AU2020101874A4 - A method for predicting high-temperature mechanical properties of heat-resistant alloys based on deep learning - Google Patents
A method for predicting high-temperature mechanical properties of heat-resistant alloys based on deep learning Download PDFInfo
- Publication number
- AU2020101874A4 AU2020101874A4 AU2020101874A AU2020101874A AU2020101874A4 AU 2020101874 A4 AU2020101874 A4 AU 2020101874A4 AU 2020101874 A AU2020101874 A AU 2020101874A AU 2020101874 A AU2020101874 A AU 2020101874A AU 2020101874 A4 AU2020101874 A4 AU 2020101874A4
- Authority
- AU
- Australia
- Prior art keywords
- heat
- deep learning
- resistant alloys
- mechanical properties
- temperature mechanical
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C60/00—Computational materials science, i.e. ICT specially adapted for investigating the physical or chemical properties of materials or phenomena associated with their design, synthesis, processing, characterisation or utilisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30136—Metal
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Quality & Reliability (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Investigating Or Analyzing Materials Using Thermal Means (AREA)
Abstract
This invention relates to a method for predicting high-temperature mechanical properties of
heat-resistant alloys based on deep learning. The invention can realize the direct prediction
from microstructure to high-temperature mechanical properties of heat-resistant alloys,
improve the efficiency of heat-resistant alloys high-temperature performance detection, save
heat-resistant alloys high-temperature testing costs, and improve the accuracy of existing
heat-resistant alloys high-temperature mechanical properties prediction methods.
-1/8
Original_____s Normalization and Orignalataasesplitting of database
0.
Read and
Traiing alidtion convert
[cRava Evaluate
erformanc
7 New
- microstructures
Iteration:'
Training:
* and Daap:
,validation:
---------- g-----------p
Prediction 1izV results <c21]
Figure1I
Description
-1/8
Original_____s Normalization and Orignalataasesplitting of database
0.
Read and Traiing alidtion convert
[cRava Evaluate erformanc
7 New - microstructures Iteration:' Training: * and Daap: ,validation: ---------- g-- - - - - p
Prediction results <c21] 1izV
Figure1I
PATENTS ACT 1990
A method for predicting high-temperature mechanical properties of heat-resistant alloys
based on deep learning
The invention is described in the following statement:-
A method for predicting high-temperature mechanical properties of heat-resistant
alloys based on deep learning
The invention relates to the field of high-temperature performance analysis of heat
resistant alloys, in particular to a method for predicting high-temperature mechanical
properties of heat-resistant alloys based on deep learning.
The mechanical properties of materials are directly determined by their microstructures.
Heat-resistant alloys are high-end products of metal materials, which are widely used in
petroleum, chemical, power generation, aerospace and other industries associated with
national security and national economy. Meanwhile, the development of heat-resistant
alloys is also closely related to the progress in these fields. The rapid development of
world industrialization raises the application of heat-resistant alloys in today's society. As
a result, the testing demand for high-temperature mechanical properties of heat-resistant
alloys has also increased.
The long-term conventional tests are always used to detect the high-temperature
mechanical properties of heat-resistant alloys, which are energy and time consuming. For
example, the design life of a jet turbine is often 104 hours (1 year), the design life of an
ethylene pyrolysis furnace tube is 10' hours (10 years), and the design life of nuclear
reactor is as long as 40 years. Although the long-term service performance of heat- resistant alloys can be predicted by extrapolation method based on the results of short term high-temperature experiments, the short-term experiments still often last dozens to hundreds of hours. The existing methods are impossible to achieve efficient prediction of the high-temperature mechanical properties of heat-resistant alloys. On the other hand, there are many accumulated experimental data on the micro-structures and properties of heat-resistant alloys, but the valuable data are considered inefficient to guide new predictions.
