epsilon m All can be equivalent to a multi-channel picture A, a convolution kernel is set as a multi-dimensional array C, and the two-dimensional convolution calculation of three channels can be expressed as:
where ζ and ξ are the receptive field positions, k is the number of channels, and ×, which represents the convolution.
Calculating the strain coordination error of the strain image by adopting a difference method, wherein any pixel point P ζ,ξ The second order partial derivatives in each direction are noted as:
formula (5) can be represented as:
introduction of custom fixed convolution kernel C p And is recorded as:
strain coordination error L for calculating strain field by adopting convolution operation C And is recorded as:
wherein f is a pixel side length size correction coefficient.
When the prediction target is a strain field, the FEM stress distribution is set as a true reference, and a stress-strain relation error L is calculated based on the equation (4) H,12 And L H,3 And is recorded as:
wherein L is H,12 And L H,3 For constraining the axial strain and shear strain components, respectively.
L C ，L H,12 And L H,3 And meanwhile, the weight optimization is used for the constraint generator. Discriminator D loss L of true strain load distribution S,D And is recorded as:
L S,D ＝log D(r c ,σ t ,L C,t ) (14)
wherein r is c For semantically segmented structural images,. epsilon t For true strain distribution corresponding to the structure, L C,t For strain coordination errors of the true strain distribution,
the strain distribution generated by the generator G, the discriminator loss L thereof SG,D And is recorded as:
L SG,D ＝log(1-D(r c ,G(r c )),L C,m ) (15)
wherein L is C,m Is the strain coordination error of the composite strain distribution; discriminator loss L D Can be written as:
where b is the sample batch size.
When the predicted target is a stress distribution, the FEM strain distribution is set as a true reference, and a stress-strain relation error L is calculated based on equation (3) H,12 And L H,3 And is recorded as:
wherein L is H,12 And L H,3 For constraining the axial stress and the shear stress components, respectively.
L C ，L H,12 And L H,3 And meanwhile, the weight optimization is used for the constraint generator. Discriminator D loss L of true strain load distribution S,D And is recorded as:
L S,D ＝log D(r c ,σ t ,L C,t ) (18)
stress distribution generated by generator G, and discriminator loss L thereof SG,D And is recorded as:
L SG,D ＝log(1-D(r c ,G(r c )),L C,m ) (19)
wherein L is C,m Synthesizing strain coordination error of strain distribution, calculating the generated stress distribution as strain distribution according to formula (3), and then adopting convolution calculation based on formula (12); discriminator loss L D Can be written as:
generator loss L G Including L1 norm L 1 And elastomechanically constrained L C ，L S,12 And L S,3 And is recorded as:
therefore, the generation of the elastic-mechanical constraint opposes the net total loss function, noted as:
wherein,
is a loss of elastodynamics L E Coefficient matrix of, L E Is recorded as:
L E ＝[L C,m ,L S,12 ,L S,3 ] T (23)
and 6, constructing a training sample set by taking the semantic segmentation structure chart obtained in the step 2 as model input and the load distribution obtained in the step 4 as model output. Dividing the semantic division structure chart obtained in the step 2 and the load distribution chart such as stress, strain and the like calculated in the step 4 into a plurality of local views, and establishing a training data set based on the divided structure chart and the load chart at the corresponding position: cutting the high-resolution composite structure picture and the corresponding stress and strain pictures into a plurality of groups of local pictures meeting the size condition of the generated confrontation network input picture established in the step 5, and establishing a training sample set; establishing a corresponding material attribute and boundary condition matrix according to the structural picture for calculating the elastic mechanical loss in the training process; approximately 70% of the samples were selected for training of the network, and the remaining samples were used for validation.
And 7, training the generated countermeasure network established in the step 6 based on the training set established in the step 5, wherein the load distribution is stress distribution and strain field of the multiphase composite structure, and the stress and strain of the multiphase composite structure are respectively used as output to train the generated countermeasure network.
