CN112699545A - Method and system for establishing three-dimensional processing diagram based on improved artificial neural network model - Google Patents
Method and system for establishing three-dimensional processing diagram based on improved artificial neural network model Download PDFInfo
- Publication number
- CN112699545A CN112699545A CN202011544014.1A CN202011544014A CN112699545A CN 112699545 A CN112699545 A CN 112699545A CN 202011544014 A CN202011544014 A CN 202011544014A CN 112699545 A CN112699545 A CN 112699545A
- Authority
- CN
- China
- Prior art keywords
- dimensional
- neural network
- processing
- distribution
- 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
Images
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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/08—Thermal analysis or thermal optimisation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Geometry (AREA)
- Computer Hardware Design (AREA)
- Computer Graphics (AREA)
- Powder Metallurgy (AREA)
Abstract
The invention belongs to the technical field of data processing, and discloses a method and a system for establishing a three-dimensional processing diagram based on an improved artificial neural network model, which are used for acquiring material stress-strain data based on the thermal processing range of a material; carrying out generalized supplementation on stress distribution of a three-dimensional space by utilizing a neural network based on acquired data, and analyzing and processing the data to obtain three-dimensional space distribution of stress in three dimensions and other three-dimensional distribution conditions; based on the determined three-dimensional distribution, the energy dissipation and the generated heat are calculated, and a processing map is constructed. The invention optimizes the traditional processing diagram composition method by means of the intelligent thought of the neural network, not only can realize better correspondence between the microstructure and the processing diagram in the actual production application, but also can play a great role in promoting in industry 4.0 to assist the intelligent production and product control of the industry.
Description
Technical Field
The invention belongs to the technical field of data processing, and particularly relates to a method and a system for establishing a three-dimensional processing diagram based on an improved artificial neural network model.
Background
At present, the classical Process Map (PM) established based on dynamic material model theory (DMM) is considered to be an excellent tool to obtain a suitable process window and to track the evolution of the deformed microstructure. Due to its excellent predictability and effectiveness, the method has not only been widely used in practical production, but also is continuously advanced in theoretical research, and is now applied to various metals and alloys, including zirconium alloys, copper alloys, aluminum alloys, and nickel-based superalloys. Conventional PM generally considers that strain has little effect on processability, but this assumption affects the general utility of the model. In order to better achieve the combination of PM and finite element simulation, prior art 1 proposes a three-dimensional machining map on the basis of the relevant theory. To describe the relationship between true stress and different deformation conditions, the three-dimensional machining map is created under two assumptions 1 that the stress distribution can be expressed as a power law property (i.e., σ ═ kepsilon)′m) (ii) a 2, using an interpolation method (i.e. sigma)i(x)＝ai+bi(x-xi)+ci(x-xi)2+di(x-xi)3) And obtaining the three-dimensional contour line. Based on the simplified model, the prior art 2 establishes a three-dimensional machining map of the U720LI superalloy and verifies a suitable machining window. In order to improve numerical stability, prior art 3 suggests using polynomial functions and least squares regression more than 3 times.
However, thermal deformation involves a complex series of non-linear and interactive effects, so a great deal of research finds that conventional process maps based on fitting do not combine well with practical experimental phenomena. Later scholars further introduced Arrhenbus constitutive equation or JC equation to describe the stress-strain relationship, but finally, differences still exist in some special places.
In general, conventional curve fitting techniques may not be suitable for modeling these highly complex interaction models. To address this problem, artificial neural network models have been developed as an effective alternative to describe highly accurate nonlinear data and interactions. The neural network is characterized in that even if the form of the nonlinear relation is unknown, some experimental results are wrong, and the functional relation among the variables can be obtained. The neural network technology enables the functional relationship modeling of any engineering problem to be possible, and has wide application prospect.
Through the above analysis, the problems and defects of the prior art are as follows: 1) the traditional processing diagram has no wide applicability, particularly in some special areas, such as high-temperature high-strain-rate areas, the processability along with the occurrence of dynamic recrystallization is greatly improved, and the traditional processing diagram cannot effectively identify the transformation, so that the processing efficiency is reduced; 2) the traditional method causes data errors and de-characterization to a certain extent, on one hand, curve fitting technology may not be suitable for modeling of the highly complex interaction models, and on the other hand, the data interpolation method reduces the real effectiveness of data. 3) The generalization capability of the traditional constitutive equation has defects.