The Materials Genome Initiative (MGI), which has been widely developed in the world
since 2011, emphasizes finding and establishing the interrelationship of atomic
arrangement, phase formation, micro-structure formation and material properties. MGI
technology provides a new way of innovating materials research and improvement,
accelerates the process of materials from design to application. Consequently, MGI
technology will support the development of cutting-edge materials, advanced
manufacturing and high-tech ultimately. Although the MGI technology has been
developed for many years, the prediction of mechanical properties has always been the
shortcoming. The difficulty lies in the fact that materials are composed of a large number
of atoms, their mechanical properties are usually the result of complex coupling of
multiple physical mechanisms. For example, the formation and interaction of
dislocations, the interaction between dislocations and micro-structures and other defects,
the formation and growth of defects and cracks, the interaction between materials/defects
and the environment, etc. Therefore, there is no universal and widely used method to
predict the mechanical properties of heat-resistant alloys directly from micro-structures. It is still a very complicated and difficult task to establish a prediction model of material mechanical properties.
The invention proposes a method for predicting the high-temperature mechanical
properties of heat-resistant alloys based on deep learning. A convolution neural network
(CNN) is adopted as the deep learning model. The normalized microstructure images
(including the alloys before and after service) of heat-resistant alloys and their
corresponding experimental high-temperature mechanical properties are collected to
construct the original experimental database. The database is used optimize the deep
learning model, in order to achieve the purpose of predicting the high temperature
mechanical properties of the heat-resistant alloys based on the micro-structure images
directly. The invention explores the potential value of a large amount of accumulated
experimental data of the heat-resistant alloys, and makes full use of them to develop a
simple and quick prediction method for the heat-resistant alloys from micro-structures to
high-temperature mechanical properties. The invention is important for improving the
detection and prediction efficiency of mechanical properties of heat-resistant alloys.
In order to achieve the above objectives, the present invention adopts the following
technical solutions:
A method for predicting high-temperature mechanical properties of heat-resistant alloys
based on deep learning, including the following steps:
Si. For the same type of heat-resistant alloys, collect different micro-structure
images and their corresponding high-temperature mechanical properties under specific
testing conditions to form the original experimental database.
S2. Perform data pre-processing on the micro-structure images in the original
experimental database.
S3. According to the distribution of experimental high-temperature mechanical
properties of heat-resistant alloys, the original experimental database is divided into
continuous N categories. The class associated with a specific category is the label of the
corresponding images. Then all the labelled image data are segregated randomly into 3
groups: training dataset, validation dataset, and test dataset.
S4. Read each image dataset, and convert all the images into floating-point tensors.
S5. Build a deep learning model, configure the model architecture and parameters.
The training dataset is used to train and optimize the model's parameters in deep learning
process. The validation dataset is used to monitor the over fitting and identify the best
performance of the model. And the test dataset is used as the final performance
evaluation of the chosen model on the remainder of the withheld data.
S6. Use the optimized deep learning model, input the new micro-structure image of
the heat-resistant alloys to obtain the output of the predicted high-temperature mechanical
properties.
The main advantage of this invention is that the deep learning model can extract feature
automatically without the intervention of artificial subjective judgment. In the deep
learning model, each layer learns to recognize a set of specific features on the basis of integrating and reorganizing the output of the previous layer. As the depth of the neural network increases, the features extracted by the layer become increasingly abstract and complex. The activation of higher layers carry less information about the specific input, and more information about the prediction target. Therefore, the deep learning model can process large-scale high-dimensional micro-structure data of materials. The invention can realize the direct prediction of the high temperature mechanical properties of heat resistant alloys from their micro-structures, which supplements the shortcomings of MGI technology in the field of mechanical properties prediction. On the basis of the improved property prediction accuracy, the model will reduce a lot of time and cost for the testing of heat-resistant alloys. And the trained deep learning model can also be applied to the properties prediction of other materials through the transfer learning method.
Figure 1 is a work-flow schematic of the main steps of the method of the present
invention.
Figure 2 is a typical architecture of deep learning network.
Figure 3 is the architecture of the deep learning network in a specific embodiment of
the present invention.
Figure 4 is the training and verification accuracy of the original deep learning
network in an embodiment of the present invention.
Figure 5 is the training and verification loss of the original deep learning network in
an embodiment of the present invention.
