CN112329322A - Thermal spraying method and device based on convolutional neural network - Google Patents
Thermal spraying method and device based on convolutional neural network Download PDFInfo
- Publication number
- CN112329322A CN112329322A CN202011073655.3A CN202011073655A CN112329322A CN 112329322 A CN112329322 A CN 112329322A CN 202011073655 A CN202011073655 A CN 202011073655A CN 112329322 A CN112329322 A CN 112329322A
- Authority
- CN
- China
- Prior art keywords
- neural network
- convolutional neural
- network model
- training
- testing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 142
- 238000013527 convolutional neural network Methods 0.000 title claims abstract description 118
- 238000007751 thermal spraying Methods 0.000 title claims abstract description 39
- 230000008569 process Effects 0.000 claims abstract description 115
- 238000012549 training Methods 0.000 claims abstract description 74
- 238000012360 testing method Methods 0.000 claims abstract description 70
- 238000005507 spraying Methods 0.000 claims abstract description 64
- 239000011248 coating agent Substances 0.000 claims abstract description 48
- 238000000576 coating method Methods 0.000 claims abstract description 48
- 239000002245 particle Substances 0.000 claims abstract description 46
- 238000007781 pre-processing Methods 0.000 claims abstract description 10
- 238000013508 migration Methods 0.000 claims abstract description 8
- 230000005012 migration Effects 0.000 claims abstract description 8
- 239000007921 spray Substances 0.000 claims description 23
- 238000013461 design Methods 0.000 claims description 17
- 238000004088 simulation Methods 0.000 abstract description 8
- 238000002474 experimental method Methods 0.000 abstract description 7
- 238000010586 diagram Methods 0.000 description 13
- 238000005457 optimization Methods 0.000 description 12
- 238000004458 analytical method Methods 0.000 description 9
- 239000007789 gas Substances 0.000 description 3
- 239000000463 material Substances 0.000 description 3
- 239000000843 powder Substances 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 239000000919 ceramic Substances 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 239000002184 metal Substances 0.000 description 2
- 230000001105 regulatory effect Effects 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 239000012159 carrier gas Substances 0.000 description 1
- 230000001276 controlling effect Effects 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000009413 insulation Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000003647 oxidation Effects 0.000 description 1
- 238000007254 oxidation reaction Methods 0.000 description 1
- 238000007789 sealing Methods 0.000 description 1
- 238000005728 strengthening Methods 0.000 description 1
- 238000000547 structure data Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 239000000758 substrate Substances 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
- G06F30/25—Design optimisation, verification or simulation using particle-based methods
-
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/10—Numerical modelling
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Coating By Spraying Or Casting (AREA)
Abstract
The invention discloses a thermal spraying method and a device based on a convolutional neural network, wherein the method comprises the following steps: under the condition of given process parameters, acquiring the characteristic spatial distribution of flying particles or the microstructure of a coating, and preprocessing process parameter data and particle data or coating data; designing a framework of a convolutional neural network model or carrying out migration and fine tuning on a pre-trained convolutional neural network model; dividing the preprocessed data into a training set and a testing set, training the parameters of the convolutional neural network model by using the training set, and testing and evaluating the trained model by using the testing set to obtain a model which is trained and tested; and designing a required particle characteristic distribution or coating microstructure, predicting corresponding process parameters according to the evaluated convolutional neural network model, and acquiring a corresponding structure by combining the process parameters through methods such as experiments, simulation and the like. The method provides good guidance for optimizing the spraying process, and greatly reduces the cost of manually exploring the optimal process parameters.
Description
Technical Field
The invention relates to the technical field of thermal spraying, in particular to a thermal spraying method and device based on a convolutional neural network.
Background
The thermal spraying technology is a mature surface strengthening and modifying technology at present, in the thermal spraying process, coating materials such as metal, ceramic, metal ceramic and the like are heated to a molten and semi-molten state by a high-temperature and high-speed heat source, are accelerated to the surface of a part needing protection treatment, and then collide, deform and deposit to form a coating, so that the surface of a workpiece obtains different physical and chemical properties such as wear resistance, corrosion resistance, heat resistance, oxidation resistance, heat insulation, electric conduction, sealing and the like. In the spraying process, the required microstructure and characteristic parameters of the coating can be obtained by controlling the parameters of the spraying process, so that the expected coating performance is obtained. However, the spraying process involves many process parameters including the spray gun power, the plasma gas flow, the powder feeding rate, the spraying distance, the spraying angle, the spraying environment, the spraying speed, the substrate temperature, etc., and different process parameters may have internal correlation; and the performance characteristics of the sprayed coating, such as the microstructure of the coating, the porosity of the coating, the bonding strength and the like, are also many, which makes the control and optimization of the spraying process very complicated.