The difficulty in solving the above problems and defects is: the conventional constitutive equation is developed for decades, but due to the complex influence of deformation conditions and the coupling among different deformation conditions, the uncertainty of stress is caused, the fitting accuracy is reduced, and the correlation can only be kept between 3% and 5% in the past. To solve this problem, artificial neural network models have been developed as an effective alternative to describe highly accurate nonlinear data and interactions.
The significance of solving the problems and the defects is as follows: the effectiveness of the machining diagram technology is recognized in the related field, and not only for theoretical research, the machining diagram can be used for researching the structural evolution of the material in the thermal deformation process, but also for actual production, the machining diagram can guide the selection of the production process in real time. A higher-precision processing diagram means that in actual production, a destabilization area can be effectively avoided, so that the production cost is reduced, and the production safety is improved; and more efficient deformation parameters can be selected in a proper processing window, the capacity is increased, and the energy consumption is controlled. For scientific research, the new processing diagram construction method also provides a new idea.
Disclosure of Invention
Aiming at the problems in the prior art, the invention provides a method and a system for establishing a three-dimensional processing diagram based on an improved artificial neural network model.
The invention is realized in such a way that the method for establishing the three-dimensional processing diagram based on the improved artificial neural network model comprises the following steps:
selecting a proper thermal processing range, including a temperature range and a strain rate range, wherein the temperature and strain rate interval cannot be too large in consideration of the generalization capability of a neural network;
acquiring material stress strain data based on the hot processing range of the material;
thirdly, constructing a matrix based on deformation conditions (temperature, strain rate and strain) and stress, and performing special normalization treatment;
step four, learning part (75%) of data by utilizing a neural network, and constructing a network;
step five, in order to ensure the effectiveness of learning and the generalization ability of the network, the network is compared with the rest (25%) data, and a data accuracy report is given;
step six, generalization of stress distribution in a three-dimensional space is realized by using a trained network, and anti-normalization treatment is carried out;
and step seven, calculating energy dissipation and generated heat based on the three-dimensional distribution condition of the stress, and constructing a processing diagram.
Another object of the present invention is to provide an improved artificial neural network model-based three-dimensional machining map creation system for implementing the improved artificial neural network model-based three-dimensional machining map creation method, the improved artificial neural network model-based three-dimensional machining map creation system including:
the data acquisition module is used for acquiring material stress strain data based on the hot processing range of the material;
the neural network module is used for carrying out generalized supplementation on stress distribution of a three-dimensional space by utilizing a neural network based on the acquired data, and acquiring the three-dimensional space distribution and other three-dimensional distribution conditions of stress in three dimensions based on the acquired stress strain data;
and the processing diagram module is used for calculating energy dissipation and generated heat based on the determined three-dimensional distribution condition and constructing a processing diagram.
Further, the material stress strain data acquisition amount is more than 24.
Further, the three dimensions of the stress are respectively: temperature, amount of strain, and strain rate.
Further, the other three-dimensional distribution conditions include: three-dimensional distribution of variable rate sensitivity coefficients, three-dimensional distribution of power dissipation and three-dimensional distribution of instability diagrams.
Further, the three-dimensional distribution of the instability map includes: and the instability condition of the instability diagram is determined by the visual selection of a plurality of integrated instability condition models by a user according to materials.
It is a further object of the invention to provide a computer device comprising a memory and a processor, the memory storing a computer program which, when executed by the processor, causes the processor to perform the steps of:
acquiring material stress strain data based on the hot working range of the material;
carrying out generalized supplementation on stress distribution of a three-dimensional space by utilizing a neural network based on acquired data, and analyzing and processing the data to obtain three-dimensional space distribution of stress in three dimensions and other three-dimensional distribution conditions;
based on the determined three-dimensional distribution, the energy dissipation and the generated heat are calculated, and a processing map is constructed.
Another object of the present invention is to provide a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a processor, the processor executes the method for establishing a three-dimensional processing map based on an improved artificial neural network model.
The invention also aims to provide an information data processing terminal which is used for realizing the method for establishing the three-dimensional processing diagram based on the improved artificial neural network model.
The invention further aims to provide a device for tracking the evolution of the deformed microstructure in various metal and alloy fields by carrying the improved artificial neural network model-based three-dimensional processing map building system.