-'7
Figure 6 is the training and verification accuracy of an optimized deep learning
network in an embodiment of the present invention.
Figure 7 is the training and verification loss of an optimized deep learning network
in an embodiment of the present invention.
Figure 8 is the data processing of the optimized deep learning network and the high
temperature mechanical property prediction result in a specific embodiment of the present
invention.
As shown in Figure 1, a method for predicting high-temperature mechanical properties of
heat-resistant alloys based on deep learning includes the following steps.
Si. For the same type of heat-resistant alloys, collect different micro-structure images
and their corresponding high-temperature mechanical properties test results of the same
series of heat-resistant alloys under specific testing conditions to form the original
experimental database. The micro-structure images can be taken from heat-resistant
alloys that before or after high temperature service, all the micro-structures used in one
deep learning model must be taken from the heat-resistant alloys either before or after
high-temperature service. In this embodiment, the material is selected as HP40Nb type
heat-resistant alloys for ethylene pyrolysis furnace tube. The micro-structure images are
taken from the heat-resistant alloys that have not been serviced at high-temperatures. The
micro-structures are scanning electron microscope (SEM) images. The stress rupture time
of the heat-resistant alloys is the high-temperature mechanical property, which is
obtained in the experiment at a temperature of 1100 °C and a stress of 17 MPa.
S2. Perform data pre-processing on the micro-structure images in the original
experimental database. The micro-structure images in the original experimental database
are shown in different in magnifications and/or sizes, so all the images in the original
experimental database are adjusted by scaling and cutting to insure the normalized micro
structure magnification and image size. In this embodiment, all the images in the original
experimental database are first adjusted to 500 times magnification of the micro
structure, and then all the adjusted images are cut into the size of 170x170 pixels.
S3. According to the distribution of experimental high-temperature mechanical properties
of heat-resistant alloys, divide the original experimental database into continuous N
categories. The class associated with a specific category is the label Use the divided
category value as the category label of the corresponding images. In this embodiment,
according to the distribution of stress rupture times of heat-resistant alloys in the original
experimental database, the original experimental database is divided into three categories:
-100 hours, 100-200 hours, and 200-250 hours. And the category name is used as the
label of the micro-structure images therein.
The labeled image data are segregated randomly into training dataset, validation dataset,
and test dataset. The number of training group data is much higher than that of validation
dataset, and the number of validation group data is not less than the number of test group
data. In this embodiment, the ratio of training dataset: validation dataset: test dataset is
8:1:1.
S4. Read each group of image dataset, and convert all the images into floating-point
tensors according to the following steps:
S41. Decode every image file into a gray-scale grid.
S42. Rescale the grid gray values ([0,255] interval) to the [0,1] interval, and adjust the
input image size to 150x150x3.
S5. Build a deep learning model, configure the model architecture and parameters. The
training dataset is used to train and optimize the model's parameters in deep learning
process. The validation dataset is used to monitor the over fitting and identify the best
performance of the model. And the test dataset is used as the final performance
evaluation of the chosen model on the remainder of the withheld data.
Specifically, as shown in Figure 2, the deep learning adopts a CNN model. The deep
learning model is composed of A continuous convolutional blocks and B fully connected
layers. Each convolutional block has C continuous convolutional layers and D pooling
layers. The specific calculation content is as follows:
The convolution operation of the input image feature map is:
S(i, j) = (I * W)(i,j) = YYI(m, n)W(i - mj - n) (1) m n
wherein I is the image feature I E Rm "n, W is the two-dimensional convolution
kernel W R i.
For image tensors X Rmxnx D, the convolution operation is:
X = WD *X+b= WD,d * Xd+b (2)
Yc = ReLU (X) (3) wherein W is the three-dimensional convolution kernel (W ERi x D ). The above m, n, i, j and D are the dimensions of the tensor on each axis, b is the bias matrix,
WDd and Xdare the slice matrices of Wand X' respectively. The rectified linear unit
function ReLU () is the activation function, and its expression is:
ReLU(x)= max{0, x} (4)
The pooling layer performs the maximum pooling operation, and its expression is:
yd = aMn {xa} (5)
wherein, xa is the activation value of each neuron in the R area;
The softmax function is used as the classifier. For S scalars, the expression
is: softmax(x) exp (x) (6) i:i exp (xi)
Configuration of the deep model includes loss function, optimizer and monitoring
indicators.