For the control and optimization of the spraying process, in the actual spraying, the process parameter adjustment is mainly carried out based on manual experience; in scientific research, modeling analysis is mainly performed on process optimization based on traditional design experiments and numerical simulation. However, because of the high cost, long time and low efficiency of conventional experimental research and numerical analysis, and because of the limitations of conventional methods, it is difficult to perform comprehensive analysis and optimization of spraying processes with multivariable and inherently complex correlation. While in some current studies, rapidly evolving machine learning, artificial neural network methods have been applied to modeling and analysis of the overall spray coating process, in actual spray coating, some of the research has focused on achieving the desired process parameters based on desired coating properties for optimization of the process parameters. Therefore, the modeling analysis for the forward process of spraying cannot meet the above requirements, and the reverse process of spraying needs to be analyzed. Meanwhile, the flight particle state parameters and the coating microstructure are directly or indirectly related to the coating performance by combining the data result in actual spraying. Therefore, in the modeling of the reverse spraying process, the analysis of the correlation among the flight particle state parameters, the coating microstructure and the spraying process parameters is very important. And considering that the flight particle characteristics and the coating microstructure have the characteristics of spatial distribution and the characteristics of the convolutional neural network, the method has higher industrial practical value for the modeling analysis of the spraying reverse process by adopting the convolutional neural network model.
Disclosure of Invention
The present invention is directed to solving, at least to some extent, one of the technical problems in the related art.
Therefore, the invention aims to provide a thermal spraying method based on a convolutional neural network, which provides good guidance for optimizing a spraying process and greatly reduces the cost of manually exploring optimal process parameters.
Another object of the present invention is to provide a thermal spray device based on a convolutional neural network.
In order to achieve the above object, an embodiment of an aspect of the present invention provides a thermal spraying method based on a convolutional neural network, including the following steps: step S1, acquiring experimental data of the thermal spraying process; step S2, preprocessing the experimental data to construct a database; step S3, designing a convolutional neural network model according to preset requirements; step S4, dividing the database into a training set and a testing set, wherein the training set is used for training the convolutional neural network model, and the testing set is used for testing the prediction accuracy of the trained convolutional neural network model; step S5, judging whether the training error in the training process and the testing error in the testing process are in a preset range, if the training error and the testing error are beyond the preset range, iteratively executing the steps S1-S5 until the training error and the testing error meet the preset range; step S6, designing a demand target, and acquiring ideal target data of the demand target; step S7, inputting the ideal target data into a tested convolutional neural network model, and outputting corresponding spraying process parameters; and step S8, combining the spraying process parameters with the experimental data to obtain the actual flying particle characteristic distribution or coating microstructure.
According to the thermal spraying method based on the convolutional neural network, the convolutional neural network model is used for modeling and analyzing the internal relation among the coating microstructure, the flight particle space distribution and the spraying process parameters in the thermal spraying system, so that the trained model is used for regulating and optimizing the thermal spraying process, the required process parameters are predicted according to the model by combining the actual demand result, good guidance is provided for the optimization of the spraying process, and the cost for manually searching the optimal process parameters is greatly reduced.
In addition, the thermal spraying method based on the convolutional neural network according to the above embodiment of the present invention may also have the following additional technical features:
further, in one embodiment of the present invention, the experimental data are spray process parameters, flight particle status parameters, and coating parameters accumulated during the thermal spray process.
Further, in an embodiment of the present invention, in the convolutional neural network model in step S3, the flying particle characteristic distribution or the coating microstructure is used as an input layer, and the spraying process parameters are used as an output layer.
Further, in an embodiment of the present invention, if the training error and the test error exceed the preset ranges, the database is added to expand the training set and the test set, or the architecture of the convolutional neural network model is trimmed, the migration manner of the convolutional neural network model is changed, or the number of training iterations of the convolutional neural network model is increased, and the convolutional neural network model is further adjusted until the error is within the error range.