By combining all the technical schemes, the invention has the advantages and positive effects that: the invention optimizes the traditional processing diagram composition method by means of the intelligent thought of the neural network, not only can realize better correspondence between the microstructure and the processing diagram in the actual production application, but also can play a great role in promoting in industry 4.0 to assist the intelligent production and product control of the industry. And the invention has realized the complete programming and application based on matlab, can dock many other application environments through the api, facilitate the supervisor to carry on direct data interaction and product control.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings used in the embodiments of the present application will be briefly described below, and it is obvious that the drawings described below are only some embodiments of the present application, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a flow chart of a method for establishing a three-dimensional processing diagram based on an improved artificial neural network model according to an embodiment of the present invention.
FIG. 2 is a schematic structural diagram of a system for building a three-dimensional processing diagram based on an improved artificial neural network model according to an embodiment of the present invention;
in the figure: 1. a data acquisition module; 2. a neural network module; 3. and (5) processing a drawing module.
FIG. 3 is a process diagram of a nickel-based GH925 alloy constructed according to an embodiment of the invention under different strain levels.
FIG. 4 is a schematic diagram comparing process maps constructed in different ways provided by embodiments of the present invention.
Fig. 5-9 are process diagrams of different conditions provided by embodiments of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail with reference to the following embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Aiming at the problems in the prior art, the invention provides a method for establishing a three-dimensional processing diagram based on an improved artificial neural network model, and the invention is described in detail below with reference to the accompanying drawings.
As shown in fig. 1, the method for establishing a three-dimensional processing diagram based on an improved artificial neural network model according to an embodiment of the present invention includes the following steps:
s101, collecting material stress strain data based on the hot working range of the material;
s102, performing generalized supplementation on stress distribution of a three-dimensional space by utilizing a neural network based on acquired data, and analyzing and processing the data to obtain three-dimensional space distribution and other three-dimensional distribution conditions of stress in three dimensions;
and S103, calculating energy dissipation and generated heat based on the determined three-dimensional distribution condition, and constructing a processing diagram.
As shown in fig. 2, the system for establishing a three-dimensional processing diagram based on an improved artificial neural network model according to an embodiment of the present invention includes:
the data acquisition module 1 is used for acquiring material stress strain data based on the hot processing range of the material;
the neural network module 2 is used for carrying out generalization supplement on stress distribution of a three-dimensional space by utilizing a neural network based on the acquired data, and acquiring the three-dimensional space distribution and other three-dimensional distribution conditions of stress in three dimensions based on the acquired stress strain data;
and a processing diagram module 3 for calculating energy dissipation and generated heat based on the determined three-dimensional distribution and constructing a processing diagram.
The material stress strain data acquisition amount provided by the embodiment of the invention is more than 24.
The three dimensions of the stress provided by the embodiment of the invention are respectively as follows: temperature, amount of strain, and strain rate.
Other three-dimensional distribution conditions provided by the embodiment of the invention include: three-dimensional distribution of variable rate sensitivity coefficients, three-dimensional distribution of power dissipation and three-dimensional distribution of instability diagrams.
The three-dimensional distribution of the instability image provided by the embodiment of the invention comprises the following steps: and the instability condition of the instability diagram is determined by visually selecting the integrated multiple instability condition models by a user according to materials.
The technical effects of the present invention will be further described with reference to specific embodiments.
Example 1:
as shown in the figure, the overall experimental thought of the invention is composed of two parts, one part is a neural network module, and the other part is a processing diagram module. The processing diagram module is built on the basis of a traditional dynamic model, but similarity simplification is reduced as much as possible in the calculation process so as to obtain the most accurate energy dissipation and heat generation. And the other part is a new preprocessing module, namely a neural network module, different from the traditional machining diagram calculation, wherein the invention introduces a three-dimensional distribution diagram of the genetic algorithm optimized feedforward neural network corresponding force in three dimensions, including temperature, strain and strain rate, and the machining diagram is constructed by combining the machining diagram construction module on the basis of the distribution diagram.