The categorical cross-entropy is adopted as the loss function, and its expression is:
H(L, P) = - 1 L, log Pn (7)
wherein L is the distribution of sample labels and P is the distribution of model prediction
values. They both belong to the N-dimensional vectors;
The optimizer is an algorithm based on the stochastic gradient descent method. The
algorithm can be any one of RMSProp algorithm, AdaGrad algorithm, and AdaDelta
algorithm, etc.. The RMSProp algorithm is used in this solution.
The stochastic gradient descent algorithm is:
1) Input: trainingdatasetTr ={(xy(t)} T validation dataset V, learning rate a
and initialization parameter 0.
2) Initialize 0 randomly.
3) Repeat the following steps.
The samples in the training set Tr are randomly reordered. When t = 1, ... , T, collect a
mini-batch containing b samples from the training set Tr, select the samples (x(,y(0),
calculate the gradient estimation§- + Ve L(f(x(); 0), y()), apply the update
o <- 0 - a§;
4) The error rate of modelf(x; 0) on the validation set V no longer drops and finally it
outputs 0. The input and output in this embodiment are shown in Figure 3.
The monitoring index is the accuracy and loss of training and verification dataset in this
embodiment.
Error! Reference source not found.Error! Reference source not found.
Error! Reference source not found.Error! Reference source not found.
(1) Tune hyper parameters of the deep learning model.
(2) Use data augmentation on labeled images to generate more training data from
existing training samples. Data augmentation methods include random rotation of
the image from 0-180°, horizontal and vertical flipping, shifting in height and
width of 0.5 ratio, enclosing filling, and reflection filling.
(3) Use weight regularization or dropout methods to limit the complexity of the deep
learning model. In this embodiment, the dropout is set to 0.5.
S6. Input the microstructure images of the heat resistant alloys into the optimized deep
learning model, then the model outputs the stress rupture time predictions. The
microstructure magnification and image size of the tested data should be normalized
according to step S2.
Error!Reference source notfound.
Based on the stress rupture time of the heat-resistant alloys predicted by the deep learning
model, the corresponding stress rupture times of the heat-resistant alloys at other
temperatures can be obtained according to the extrapolation methods. In this
embodiment, the Larson-Miller method is used to extrapolate the stress rupture times of
the tested heat-resistant alloys at other temperatures. According to the above
embodiments, the present invention has high efficiency and accuracy in predicting the
high-temperature mechanical properties of heat-resistant alloys, and the optimization
approaches have a significant effect on the improvement of model performance. With the
further optimization of the model and the increase of experimental data, the prediction
accuracy of the deep learning method will be further improved. The invention explores
the potential value of a large amount of accumulated experimental data of the structure
and performance of the heat-resistant alloys, and makes full use of them to develop a
quantitative prediction method for the heat-resistant alloys from microstructure to high
temperature mechanical properties. The invention is important for improving the
detection and prediction efficiency of mechanical properties of heat-resistant alloys.
In addition, high-temperature mechanical properties also include creep properties, fatigue
properties, etc. Other types of data can be used to predict the corresponding properties.
The above are only the preferred embodiments created by the present invention and are
not intended to limit the creation of the present invention. Any modification, equivalent
replacement and improvement made within the spirit and principle of the present
invention, should be included in the protection scope of the present invention and
creation.
Claims (9)
1.A method for predicting high-temperature mechanical properties of heat-resistant alloys
based on deep learning is characterized in that it comprises the following steps:
Si. For the same type of heat-resistant alloys, collect different micro-structure images
and their corresponding high-temperature mechanical properties under specific testing
conditions to form the original experimental database.
S2. Perform data pre-processing on the micro-structure images in the original
experimental database.