In order to achieve the above object, another embodiment of the present invention provides a thermal spraying apparatus based on a convolutional neural network, including: the acquisition module is used for acquiring experimental data preprocessing module of the thermal spraying process and preprocessing the experimental data to construct a database; the first design module is used for designing a convolutional neural network model according to a preset requirement; the training and testing module is used for dividing the database into a training set and a testing set, wherein the training set is used for training the convolutional neural network model, and the testing set is used for testing the prediction accuracy of the trained convolutional neural network model; the judging iteration module is used for judging whether the training error in the training process and the testing error in the testing process are within a preset range or not, and if the training error and the testing error exceed the preset range, the judging iteration module executes the modules in an iteration mode until the training error and the testing error meet the preset range; the second design module is used for designing a demand target and acquiring ideal target data of the demand target; the prediction module is used for inputting the ideal target data into a tested convolutional neural network model and outputting corresponding spraying process parameters; and the actual result obtaining module is used for combining the spraying process parameters with the experimental data to obtain the actual flying particle characteristic distribution or coating microstructure.
According to the thermal spraying device based on the convolutional neural network, the convolutional neural network model is used for modeling and analyzing the internal relation among the coating microstructure, the flight particle space distribution and the spraying process parameters in the thermal spraying system, so that the trained model is used for regulating and optimizing the thermal spraying process, the required process parameters are predicted according to the model by combining the actual demand result, good guidance is provided for the optimization of the spraying process, and the cost for manually searching the optimal process parameters is greatly reduced.
In addition, the thermal spraying device based on the convolutional neural network according to the above embodiment of the present invention may also have the following additional technical features:
further, in one embodiment of the present invention, the experimental data are spray process parameters, flight particle status parameters, and coating parameters accumulated during the thermal spray process.
Further, in one embodiment of the present invention, in the convolutional neural network model of the first design module, the flight particle characteristic distribution or coating microstructure is used as an input layer, and the spraying process parameters are used as an output layer.
Further, in an embodiment of the present invention, if the training error and the test error exceed the preset ranges, the database is added to expand the training set and the test set, or the architecture of the convolutional neural network model is trimmed, the migration manner of the convolutional neural network model is changed, or the number of training iterations of the convolutional neural network model is increased, and the convolutional neural network model is further adjusted until the error is within the error range.
Additional aspects and advantages of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.
Drawings
The foregoing and/or additional aspects and advantages of the present invention will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:
FIG. 1 is a flow chart of a convolutional neural network based thermal spray method according to one embodiment of the present invention;
FIG. 2 is a block flow diagram of a thermal spray optimization method based on a convolutional neural network model;
FIG. 3 is a temperature distribution diagram of the flying particles, wherein the left side is the temperature distribution diagram of the flying particles measured in the experiment, and the right side is the temperature distribution diagram of the flying particles processed by the simulation result;
FIG. 4 is a microstructure view of a coating;
FIG. 5 is a model diagram of the relationship between the distribution of the flight particle state and the parameters of the spraying process;
fig. 6 is a schematic structural diagram of a thermal spraying device based on a convolutional neural network according to an embodiment of the present invention.
Detailed Description
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are illustrative and intended to be illustrative of the invention and are not to be construed as limiting the invention.
Hereinafter, a convolutional neural network-based thermal spraying method and apparatus according to an embodiment of the present invention will be described with reference to the accompanying drawings, and first, a convolutional neural network-based thermal spraying method according to an embodiment of the present invention will be described with reference to the accompanying drawings.
FIG. 1 is a flow chart of a convolutional neural network based thermal spray method of one embodiment of the present invention.
As shown in fig. 1, in the thermal spraying process, the optimization of the spraying process parameters is complicated and tedious, and the corresponding process parameters can only be manually searched for aiming at the target coating with ideal requirements. In order to improve the spraying work efficiency and save the test cost, on the basis of accumulating a large amount of experimental or simulation data, a convolutional neural network model is used for modeling the reverse process of thermal spraying, and an incidence relation model of the coating microstructure, the flight particle state and the spraying process parameters is established. Modeling the relationship between the state distribution of the flying particles and the spraying process parameters in the reverse process of thermal spraying is taken as an example, and a convolution extending network model connecting the state distribution of the flying particles and the spraying process parameters is constructed, so that the corresponding process parameters are predicted according to the required state distribution of the particles. The method comprises the following specific steps: the thermal spraying method based on the convolutional neural network comprises the following steps:
in step S1, experimental data of the thermal spray process is acquired.