In order to realize the construction of the processing diagram, the invention needs to acquire deformation data of the material within a certain range. The data acquisition range is consistent with the thermal processing range of the material, such as 900-1150 ℃ if required, and the deformation rate is 0.01-10s-1The strain amount is 0-0.9, at least four pieces of stress strain data are required to be introduced, wherein the stress strain data are respectively (1)900 ℃ and 0.01s-1Deformation amount 0.9; (2)900 ℃ for 10s-1Deformation amount 0.9; (3)1150 deg.C, 0.01s-1Deformation amount 0.9; (4)1150 deg.C for 10s-1And the deformation amount was 0.9. Stress distribution of the whole three-dimensional space is supplemented in a generalization mode through a neural network, however, due to the limitation of artificial intelligence learning, the purpose of high precision is difficult to achieve by only using four pieces of data for learning, and related researches show that the high-precision learning effect can be obtained under the condition of using 24 pieces of deformation data, namely the temperature range is 900-; strain rate in the range of 0.01-10s-1At intervals of 10 times; the strain amount is continuous deformation data. Such data conditioning works best, although subsequent research would hopefully continue to increase the predictability and generalization ability of neural networks, thereby reducing the data requirements. After data is imported, corresponding stress three-dimensional distribution, variable rate sensitivity coefficient three-dimensional distribution, power dissipation three-dimensional distribution and three-dimensional distribution of a instability graph can be obtained through a program, multiple models on the market are integrated under the instability condition, and visual selection can be carried out according to materials.
Example 2:
as shown in the figure, the invention is based on phasesIt is contemplated that the modified tooling diagram provides a suitable heat treatment window for the construction of the nickel-based GH925 alloy. Since the three-dimensional distribution is difficult to observe and describe well, the processing diagrams under different strain quantities are sliced so as to be convenient to observe the processing diagrams under different strain quantities. The GH925 alloy has the strain rate of 0.01-10s in the temperature range of 900-1150 DEG C-1The PM at 0.05 intervals is shown, with true strain of 0.1-0.85. The gray areas indicate the destabilizing zones. The total area of the destabilized region decreases and then increases with increasing strain. When strain<At 0.25, the destabilization region can be divided into three regions, I region (about 900 ℃ 1150 ℃ for 4.26-10 s)-1) Region II (about 900--1) And region III (about 996--1). With increasing strain, zone I shows a slight tendency to expand and zone II reverses. Furthermore, when the strain reaches 0.45, the III region substantially disappears. In addition, the present invention can be readily observed that the constructed power dissipation diagram and the distribution of the buckling domains are approximately the same in the strain range of 0.5-0.85, indicating that the strain has little effect on the hot working of large deformations. Interestingly, when strain>0.65, another high temperature and high strain rate (1150 deg.C, 10 s)-1) The lower stable region appears and continues to expand because the larger sized recrystallized grains completely replace the deformed grains, i.e., complete dynamic recrystallization.
In order to better and more intuitively illustrate the predictability of the processing diagram, the processing diagram with the deformation of 0.85 and the corresponding microstructure are selected for analysis, and the following figure shows two processing diagrams constructed in different modes. The left figure is an improved processing diagram, and the right figure is a traditional classical processing diagram, which are different in size in the overall composition, but have some differences in details and precision, such as 5# and 4# and 1# of high temperature and high strain rate, and also include some edge positions.
And a region A, which is defined as a destabilizing region. The destabilization zone accounted for almost 36.13% of the overall treatment profile. As shown in the figure, at 900 deg.C/0.1 s-1The formation of an adiabatic shear band, marked by the area within the yellow dashed line, was clearly observed for representative microstructures obtained under deformation conditions. At higher temperatures and strain rates (1050 ℃/10 s)-1) An inhomogeneous structure (as indicated by the yellow arrows) with a large number of deformed grain boundaries may occur, indicating that a large orientation gradient remains inside the unrecrystallized grain boundaries. Furthermore, nucleation of DRX grains can also be observed at grain boundaries and twin boundaries. 2# (1050 ℃/10 s) with increasing temperature-1) Recrystallization volume ratio of (1 # (1050 ℃/10 s)-1) The recrystallization volume of (a) was significantly increased to 2.3% and 18.67%, respectively. For nickel-based alloys with lower stacking fault energy, the increase in deformation temperature seriously affects the mobility of dislocations and promotes the cross slip and rise of dislocations. However, a large number of deformed grains can still be observed in the matrix, showing severe local deformation characteristics. Even under moderate deformation conditions (950 ℃/1 s)-1) The present invention also allows a large number of entangled dislocations to be observed, which are piled up at the grain boundaries and converted into dislocation cells. In addition, the diffusion and entanglement of dislocations inside the crystal grains can also be observed.