S3. According to the distribution of experimental high-temperature mechanical properties
of heat-resistant alloys, the original experimental database is divided into continuous N
categories. The class associated with a specific category is the label of the corresponding
images. Then all the labelled image data are segregated randomly into 3 groups: training
dataset, validation dataset, and test dataset.
S4. Read each image dataset, and convert all the images into floating-point tensors.
S5. Build a deep learning model, configure the model architecture and parameters. The
training dataset is used to train and optimize the model's parameters in deep learning
process. The validation dataset is used to monitor the over fitting and identify the best
performance of the model. And the test dataset is used as the final performance
evaluation of the chosen model on the remainder of the withheld data.
S6. Use the optimized deep learning model, input the micro-structure image of the heat
resistant alloys into the model to obtain the output of the predicted high-temperature
mechanical properties.
2.The method for predicting high-temperature mechanical properties of heat-resistant
alloys based on deep learning according to claim 1 is characterized in Si that the micro
structure images can be taken from heat-resistant alloys that before or after high
temperature service, all the micro-structures used in one deep learning model must be
taken from the heat-resistant alloys either before or after high-temperature service.
3.The method for predicting high-temperature mechanical properties of heat-resistant
alloys based on deep learning according to claim 1 is characterized in S2 that the micro
structure images in the original experimental database are shown in different in
magnifications and/or sizes. All the images in the original experimental database are
adjusted by scaling and cutting to insure the normalized micro-structure magnification
and image size.
4.The method for predicting high-temperature mechanical properties of heat-resistant
alloys based on deep learning according to claim 1 is characterized in S3 that the number
of samples in training dataset is much more than the that of verification dataset.
5.The method for predicting high-temperature mechanical properties of heat-resistant
alloys based on deep learning according to claim 1 is characterized in S4 that the floating
point tensors processing includes the following steps:
S41. Decode every image file into pixels or gray-scale grid.
S42. Convert the pixels or the gray scale grid into a three-dimensional floating
point tensor.
6.The method for predicting high-temperature mechanical properties of heat-resistant
alloys based on deep learning according to claim 1 is characterized in S5 that the deep
learning model adopts a CNN model. The deep learning model is composed of A
continuous convolutional blocks and B fully connected layers with a stacked way. Each
convolutional block has continuous C convolutional layers and D pooling layers.
7.The method for predicting high-temperature mechanical properties of heat-resistant
alloys based on deep learning according to claim 6 is characterized in S5 that the
configuration of the model structure and model parameters includes a loss function, an
optimizer and a monitoring index. The loss function is a cross-entropy, the optimizer is
an algorithm based on a stochastic gradient descent method, and the monitoring index is
the accuracy and loss of training and verification datasets.
8.The method for predicting high-temperature mechanical properties of heat-resistant
alloys based on deep learning according to claim 1 is characterized in S5 that the
optimization method for improving and enhancing the model performance includes any
one or any combination of the following:
(1) Obtain more experimental data to increase the number of samples in the database.
(2) Tune hyper parameters of the deep learning model.
(3) Use data augmentation on labeled images to generate more training data from
existing training samples. Data augmentation methods include image rotation, flipping,
shifting, enclosing filling, and reflection filling, etc.
(4) Use weight regularization or dropout methods.
(5) Keep the test dataset constant, merge the training dataset with the validation dataset
and segregate the merged dataset randomly into K groups (K>3). For every i-th group (i=
1, 2,..., K). Train the deep learning model on the remaining K-i groups, and then verify
the deep learning model on group i.
9. The method for predicting high-temperature mechanical properties of heat-resistant
alloys based on deep learning according to claim 1 is characterized in S6 that the
microstructure magnification and image size of the tested data should be normalized
according to step S2. Specially, for the prediction of the stress rupture time of the heat
resistant alloys by the deep learning model, the stress rupture times at other temperatures
can be further predicted according to the extrapolation methods.