That is, as shown in fig. 2, step S1 is data acquisition. Specifically, data accumulated in the thermal spraying process, namely spraying process parameters, flight particle state parameters and coating parameters, are collected by a design experiment or numerical simulation method. The spraying process parameters comprise spray gun current, voltage, main gas flow, auxiliary gas flow, powder feeding rate, carrier gas flow, spraying distance, spray gun moving speed, powder particle size distribution and the like; flight particle state parameters include position, temperature, velocity, size, etc.; the microstructure of the coating comprises microstructure pictures and the like; coating performance parameters include porosity, thermal conductivity, lifetime, bond strength, hardness, and the like.
In step S2, the experimental data is preprocessed to build a database.
That is, as shown in fig. 2, step S2 is a database construction. Specifically, in this step, during the training of the convolutional neural network model, the spraying process parameters need to be normalized, that is, the obtained raw data needs to be processed, so as to eliminate the influence caused by the difference of the magnitude of the order, and then the preprocessed data is reasonably divided into a training set and a test set, so that the model can be trained and tested later, and the method is better suitable for the convolutional neural network model.
It can be understood that, as shown in fig. 3, the flight particle state parameters have a certain spatial distribution rule, and when the flight particle state parameters are used as an input layer of a convolutional neural network, the acquired particle state data needs to be sorted into two-dimensional structure data similar to pictures. As shown in fig. 4, a relation model of the coating microstructure and the spraying process parameters can be constructed as an input layer of the convolutional neural network.
In step S3, the convolutional neural network model is designed according to preset requirements.
In other words, as shown in fig. 2, step S3 is to design or migrate a convolutional neural network model, and combine the above characteristics of the obtained data to design a specific architecture of the convolutional neural network, where, as shown in fig. 5, a flight particle state distribution diagram or a coating microstructure diagram is used as an input layer of the model, and a spraying process parameter is used as an output layer of the model;
it should be noted that the architecture of the convolutional neural network may be designed by itself, or the trained convolutional neural network model in other cases may be combined with the problem to perform fine tuning and application in different migration manners.
In step S4, the database is divided into a training set and a test set, wherein the training set is used to train the convolutional neural network model, and the test set is used to test the prediction accuracy of the trained convolutional neural network model.
In step S5, it is determined whether the training error in the training process and the testing error in the testing process are within the preset range, and if the training error and the testing error are outside the preset range, the steps S1-S5 are iteratively executed until the training error and the testing error meet the preset range.
In short, as shown in fig. 2, steps S4 and S5 are: training and testing the model, namely training and testing the constructed convolutional neural network model by utilizing the divided training set and test set; and judging whether the training error and the testing error of the convolutional neural network model are within a set acceptable range.
Further, if the error is within an acceptable range, the obtained convolutional neural network model is proved to have the required target in applicability and prediction precision, and can be directly applied to the subsequent spraying process parameter prediction and optimization process.
If the error is beyond the set range, further adjustment is needed to the convolutional neural network model, such as: increasing the data volume in the database, and expanding a training set and a test set; or fine tuning the architecture of the convolutional neural model and changing the migration mode of the convolutional neural network model; or increasing the training iteration times of the convolutional neural network model; and training and testing the obtained convolutional neural network model again by combining different methods for adjusting the parameters of the convolutional neural network model until the convolutional neural network model conforms to the preset range.
In step S6, a demand target is designed, and ideal target data of the demand target is acquired.
In step S7, the ideal target data is input to the tested convolutional neural network model, and the corresponding spraying process parameters are output.
In step S8, the spray process parameters are combined with the experimental data to obtain the actual flying particle feature distribution or coating microstructure.
Specifically, as shown in fig. 2, the obtained convolutional neural network model is used for modeling a reverse process of the spraying process, and predicting process parameters required by the spraying process for the input target flight particle state distribution or the coating microstructure. Designing a demand target, acquiring required ideal target data, and obtaining an ideal flight particle state distribution diagram or a coating microstructure diagram by combining means such as experimental data analysis or numerical model construction. And inputting the obtained ideal target value into a convolutional neural network model after training and testing are finished, and outputting a result by the model to obtain corresponding spraying process parameters. After the convolutional neural network model predicts the spraying process parameters, corresponding experiments are designed or corresponding examples are calculated in numerical simulation, and then an actual flight particle state distribution diagram or a coating microstructure diagram corresponding to the rational target values can be obtained.