Step 8, the composite structure to be tested is preprocessed in the step 2, and the strain field of the structure to be tested is rapidly predicted by adopting the network model trained in the step 7: preprocessing the high-resolution picture of the multiphase composite structure to be detected according to the steps 1 and 2, and adjusting the image scale to be the same as the size of the image in the training set; cutting the preprocessed structure into a plurality of groups of local pictures meeting the size condition of the generated confrontation network input picture established in the step 5, and sequentially predicting the corresponding strain distribution by adopting the network model trained in the step 7; and combining the predicted multiple groups of local strain distributions, and removing the overlapping area to obtain complete strain distribution.
In the present example, a typical complex two-phase microstructure of CaO-MgO-Al 2O 3-SiO 2(CMAS) infiltrated thermal barrier coating (TC-CMAS) was used as the strength evaluation target. During the cooling phase at the end of the gas turbine operation, the solidified CMAS develops a more complex state of strain in the coating due to the different coefficient of thermal expansion from the ceramic matrix. Meanwhile, the coating is mostly prepared by adopting an air plasma spraying method, so that the microstructure of the coating is complex and changeable, and a large amount of heat load analysis is required for the changeable structure. In this embodiment, a method for quickly evaluating a load state of a multiphase composite structure image is specifically provided, and modeling for a new structure morphology is not required each time, and an overall architecture is shown in fig. 2. The specific implementation mode of the method is mainly divided into a training stage and a thermal strain evaluation stage:
a training stage:
as shown in fig. 3, a sample based on a thermal barrier coating is subjected to electron microscope shooting of a section microstructure image, scales of different images are adjusted to be the same, and the size of the scale is recorded. Adopting Gaussian filtering to process the shot SEM image, and reducing the noise level of the SEM image; the contrast is properly enhanced, so that the CMAS inclusion and the gray scale difference of the ceramic matrix are different; and then, carrying out binarization processing on the picture based on a local threshold method, and marking the CMAS and the ceramic substrate by different colors to finish semantic segmentation of the structural image. And carrying out boundary identification on the segmented microstructure, removing the sawtooth effect of the boundary by adopting a CA (conditional access) method, and obtaining the structural boundary established by the image restoration finite element model.
And obtaining the temperature boundary condition of coating cooling, and obtaining the basic mechanical property parameters of the CMAS and the ceramic matrix. And establishing an image restoration geometric model based on the structural boundary, dividing grids, setting boundary conditions and material parameters, and calculating the strain distribution of the multiple groups of coatings. And establishing a corresponding material attribute and boundary condition matrix based on the distribution of the CMAS and the ceramic matrix, and establishing a training data set for generating the countermeasure network by combining the microstructure image and the strain load distribution. And establishing a generation countermeasure network model for predicting the coating strain distribution, and constraining the network model by adopting the generalized Hooke's law and a strain coordination condition. The calculation of finite element data is often time consuming and has a lot of repetitive work. Therefore, the input picture resolution for generating the countermeasure network is generally set lower, 256 × 256 in this example. Therefore, the high-resolution image can be divided into a plurality of low-resolution local images, and the training data set is established based on the small-scale high-definition image.
And training and generating a confrontation network model based on the established training data set. Wherein generalized hooke's law considering thermal expansion and strain coordination conditions constrain the gradient update of the generator during training. The strain coordination error, the strain distribution and the structural image of the generated strain field are input into a discriminator together, and the discriminator outputs the discrimination error. The network is trained until the error meets a threshold or a predetermined number of steps is reached.
A prediction stage:
and shooting a microstructure image of the coating sample to be detected by adopting an electron microscope. And performing semantic segmentation by adopting the same processing mode as the images in the training set, and adjusting the image scale to be the same as the training set. And establishing a corresponding coating material attribute and boundary condition matrix according to the segmented structural image, and taking the matrix and the structural image as input data. In the prediction stage, a model does not need to be established for each sample, and a trained generation countermeasure network is directly adopted as a structure-strain fitting model. And inputting the structural picture and the parameter matrix into the fitting model, so that the strain distribution of the coating can be rapidly predicted. And when the number of the electron microscope images of the sample to be detected is large, repeating the process of establishing the images and the parameter matrix until the data quantity requirement of the database is met.