And a region B: the zone is distributed at the temperature of 1111 ℃ and 1150 ℃, and the strain rate is 4-10s-1. Traditional PMs define this region as a destabilizing region, while neural network-based PMs exhibit different results. The invention is carried out at 1150 ℃ for 10s-1Microscopic structure observation under deformation conditions proves the effectiveness of improving PM. According to the invention, the large DRX appears in the No. 3, and the unstable thermal deformation behavior, such as adiabatic shearing deformation or the initiation and the expansion of micro-cracks, does not exist. 3# (1150 ℃/10 s)-1) Has a recrystallization rate of higher than 4# (1150 ℃/1 s)-1) The recrystallization rates of (a) and (b) were 0.53 and 0.32, respectively. This indicates that continuous recrystallization is occurring and is possible at high temperatures and high strain rates, which has been explained in previous studies. Samantray believes that this false determination is due to an imbalance between energy input and energy dissipation.
And (3) area C: this area shows similar power dissipation profiles in both processing diagrams and is not described herein in detail.
And (3) area D: same as above
And (4) area E: region E is located at a strain rate and temperature range of 0.01-0.1s-1And 970-. In particular, the zones have different dissipation factor distributions at different PMs. Specifically, classical PM has efficiency values of about 0.2-0.3, which is related to the DRV mechanism. To assess correctness, the present invention shows that this region is in 8# (1050 ℃/0.01 s)-1) Typical microstructure of (a). As can be seen from the yellow dashed box in the figure, only a small portion of the elongated grains remain. The lower strain rate provides more time for grain growth than # 6, so that most of the deformed parent material is replaced by unstrained DRX grains. Furthermore, EBSD analysis of sample # 8 showed the presence of equiaxed particles (volume fraction of about 89.05%), twin boundaries of about 10.74% (Σ3), and concluded that twin boundary formation was accompanied by DRX process or subsequent growth. It can be concluded from this that the deformation mechanisms of the investigated material in this region are mainly recrystallization and grain growth. In contrast, the improved PM has better generalization capability.
And a region F: region F corresponds to a lower dissipation value range of 29-37% at a temperature and strain rate of 1089--1Within the range of (1). The efficiency at higher temperatures and lower strain rates is relatively low, which may be caused by coarse grain growth after full dynamic recrystallization. At 1150 ℃/0.01s-1Under the conditions, the typical deformed structure shows obvious equiaxial crystals, the average grain size is 84.49 degrees m, and the deformed structure is completely replaced. The main reason for grain growth is that an increase in deformation temperature reduces the solute resistance effect and increases the mobility of grain boundaries. On the other hand, a lower strain rate allows sufficient time for nucleation and growth. Generally, the occurrence of full DRX at this stage consumes energy stored in the material, thereby suppressing the consumption of energy. Dynamic recovery plays an important role in achieving the accumulation of threading dislocations.
The above description is only for the specific embodiments of the present invention, but the scope of the present invention is not limited thereto, and any modification, equivalent replacement, and improvement made by those skilled in the art within the technical scope of the present invention disclosed in the present invention should be covered within the scope of the present invention.
Claims (10)
1. A method for establishing a three-dimensional processing diagram based on an improved artificial neural network model is characterized by comprising the following steps of:
acquiring material stress strain data based on the hot working range of the material;
carrying out generalized supplementation on stress distribution of a three-dimensional space by utilizing a neural network based on acquired data, and analyzing and processing the data to obtain three-dimensional space distribution of stress in three dimensions and other three-dimensional distribution conditions;
based on the determined three-dimensional distribution, the energy dissipation and the generated heat are calculated, and a processing map is constructed.
2. A three-dimensional processing diagram building system based on an improved artificial neural network model is characterized by comprising:
the data acquisition module is used for acquiring material stress strain data based on the hot processing range of the material;
the neural network module is used for carrying out generalized supplementation on stress distribution of a three-dimensional space by utilizing a neural network based on the acquired data, and acquiring the three-dimensional space distribution and other three-dimensional distribution conditions of stress in three dimensions based on the acquired stress strain data;
and the processing diagram module is used for calculating energy dissipation and generated heat based on the determined three-dimensional distribution condition and constructing a processing diagram.
3. The improved artificial neural network model-based three-dimensional processing map creation system of claim 2, wherein the material stress-strain data collection amount is greater than 24.