According to the thermal spraying method based on the convolutional neural network, provided by the embodiment of the invention, the convolutional neural network model is utilized to perform modeling analysis on the reverse process of the spraying process, a database and a specific framework of the convolutional neural network model are firstly constructed, and the trained and tested convolutional neural network model is obtained after the training and testing errors of the model reach a certain range. In the application process of the model, the obtained convolutional neural network model is applied to the process parameter prediction and optimization in the spraying process; the ideal flying particle state parameter or coating microstructure is designed, the spraying process parameter is predicted by utilizing the convolutional neural network model, and then the actual flying particle state parameter or coating microstructure can be obtained by utilizing experiments or simulation.
Next, a thermal spraying apparatus based on a convolutional neural network proposed according to an embodiment of the present invention will be described with reference to the accompanying drawings.
Fig. 6 is a schematic structural diagram of a thermal spraying device based on a convolutional neural network according to an embodiment of the present invention.
As shown in fig. 6, the apparatus 60 includes: an obtaining module 601, a preprocessing module 602, a first design module 603, a training test module 604, a judgment iteration module 605, a second design module 606, a prediction module 607 and an actual result obtaining module 608.
The obtaining module 601 is configured to obtain experimental data of a thermal spraying process. The preprocessing module 602 is used for preprocessing the experimental data to construct a database. The first design module 603 is configured to design a convolutional neural network model according to a preset requirement. The training and testing module 604 is configured to divide the database into a training set and a testing set, where the training set is used to train the convolutional neural network model, and the testing set is used to test the prediction accuracy of the trained convolutional neural network model. The iteration determining module 605 is configured to determine whether a training error in the training process and a testing error in the testing process are within a preset range, and if the training error and the testing error are outside the preset range, iteratively execute the above modules until the training error and the testing error meet the preset range. The second design module 606 is used to design the demand targets and obtain ideal target data for the demand targets. The prediction module 607 is used for inputting the ideal target data to the tested convolutional neural network model and outputting the corresponding spraying process parameters. The get actual results module 608 is used to combine the spray process parameters with the experimental data to get the actual fly particle feature distribution or coating microstructure.
Further, in one embodiment of the present invention, the experimental data are spray process parameters, flight particle status parameters, and coating parameters accumulated during the thermal spray process.
Further, in one embodiment of the present invention, the first design module convolution neural network model has a flight particle feature distribution or coating microstructure as an input layer and spray coating process parameters as an output layer.
Further, in an embodiment of the present invention, if the training error and the test error exceed the preset ranges, the database is added to expand the training set and the test set, or the architecture of the convolutional neural network model is trimmed, the migration manner of the convolutional neural network model is changed, or the number of training iterations of the convolutional neural network model is increased, and the convolutional neural network model is further adjusted until the error is within the error range.
According to the thermal spraying device based on the convolutional neural network, provided by the embodiment of the invention, the convolutional neural network model is utilized to perform modeling analysis on the reverse process of the spraying process, the specific framework of the database and the convolutional neural network model is firstly constructed, and the trained and tested convolutional neural network model is obtained after the training and testing errors of the model reach a certain range. In the application process of the model, the obtained convolutional neural network model is applied to the process parameter prediction and optimization in the spraying process; the ideal flying particle state parameter or coating microstructure is designed, the spraying process parameter is predicted by utilizing the convolutional neural network model, and then the actual flying particle state parameter or coating microstructure can be obtained by utilizing experiments or simulation.
Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of the present invention, "a plurality" means at least two, e.g., two, three, etc., unless specifically limited otherwise.
In the description herein, references to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above are not necessarily intended to refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, various embodiments or examples and features of different embodiments or examples described in this specification can be combined and combined by one skilled in the art without contradiction.
Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention, and that variations, modifications, substitutions and alterations can be made to the above embodiments by those of ordinary skill in the art within the scope of the present invention.
Claims (8)
1. A thermal spraying method based on a convolutional neural network is characterized by comprising the following steps:
step S1, acquiring experimental data of the thermal spraying process;
step S2, preprocessing the experimental data to construct a database;
step S3, designing a convolutional neural network model according to preset requirements;
step S4, dividing the database into a training set and a testing set, wherein the training set is used for training the convolutional neural network model, and the testing set is used for testing the prediction accuracy of the trained convolutional neural network model;
step S5, judging whether the training error in the training process and the testing error in the testing process are in a preset range, if the training error and the testing error are beyond the preset range, iteratively executing the steps S1-S5 until the training error and the testing error meet the preset range;
step S6, designing a demand target, and acquiring ideal target data of the demand target;
step S7, inputting the ideal target data into a tested convolutional neural network model, and outputting corresponding spraying process parameters;
and step S8, combining the spraying process parameters with the experimental data to obtain the actual flying particle characteristic distribution or coating microstructure.