It should be noted that the training phase in the embodiment includes a large amount of data collection, data analysis, data extraction, and the like, and the good computing hardware and parallel software design can provide a more convenient data processing and analysis means, and the work content is a preparation premise for the on-line phase thermal state evaluation. And the work in the online stage is insensitive to the requirement of computing hardware, and can be completely and quickly realized by relying on a common single-core computer.
The electronic device of the present invention includes a Central Processing Unit (CPU) that can perform various appropriate actions and processes according to computer program instructions stored in a Read Only Memory (ROM) or computer program instructions loaded from a storage unit into a Random Access Memory (RAM). In the RAM, various programs and data required for the operation of the device can also be stored. The CPU, ROM, and RAM are connected to each other via a bus. An input/output (I/O) interface is also connected to the bus.
A plurality of components in the device are connected to the I/O interface, including: an input unit such as a keyboard, a mouse, etc.; an output unit such as various types of displays, speakers, and the like; storage units such as magnetic disks, optical disks, and the like; and a communication unit such as a network card, modem, wireless communication transceiver, etc. The communication unit allows the device to exchange information/data with other devices via a computer network such as the internet and/or various telecommunication networks.
The processing unit performs the various methods and processes described above, such as steps 1-8. For example, in some embodiments, steps 1-8 can be implemented as a computer software program tangibly embodied in a machine-readable medium, such as a memory unit. In some embodiments, part or all of the computer program may be loaded and/or installed onto the device via ROM and/or a communication unit. When the computer program is loaded into RAM and executed by the CPU, one or more of steps 1-8 described above may be performed. Alternatively, in other embodiments, the CPU may be configured to perform steps 1-8 in any other suitable manner (e.g., by way of firmware).
The functions described herein above may be performed, at least in part, by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: a Field Programmable Gate Array (FPGA), an Application Specific Integrated Circuit (ASIC), an Application Specific Standard Product (ASSP), a system on a chip (SOC), a load programmable logic device (CPLD), and the like.
Program code for implementing the methods of the present invention may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the program codes, when executed by the processor or controller, cause the functions/operations specified in the flowchart and/or block diagram to be performed. The program code may execute entirely on the machine, partly on the machine, as a stand-alone software package partly on the machine and partly on a remote machine or entirely on the remote machine or server.
In the context of the present invention, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
While the invention has been described with reference to specific embodiments, the invention is not limited thereto, and various equivalent modifications and substitutions can be easily made by those skilled in the art within the technical scope of the invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (10)
1. A method for rapidly evaluating the loading state of a multiphase composite structure image is characterized by comprising the following steps:
acquiring a profile image of the multiphase composite structure, establishing an image restoration finite element analysis model, and calculating load distribution under different boundary conditions and geometric characteristics;
constructing a generation countermeasure network constrained by an elastic mechanical model, and training based on data acquired by a finite element analysis model;
and evaluating the load state of the multiphase composite structure to be detected based on the trained generated countermeasure network.
2. The method for rapidly evaluating the loading state of the multiphase composite structure image according to claim 1, characterized in that the method specifically comprises the following steps:
s1: acquiring a section morphology image of a multiphase composite structure, and constructing an image library;
s2: performing semantic segmentation on images in an image library according to different phases to obtain a semantic segmentation structure chart;
s3: performing boundary identification and anti-aliasing optimization on the semantically segmented image;
s4: establishing an image restoration finite element model based on the optimized structure boundary coordinates, and calculating the load distribution of the structure under the preset working condition parameters;
s5: establishing an elastic mechanical constraint generation confrontation network model for fitting a structural image and load distribution relation;
s6: constructing a training sample set by taking the semantic segmentation structure chart obtained in the step S2 as model input and the load distribution obtained in the step S4 as model output;
s7: training the generated confrontation network model based on the training sample set;
s8: and acquiring a multi-phase structural profile image to be detected, preprocessing the multi-phase structural profile image based on S2, and sending the multi-phase structural profile image to a trained generated confrontation network model to acquire a load state.