4. The improved artificial neural network model-based three-dimensional processing map creation system according to claim 2, wherein the three dimensions of the stress are respectively: temperature, amount of strain, and strain rate.
5. The improved artificial neural network model-based three-dimensional processing map creation system of claim 2, wherein the other three-dimensional distribution comprises: three-dimensional distribution of variable rate sensitivity coefficients, three-dimensional distribution of power dissipation and three-dimensional distribution of instability diagrams.
6. The improved artificial neural network model-based three-dimensional processing map creation system according to claim 2, wherein the three-dimensional distribution of the instability map comprises: and the instability condition of the instability diagram is determined by the visual selection of a plurality of integrated instability condition models by a user according to materials.
7. A computer device, characterized in that the computer device comprises a memory and a processor, the memory storing a computer program which, when executed by the processor, causes the processor to carry out the steps of:
acquiring material stress strain data based on the hot working range of the material;
carrying out generalized supplementation on stress distribution of a three-dimensional space by utilizing a neural network based on acquired data, and analyzing and processing the data to obtain three-dimensional space distribution of stress in three dimensions and other three-dimensional distribution conditions;
based on the determined three-dimensional distribution, the energy dissipation and the generated heat are calculated, and a processing map is constructed.
8. A computer-readable storage medium storing a computer program which, when executed by a processor, causes the processor to execute the improved artificial neural network model-based three-dimensional machining map building method of claim 1.
9. An information data processing terminal, characterized in that the information data processing terminal is used for implementing the method for establishing the three-dimensional processing diagram based on the improved artificial neural network model of claim 1.
10. The device for tracking the evolution of the deformed microstructure in various metal and alloy fields by carrying the improved artificial neural network model-based three-dimensional processing diagram establishing system of claims 2-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011544014.1A CN112699545A (en) | 2020-12-23 | 2020-12-23 | Method and system for establishing three-dimensional processing diagram based on improved artificial neural network model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011544014.1A CN112699545A (en) | 2020-12-23 | 2020-12-23 | Method and system for establishing three-dimensional processing diagram based on improved artificial neural network model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112699545A true CN112699545A (en) | 2021-04-23 |
Family
ID=75509508
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011544014.1A Pending CN112699545A (en) | 2020-12-23 | 2020-12-23 | Method and system for establishing three-dimensional processing diagram based on improved artificial neural network model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112699545A (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106055844A (en) * | 2016-07-06 | 2016-10-26 | 中南大学 | Prediction and control method of nickel-base super alloy microstructure on the basis of BP (Back Propagation) neural network |
| CN109695299A (en) * | 2018-12-29 | 2019-04-30 | 中国十七冶集团有限公司 | A kind of inorfil vacuum heat-preserving board construction method based on BIM technology |
| CN110006568A (en) * | 2019-04-12 | 2019-07-12 | 中国地质科学院地质力学研究所 | It is a kind of to obtain the method for three-dimensional ground stress using core and its obtain system |
| CN111159863A (en) * | 2019-12-17 | 2020-05-15 | 上海交通大学 | Neural network algorithm-based material deformation and failure prediction method and system |
| CN111240268A (en) * | 2020-01-14 | 2020-06-05 | 重庆大学 | Axle system thermal error modeling method and thermal error compensation system based on SLSTM neural network |
| CN111625952A (en) * | 2020-05-21 | 2020-09-04 | 中国石油大学(华东) | Method, system, storage medium, and program for detecting three-dimensional distribution of temperature and stress |
-
2020
- 2020-12-23 CN CN202011544014.1A patent/CN112699545A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106055844A (en) * | 2016-07-06 | 2016-10-26 | 中南大学 | Prediction and control method of nickel-base super alloy microstructure on the basis of BP (Back Propagation) neural network |
| CN109695299A (en) * | 2018-12-29 | 2019-04-30 | 中国十七冶集团有限公司 | A kind of inorfil vacuum heat-preserving board construction method based on BIM technology |