2. The convolutional neural network-based thermal spray method of claim 1, wherein the experimental data are spray process parameters, flight particle status parameters, and coating parameters accumulated during thermal spray.
3. The convolutional neural network based thermal spraying method of claim 1, wherein in the convolutional neural network model in step S3, the flying particle feature distribution or coating microstructure is used as an input layer, and the spraying process parameters are used as an output layer.
4. The convolutional neural network based thermal spray coating method of claim 1 wherein if the training error and the testing error are out of the preset range, the database is increased, the training set and the testing set are expanded, or the architecture of the convolutional neural network model is fine-tuned, the migration mode of the convolutional neural network model is changed, or the number of training iterations of the convolutional neural network model is increased, and the convolutional neural network model is further adjusted until the error is within the error range.
5. A convolutional neural network based thermal spray device comprising the steps of:
the acquisition module is used for acquiring experimental data of the thermal spraying process;
the preprocessing module is used for preprocessing the experimental data to construct a database;
the first design module is used for designing a convolutional neural network model according to a preset requirement;
the training and testing module is used for dividing the database into a training set and a testing set, wherein the training set is used for training the convolutional neural network model, and the testing set is used for testing the prediction accuracy of the trained convolutional neural network model;
the judging iteration module is used for judging whether the training error in the training process and the testing error in the testing process are within a preset range or not, and if the training error and the testing error exceed the preset range, the judging iteration module executes the modules in an iteration mode until the training error and the testing error meet the preset range;
the second design module is used for designing a demand target and acquiring ideal target data of the demand target;
the prediction module is used for inputting the ideal target data into a tested convolutional neural network model and outputting corresponding spraying process parameters;
and the actual result obtaining module is used for combining the spraying process parameters with the experimental data to obtain the actual flying particle characteristic distribution or coating microstructure.
6. A thermal spray device based on a convolutional neural network as claimed in claim 5, wherein the experimental data are spray process parameters, flight particle status parameters and coating parameters accumulated during thermal spray.
7. A convolutional neural network based thermal spraying device as claimed in claim 5 wherein the convolutional neural network model of the first design module has the flying particle feature distribution or coating microstructure as an input layer and the spraying process parameters as an output layer.
8. The convolutional neural network based thermal spraying device as claimed in claim 5, wherein if the training error and the testing error exceed the preset ranges, the database is added to expand the training set and the testing set, or the architecture of the convolutional neural network model is fine-tuned, the migration manner of the convolutional neural network model is changed, or the number of training iterations of the convolutional neural network model is increased, and the convolutional neural network model is further adjusted until the errors are within the error ranges.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011073655.3A CN112329322A (en) | 2020-10-09 | 2020-10-09 | Thermal spraying method and device based on convolutional neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011073655.3A CN112329322A (en) | 2020-10-09 | 2020-10-09 | Thermal spraying method and device based on convolutional neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112329322A true CN112329322A (en) | 2021-02-05 |
Family
ID=74314706
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011073655.3A Pending CN112329322A (en) | 2020-10-09 | 2020-10-09 | Thermal spraying method and device based on convolutional neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112329322A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113051820A (en) * | 2021-03-24 | 2021-06-29 | 中国空气动力研究与发展中心超高速空气动力研究所 | Cross-basin pneumatic parameter simulation method based on convolutional neural network |
| CN115828574A (en) * | 2022-11-28 | 2023-03-21 | 江苏凯威特斯半导体科技有限公司 | Plasma spraying parameter determination method |