-1/8-
Figure 1
-2/8-
Figure 2
-3/8-
Figure 3
-4/8-
Figure 4
-5/8-
Figure 5
-6/8-
Figure 6
-7/8-
Figure 7
-8/8-
Figure 8
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| AU2020101874A AU2020101874A4 (en) | 2020-08-18 | 2020-08-18 | A method for predicting high-temperature mechanical properties of heat-resistant alloys based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| AU2020101874A AU2020101874A4 (en) | 2020-08-18 | 2020-08-18 | A method for predicting high-temperature mechanical properties of heat-resistant alloys based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| AU2020101874A4 true AU2020101874A4 (en) | 2020-09-24 |
Family
ID=72513356
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| AU2020101874A Active AU2020101874A4 (en) | 2020-08-18 | 2020-08-18 | A method for predicting high-temperature mechanical properties of heat-resistant alloys based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| AU (1) | AU2020101874A4 (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112861665A (en) * | 2021-01-25 | 2021-05-28 | 中国石油天然气集团有限公司 | Oil casing heat treatment tissue inspection method based on deep learning |
| CN113033105A (en) * | 2021-04-06 | 2021-06-25 | 东北大学 | Steel material fatigue performance prediction method based on mechanical theory guide transfer learning |
| CN113033106A (en) * | 2021-04-06 | 2021-06-25 | 东北大学 | Steel material performance prediction method based on EBSD and deep learning method |
| CN113361768A (en) * | 2021-06-04 | 2021-09-07 | 重庆科技学院 | Grain depot health condition prediction method, storage device and server |
| CN113706516A (en) * | 2021-08-31 | 2021-11-26 | 西安交通大学 | Metallographic structure identification method, system, equipment and storage medium |
| CN114580271A (en) * | 2022-02-16 | 2022-06-03 | 昆明贵金属研究所 | Method for realizing solid-liquid phase temperature prediction of multi-element precious metal alloy brazing filler metal |
| CN114821136A (en) * | 2022-04-11 | 2022-07-29 | 成都信息工程大学 | Self-adaptive cloud particle sub-image data processing method |
| CN114996658A (en) * | 2022-07-20 | 2022-09-02 | 中国空气动力研究与发展中心计算空气动力研究所 | Projection-based hypersonic aircraft aerodynamic heat prediction method |
| EP4361949A1 (en) * | 2022-10-31 | 2024-05-01 | Doosan Enerbility Co., Ltd. | Device and method for degradation prediction |
| CN118090820A (en) * | 2024-04-23 | 2024-05-28 | 常州市安贞建设工程检测有限公司 | Magnesium-based inorganic fireproof plate performance detection method and system |
| CN114580271B (en) * | 2022-02-16 | 2024-07-02 | 昆明贵金属研究所 | Method for realizing solid-liquid phase temperature prediction of multi-element noble metal alloy solder |
-
2020
- 2020-08-18 AU AU2020101874A patent/AU2020101874A4/en active Active
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112861665A (en) * | 2021-01-25 | 2021-05-28 | 中国石油天然气集团有限公司 | Oil casing heat treatment tissue inspection method based on deep learning |
| CN113033105A (en) * | 2021-04-06 | 2021-06-25 | 东北大学 | Steel material fatigue performance prediction method based on mechanical theory guide transfer learning |
| CN113033106A (en) * | 2021-04-06 | 2021-06-25 | 东北大学 | Steel material performance prediction method based on EBSD and deep learning method |
| CN113033105B (en) * | 2021-04-06 | 2023-09-19 | 东北大学 | Method for predicting fatigue performance of steel material based on mechanical theory guiding transfer learning |
| CN113033106B (en) * | 2021-04-06 | 2023-09-19 | 东北大学 | Steel material performance prediction method based on EBSD and deep learning method |
| CN113361768A (en) * | 2021-06-04 | 2021-09-07 | 重庆科技学院 | Grain depot health condition prediction method, storage device and server |
| CN113706516B (en) * | 2021-08-31 | 2024-04-02 | 西安交通大学 | Metallographic structure identification method, system, equipment and storage medium |
| CN113706516A (en) * | 2021-08-31 | 2021-11-26 | 西安交通大学 | Metallographic structure identification method, system, equipment and storage medium |