| CN110006568A (en) * | 2019-04-12 | 2019-07-12 | 中国地质科学院地质力学研究所 | It is a kind of to obtain the method for three-dimensional ground stress using core and its obtain system |
| CN111159863A (en) * | 2019-12-17 | 2020-05-15 | 上海交通大学 | Neural network algorithm-based material deformation and failure prediction method and system |
| CN111240268A (en) * | 2020-01-14 | 2020-06-05 | 重庆大学 | Axle system thermal error modeling method and thermal error compensation system based on SLSTM neural network |
| CN111625952A (en) * | 2020-05-21 | 2020-09-04 | 中国石油大学(华东) | Method, system, storage medium, and program for detecting three-dimensional distribution of temperature and stress |
Non-Patent Citations (5)
| Title |
|---|
| BAI S等: "Construction of three-dimensional extrusion limit diagram for magnesium alloy using artificial neural network and its validation", 《JOURNAL OF MATERIALS PROCESSING TECHNOLOGY》 * |
| SHEN J Y等: "Hot deformation behaviors and three-dimensional processing map of a nickel-based superalloy with initial dendrite microstructure", 《JOURNAL OF ALLOYS AND COMPOUNDS》 * |
| ZHU YULONG等: "Dynamic behavior and modified artificial neural network model for predicting flow stress during hot deformation of Alloy 925", 《MATERIALS TODAY COMMUNICATIONS》 * |
| ZHU Y等: "Three-dimensional hot processing map of a nickel-based superalloy (Alloy 925) established by modified artificial neural network model", 《INTERMETALLICS》 * |
| 李玉亮等: "高强GWZ1042镁合金三维热加工图及可加工性研究", 《热加工工艺》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Ding et al. | Microstructural modelling of dynamic recrystallisation using an extended cellular automaton approach | |
| Li et al. | A three-dimensional cellular automata-crystal plasticity finite element model for predicting the multiscale interaction among heterogeneous deformation, DRX microstructural evolution and mechanical responses in titanium alloys | |
| Chen et al. | Multiscale modeling of discontinuous dynamic recrystallization during hot working by coupling multilevel cellular automaton and finite element method | |
| Chen et al. | Modeling the dynamic recrystallization in austenitic stainless steel using cellular automaton method | |
| Chen et al. | A strategy to control microstructures of a Ni-based superalloy during hot forging based on particle swarm optimization algorithm | |
| Li et al. | Hot deformation behavior of LA43M Mg-Li alloy via hot compression tests | |
| Li et al. | Research on dynamic recrystallization behavior of NiFeCr based alloy | |
| JIN et al. | Identification of nucleation parameter for cellular automaton model of dynamic recrystallization | |
| CN112288139A (en) | Air conditioner energy consumption prediction method and system based on chaotic time sequence and storage medium | |
| Zhu et al. | Three-dimensional hot processing map of a nickel-based superalloy (Alloy 925) established by modified artificial neural network model | |
| Saxena et al. | Zr–Nb alloys and its hot deformation analysis approaches | |
| Zhang et al. | Thermal deformation behavior investigation of Ti–10V–5Al-2.5 fe-0.1 B titanium alloy based on phenomenological constitutive models and a machine learning method | |
| CN112699545A (en) | Method and system for establishing three-dimensional processing diagram based on improved artificial neural network model | |
| Ahmadikia et al. | Data-driven texture design for reducing elastic and plastic anisotropy in titanium alloys | |
| Chen et al. | Online optimizing hot forming parameters for alloy parts based on action-dependent heuristic dynamic programming | |
| CN116826745A (en) | Layered and partitioned short-term load prediction method and system in power system background | |
| Zheng et al. | Crystal orientation and morphology of α lamellae in wrought titanium alloys: On the role of microstructure evolution in β processing | |
| Massih et al. | Transformation kinetics of alloys under non-isothermal conditions | |
| Wen et al. | Deep learning-based modeling of the strain rate-dependent thermomechanical processing response for a novel HIPed P/M nickel-based superalloy | |
| Chen et al. | Recrystallization Microstructure Prediction of a Hot‐Rolled AZ31 Magnesium Alloy Sheet by Using the Cellular Automata Method | |
| Xue et al. | Optimization of stabilized annealing of Al-Mg alloys utilizing machine learning algorithms | |
| Sun et al. | Allowable ranges of conventional forging parameters determination for TA15 Ti-alloy to obtain tri-modal microstructure under given subsequent heat treatment | |
| Huang et al. | Simulation of dynamic recrystallization for aluminium alloy 7050 using cellular automaton | |
| CN105740513A (en) | Simulation method for GH4169 alloy dynamic recrystallization during thermal deformation | |
| Tao et al. | Research on energy dissipation and dynamic recrystallization microstructure evolution behavior of NiTi alloy during hot deformation |
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 |