3. The method for rapidly evaluating the loading state of the multiphase composite structure image according to claim 2, wherein the step S2 specifically comprises:
s21: processing the profile image by Gaussian filtering;
s22: adjusting the contrast of the image and enhancing the gray difference of each phase;
s23: the different phase components are marked with different colors.
4. The method for rapidly evaluating the loading state of the multiphase composite structure image according to claim 2, wherein the step S3 specifically comprises:
s31: identifying each phase component area in the semantic segmentation structure chart in the step S2 based on a connected component analysis method;
s32: and performing antialiasing optimization on the identified structure boundary by using a coordinate averaging method.
5. The method for rapidly estimating the loading state of the multi-phase composite structural image according to claim 2, wherein the step S4 specifically comprises:
carrying out finite element analysis on the structure boundary optimized in the step S3, and establishing a geometric model of a two-dimensional multiphase structure; setting corresponding material properties for each phase component, a base material and the like; dividing the geometric model by adopting quadrilateral meshes, wherein the sizes of the meshes of different groups of structures are the same; and setting boundary conditions according to preset working conditions, and submitting and calculating the load distribution of the multiple groups of structures under the nonlinear condition.
6. The method for rapidly evaluating the loading state of the multiphase composite structure image according to claim 2, wherein the step S5 specifically comprises:
establishing an initial generation countermeasure network for pixel-to-pixel conversion based on a pix2pix architecture: a multi-layer U-shaped convolution neural network is established as a generator G for generating a countermeasure network, and a downsampling convolution neural network is established as a discriminator D for generating the countermeasure network.
7. The method according to claim 6, wherein the load distribution is a stress distribution and a strain field of the multi-phase composite structure, the training of the countermeasure network is performed by using the stress and the strain of the multi-phase composite structure as output, and the step S5 specifically includes:
the initial generation of the loss function against the network is noted as:
wherein G is trained to maximize logD (r, G (r)) loss and D is trained to minimize logD (r, s), E e Representing the mathematical expectation, λ is the L1 norm loss L 1 R is the input of the network, s is the true reference of the output, L 1 (G) Is recorded as:
L 1 (G)＝E r,s [||s-G(r)|| 1 ]
when the predicted target is a stress distribution, s is the finite element stress field σ t (ii) a When the predicted target is a strain distribution, s is the finite element strain field ε t ，
A generalized hooke's law is introduced that considers thermal expansion, written as:
wherein E ═ E/(1-v) 2 ) ν '═ ν/(1- ν), α' ═ α (1+ ν), G ═ E/2(1+ ν), E is the elastic modulus, epsilon 11 Is the elastic strain component in the x-axis direction, ε 22 Is the strain component in the y-axis direction, ε 12 Is the shear strain in the x-y plane, v represents the Poisson's ratio, α is the coefficient of thermal expansion, σ represents the stress component in each axial direction, σ 11 As a stress component in the direction of the x-axis, σ 22 Stress component in y-axis direction, σ 12 Is the shear stress in the x-y plane, T r Represents room temperature, T 0 Represents the initial temperature of cooling;
each stress component is expressed as a strain component, and is expressed as:
the deformation coordination conditions for introducing the assumption of elastomechanics continuity are recorded as:
let the multi-phase structure distribution after semantic segmentation in step S2 be x c The corresponding strain distribution calculated based on the finite element model in step S4 is set as ε t Combining the three strain components into a three-dimensional array, denoted as:
wherein, # represents a switching layer and represents ε t Is composed of epsilon t,11 ,ε t,22 ,
Three-dimensional arrays stacked in sequence;
generator generated strain array
Is recorded as:
ε m ＝[ε m,11 #ε m,22 #ε m,12 ]
ε t and epsilon m All can be equivalent to a multi-channel picture A, a convolution kernel is set as a multi-dimensional array C, and the two-dimensional convolution calculation of three channels can be expressed as:
where ζ and ξ are the receptive field positions, k is the number of channels, which represents the convolution,
by differential methodCalculating the strain coordination error of the strain image, and calculating any pixel point P ζ,ξ The second order partial derivatives in each direction are noted as:
the deformation coordination condition for the elasto-mechanical continuity assumption can be expressed as:
introduction of custom fixed convolution kernel C p And is recorded as:
strain coordination error L for calculating strain field by adopting convolution operation C And is recorded as:
wherein f is a pixel side length correction coefficient,
when the predicted target is strain distribution, the stress load distribution calculated by finite element is set as a real reference, and the stress-strain relation error L is calculated based on the stress component H,12 And L H,3 And is recorded as:
wherein L is H,12 And L H,3 For constraining the axial and shear strain components, L, respectively C ，L H,12 And L H,3 While the weight optimization for the constraint generator is used,
determination of true strain load distributionD loss L of discriminator S,D And is recorded as:
L S,D ＝logD(r c ,ε t ,L C,t )
wherein r is c For semantically segmented structural images,. epsilon t For true strain distribution corresponding to the structure, L C,t For strain coordination errors of the true strain distribution,
the strain distribution generated by the generator G, the discriminator loss L thereof SG,D And is recorded as:
L SG,D ＝log(1-D(r c ,G(r c )),L C,m )
wherein L is C,m Is the strain coordination error of the composite strain distribution;
discriminator loss L D Can be written as:
where b is the sample batch size,
when the predicted target is stress distribution, the strain distribution calculated by finite element is set as real reference, and stress-strain relation error L is calculated based on generalized Hooke's law considering thermal expansion H,12 And L H,3 And is recorded as:
wherein L is H,12 And L H,3 For constraining the axial and shear stress components, L, respectively C ，L H,12 And L H,3 While the weight optimization for the constraint generator,
discriminator D loss L of true strain load distribution S,D And is recorded as:
L S,D ＝logD(r c ,σ t ,L C,t )
stress distribution generated by generator G, and discriminator loss L thereof SG,D And is recorded as:
L SG,D ＝log(1-D(r c ,G(r c )),L C,m )
wherein L is C,m Synthesizing strain coordination errors of strain distribution, calculating the generated stress distribution into the strain distribution according to a generalized Hooke's law considering thermal expansion, and then performing convolution calculation based on the strain coordination errors; discriminator loss L D Can be written as:
generator loss L G Including L1 norm L 1 And elastomechanically constrained L C ，L S,12 And L S,3 And is recorded as:
therefore, the generation of the elastic-mechanical constraint opposes the net total loss function, noted as:
wherein,
is a loss of elastodynamics L E Coefficient matrix of, L E Is recorded as:
L E ＝[L C,m ,L S,12 ,L S,3 ] T 。
8. the method for rapidly evaluating the loading state of the multiphase composite structure image according to claim 2, wherein the step S6 specifically comprises:
the semantic segmentation structure chart and the corresponding load distribution are segmented into a plurality of groups of local pictures meeting the size condition of the generated confrontation network input picture established in the step S5, and a training sample set is established; establishing a corresponding material attribute and boundary condition matrix according to the structural picture for calculating the elastic mechanical loss in the training process; approximately 70% of the samples were selected for training of the network, and the remaining samples were used for validation.
9. The method for rapidly evaluating the loading state of the multiphase composite structure image according to claim 2, wherein the step S8 specifically comprises:
preprocessing the high-resolution multiphase composite structure picture to be detected according to the step S2, and adjusting the image scale to be the same as the size of the image in the training set; dividing the preprocessed structure into a plurality of groups of local pictures meeting the size condition of the generated confrontation network input picture established in the step S5, and sequentially predicting the corresponding strain distribution by adopting the network model trained in the step S7; and combining the predicted multiple groups of local strain distributions, and removing the overlapped area to obtain complete strain distribution.
10. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method for fast load state estimation of the multiphase composite structure image according to any one of claims 1-9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210471459.4A CN114970240A (en) | 2022-04-28 | 2022-04-28 | Method and equipment for rapidly evaluating load state of multi-phase composite structure image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210471459.4A CN114970240A (en) | 2022-04-28 | 2022-04-28 | Method and equipment for rapidly evaluating load state of multi-phase composite structure image |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114970240A true CN114970240A (en) | 2022-08-30 |
Family
ID=82979921
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210471459.4A Pending CN114970240A (en) | 2022-04-28 | 2022-04-28 | Method and equipment for rapidly evaluating load state of multi-phase composite structure image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114970240A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116343958A (en) * | 2023-03-03 | 2023-06-27 | 湖北工业大学 | Method and system for judging connectivity of multiphase system |
| CN117347378A (en) * | 2023-08-18 | 2024-01-05 | 大连理工大学 | Composite material damage identification method based on optical fiber measurement and neural network |
-
2022
- 2022-04-28 CN CN202210471459.4A patent/CN114970240A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116343958A (en) * | 2023-03-03 | 2023-06-27 | 湖北工业大学 | Method and system for judging connectivity of multiphase system |
| CN116343958B (en) * | 2023-03-03 | 2024-05-10 | 湖北工业大学 | Method and system for judging connectivity of multiphase system |
| CN117347378A (en) * | 2023-08-18 | 2024-01-05 | 大连理工大学 | Composite material damage identification method based on optical fiber measurement and neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wang et al. | Structural fatigue life prediction considering model uncertainties through a novel digital twin-driven approach | |
| CN114970240A (en) | Method and equipment for rapidly evaluating load state of multi-phase composite structure image | |
| CN110895814B (en) | Aero-engine hole-finding image damage segmentation method based on context coding network | |
| EP3931650B1 (en) | Method and system for optimizing process parameters in an additive manufacturing process | |
| Gao et al. | An automated approach for machining allowance evaluation of casting parts | |
| EP3317852A2 (en) | Method in constructing a model of a scenery and device therefor | |
| Li et al. | Deep learning-based defects detection of certain aero-engine blades and vanes with DDSC-YOLOv5s | |
| Vitse et al. | Toward virtual design and optimization of a structural test monitored by a multi-view system | |
| Liu et al. | DeviationGAN: A generative end-to-end approach for the deviation prediction of sheet metal assembly | |
| Zhang et al. | Parametrically-upscaled continuum damage mechanics (PUCDM) model for multiscale damage evolution in bending experiments of glass-epoxy composites | |
| CN112287484B (en) | Complex engineering system reliability design method based on vector agent model | |
| CN115310311B (en) | Method for calculating fatigue life of metal material under spectral load | |
| Long et al. | An intelligent crack damage assessment method by integrating information and physics | |
| CN108304649B (en) | High-rise building deformation prediction method | |
| Qie et al. | Data-driven deviation generation for non-ideal surfaces of Skin Model Shapes | |
| Zhang et al. | Design and Research of Low‐Cost and Self‐Adaptive Terrestrial Laser Scanning for Indoor Measurement Based on Adaptive Indoor Measurement Scanning Strategy and Structural Characteristics Point Cloud Segmentation | |
| CN114972695A (en) | Point cloud generation method and device, electronic equipment and storage medium | |
| CN115018162A (en) | Method and system for predicting machining quality in industrial finish machining process in real time | |
| CN112182739A (en) | Aircraft structure non-probability credible reliability topological optimization design method | |
| Riddle et al. | Effects of Defects Part A: Stochastic Finite Element Modeling of Wind Turbine Blades with Manufacturing Defects for Reliability Estimation | |
| CN110543724A (en) | Satellite structure performance prediction method for overall design | |
| JPH11203330A (en) | Shape deformation mode generation system, shape optimization analyzing system and record medium in which program used for the same is recorded | |
| Mao et al. | An automatic detection and online quality inspection method for workpiece surface cracks based on machine vision | |
| JP7332068B1 (en) | Automatic calculation system, automatic calculation method, machine learning model, and method for generating machine learning model | |
| Zhang et al. | An Integrated Neural Network and Finite Element Method for Defect Characterization and Strength Prediction of Unidirectional Composites |
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 |