| CN114580271A (en) * | 2022-02-16 | 2022-06-03 | 昆明贵金属研究所 | Method for realizing solid-liquid phase temperature prediction of multi-element precious metal alloy brazing filler metal |
| CN114580271B (en) * | 2022-02-16 | 2024-07-02 | 昆明贵金属研究所 | Method for realizing solid-liquid phase temperature prediction of multi-element noble metal alloy solder |
| CN114821136A (en) * | 2022-04-11 | 2022-07-29 | 成都信息工程大学 | Self-adaptive cloud particle sub-image data processing method |
| CN114996658B (en) * | 2022-07-20 | 2022-11-08 | 中国空气动力研究与发展中心计算空气动力研究所 | Projection-based hypersonic aircraft aerodynamic heat prediction method |
| CN114996658A (en) * | 2022-07-20 | 2022-09-02 | 中国空气动力研究与发展中心计算空气动力研究所 | Projection-based hypersonic aircraft aerodynamic heat prediction method |
| EP4361949A1 (en) * | 2022-10-31 | 2024-05-01 | Doosan Enerbility Co., Ltd. | Device and method for degradation prediction |
| CN118090820A (en) * | 2024-04-23 | 2024-05-28 | 常州市安贞建设工程检测有限公司 | Magnesium-based inorganic fireproof plate performance detection method and system |
| CN118090820B (en) * | 2024-04-23 | 2024-06-21 | 常州市安贞建设工程检测有限公司 | Magnesium-based inorganic fireproof plate performance detection method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| AU2020101874A4 (en) | A method for predicting high-temperature mechanical properties of heat-resistant alloys based on deep learning | |
| CN110797096B (en) | Deep learning-based high-temperature mechanical property prediction method for heat-resistant alloy | |
| Wang et al. | Hybrid physics-based and data-driven models for smart manufacturing: Modelling, simulation, and explainability | |
| CN111444940A (en) | Fault diagnosis method for critical parts of fan | |
| CN108985335B (en) | Integrated learning prediction method for irradiation swelling of nuclear reactor cladding material | |
| CN112147432A (en) | BiLSTM module based on attention mechanism, transformer state diagnosis method and system | |
| CN110309867A (en) | A kind of Mixed gas identification method based on convolutional neural networks | |
| CN109359668A (en) | A kind of concurrent diagnostic method of aero-engine multiple faults | |
| Noraas et al. | Structural material property tailoring using deep neural networks | |
| Akrim et al. | Self-Supervised Learning for data scarcity in a fatigue damage prognostic problem | |
| Gao et al. | Data-driven machine learning for alloy research: recent applications and prospects | |
| CN116152146A (en) | Cast aluminum cylinder cover mechanical property prediction method based on GAN and CNN | |
| CN114926702B (en) | Small sample image classification method based on depth attention measurement | |
| CN110045691A (en) | A kind of multitasking fault monitoring method of multi-source heterogeneous big data | |
| Lööv | Comparison of undersampling methods for prediction of casting defects based on process parameters | |
| Lee et al. | A random forest‐based fatigue damage classification framework using global loading and displacement information | |
| Van Niekerk | Degradation estimation of high energy steam piping using hybrid recurrent neural networks | |
| Oluwafemi et al. | Machine Learning Approaches to Identify Environmental Damage on Superalloys | |
| CN113761703A (en) | Heat-resistant alloy creep property prediction method based on deep learning fusion model | |
| CN109740636B (en) | Stratum oil-water layer identification method based on AdaBoost combined with GRU | |
| Canyi et al. | Weights embedding Informer prediction algorithm-based fault diagnosis framework for nuclear power plant | |
| Gan et al. | Exploring Transformer Life Forecasting through an In-Depth Analysis Utilizing the Random Forest Algorithm in Research and Development | |
| Anderson | Automated Detection of Helium Bubbles in Inconel X-750 and Modeling Their Effect on its Mechanical Properties | |
| Mudabbiruddin et al. | Simulation of Ageing of Aircraft | |
| CN115238823A (en) | Energy industry-oriented time sequence data standardized treatment method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| FGI | Letters patent sealed or granted (innovation patent) |