| CN117195743A (en) * | 2023-10-16 | 2023-12-08 | 西安交通大学 | Spraying parameter optimization method for crack structure of thermal barrier coating |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109338354A (en) * | 2018-09-30 | 2019-02-15 | 武汉大学 | A kind of thin-wall construction ultrafast pulsed laser-cold spraying method of surface finish |
| CN109513557A (en) * | 2018-12-27 | 2019-03-26 | 海安科大机器人科技有限公司 | A kind of robot autonomous spraying method of ship segment spray painting of view-based access control model guidance |
| US20200257933A1 (en) * | 2019-02-05 | 2020-08-13 | Imagars Llc | Machine Learning to Accelerate Alloy Design |
-
2020
- 2020-10-09 CN CN202011073655.3A patent/CN112329322A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109338354A (en) * | 2018-09-30 | 2019-02-15 | 武汉大学 | A kind of thin-wall construction ultrafast pulsed laser-cold spraying method of surface finish |
| CN109513557A (en) * | 2018-12-27 | 2019-03-26 | 海安科大机器人科技有限公司 | A kind of robot autonomous spraying method of ship segment spray painting of view-based access control model guidance |
| US20200257933A1 (en) * | 2019-02-05 | 2020-08-13 | Imagars Llc | Machine Learning to Accelerate Alloy Design |
Non-Patent Citations (1)
| Title |
|---|
| JINWEI ZHU 等: "prediction of control parameters corresponding to in-flight particles in atmospheric plasma spray employing convolutional neural networks", 《SURFACE AND COATINGS TECHNOLOGY》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113051820A (en) * | 2021-03-24 | 2021-06-29 | 中国空气动力研究与发展中心超高速空气动力研究所 | Cross-basin pneumatic parameter simulation method based on convolutional neural network |
| CN115828574A (en) * | 2022-11-28 | 2023-03-21 | 江苏凯威特斯半导体科技有限公司 | Plasma spraying parameter determination method |
| CN115828574B (en) * | 2022-11-28 | 2023-09-26 | 江苏凯威特斯半导体科技有限公司 | Plasma spraying parameter determination method |
| CN117195743A (en) * | 2023-10-16 | 2023-12-08 | 西安交通大学 | Spraying parameter optimization method for crack structure of thermal barrier coating |
| CN117195743B (en) * | 2023-10-16 | 2024-06-04 | 西安交通大学 | Spraying parameter optimization method for crack structure of thermal barrier coating |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112329322A (en) | Thermal spraying method and device based on convolutional neural network | |
| CN107992648B (en) | Adaptive RBF neural network algorithm for estimating thrust of aircraft engine | |
| Ming et al. | Soft computing models and intelligent optimization system in electro-discharge machining of SiC/Al composites | |
| CN108573116B (en) | Aero-engine transition state thrust estimation method based on long-time and short-time memory network | |
| CN104715142B (en) | A kind of station boiler NOxDischarge dynamic soft-measuring method | |
| CN109338354A (en) | A kind of thin-wall construction ultrafast pulsed laser-cold spraying method of surface finish | |
| CN112433507A (en) | LSO-LSSVM (least squares support vector machine) -based five-axis numerical control machine tool thermal error comprehensive modeling method | |
| Ingole et al. | Build orientation analysis for minimum cost determination in FDM | |
| CN112131670B (en) | Aero-engine model iterative algorithm based on hybrid adaptive differential evolution | |
| D’Emilia et al. | The role of measurement and simulation in additive manufacturing within the frame of Industry 4.0 | |
| JP2008123354A (en) | Temperature controller, temperature control method, and temperature control program | |
| CN116377370A (en) | Plasma spraying control system and method based on digital twin | |
| CN109023209A (en) | A kind of heat spraying method and system based on artificial intelligence | |
| Kolahan et al. | Modeling and optimization of process parameters in PMEDM by genetic algorithm | |
| Rao et al. | PROCESS PARAMETERS MODELING AND OPTIMIZATION OF WIRE ELECTRIC DISCHARGE MACHINING. | |
| Aboutaleb et al. | Systematic optimization of Laser-based Additive Manufacturing for multiple mechanical properties | |
| CN115659529A (en) | Centrifugal compressor impeller redesign method comprehensively considering machining precision and aerodynamic performance | |
| CN109446721A (en) | Lathe technique interactive algorithm based on the arrangement of Identifier Software thread execution order | |
| Cai et al. | Application research and improvement of particle swarm optimization algorithm | |
| Sharma et al. | Increased reliability of gas turbine components by robust coatings manufacturing | |
| CN110814342A (en) | Preparation method of computer-formalized 3D printing metal material | |
| Bushra et al. | Orientation optimization in additive manufacturing: evaluation of recent trends | |
| CN111522226A (en) | Multi-objective optimization high-type PID optimal controller design method for servo turntable | |
| Stockton et al. | Cost model development using virtual manufacturing and data mining: part I—methodology development | |
| Niesłony et al. | Optimization procedures for machining operations on CNC machine tools |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210205 |