CN115730520A - Shape prediction and virtual simulation method and system for welding seam - Google Patents
Shape prediction and virtual simulation method and system for welding seam Download PDFInfo
- Publication number
- CN115730520A CN115730520A CN202211478685.1A CN202211478685A CN115730520A CN 115730520 A CN115730520 A CN 115730520A CN 202211478685 A CN202211478685 A CN 202211478685A CN 115730520 A CN115730520 A CN 115730520A
- Authority
- CN
- China
- Prior art keywords
- weld
- welding
- virtual simulation
- width
- welding seam
- 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
Landscapes
- Arc Welding In General (AREA)
Abstract
The invention discloses a method and a system for shape prediction and virtual simulation of a welding seam, wherein the method comprises the steps of collecting technological parameters of gas shielded welding and standardizing; inputting the standardized process parameters into a weld shape prediction neural network model to obtain the weld penetration and weld penetration; the method for virtually simulating the appearance of the welding seam comprises the following steps of according to the geometric structure parameters and the fusion width of the welding beadY w And penetration depthY d Generating a virtual simulation model diagram of the cross-sectional profile of the weld bead and a virtual simulation model diagram of the surface profile of the weld bead according to the weld widthY w And penetration depthY d Generating a fitted curve to represent the top profile of the weld beadThe curve, which represents the weld pool unit with ellipses stacked in the welding direction. The method can realize accurate prediction of the shape of the welding seam of the gas shielded welding, and realize dynamic virtual simulation of the shape of the welding seam based on the result of the shape prediction of the welding seam.
Description
Technical Field
The invention relates to CO 2 A virtual simulation technology for the shape of a welding seam of gas shielded welding, in particular to a shape prediction and virtual simulation method and system for the welding seam.
Background
Currently, the industrial technology revolution centered on digitization, networking and intelligence is emerging, and the industrial system, development mode and competitive format around the world have a major turn. At present, the manufacturing industry has more and more demands on talents, and the welding technical talents are in the state of short supply. The culture of the welding technical talents usually requires a large amount of time and capital investment, but the existing traditional welding operation has the disadvantages of severe environment, harm to human health and high labor intensity, so that fewer and fewer welding talents are caused, and the culture cost of the welding talents is higher and higher. In order to save cost, improve manufacturing environment and solve various welding problems in real life, a welding virtual simulation technology based on digitization, networking and intelligence is developed and becomes a research hotspot in the field of welding training. The training mode combining the welding virtual simulation technology and the welding training has wide application prospect and market value in the aspect of welding talent culture. However, the research on the welding pool shape simulation in the development of the existing virtual welding system is mainly based on post-welding static simulation, and the dynamic simulation modeling of the weld forming cannot be realized due to the lack of real-time dynamic interaction.
Disclosure of Invention
The technical problems to be solved by the invention are as follows: aiming at the problems in the prior art, the invention provides a welded seam shape prediction method, aiming at realizing accurate prediction of the welded seam shape of gas shielded welding and providing a data basis for virtual simulation of the welded seam shape; the invention also provides a virtual simulation method of the appearance of the welding seam, which is used for realizing dynamic virtual simulation of the appearance of the welding seam based on the result of the appearance prediction of the welding seam.
In order to solve the technical problems, the invention adopts the technical scheme that:
a method for predicting the appearance of a weld seam, comprising:
s101, collecting technological parameters of gas shielded welding, including welding voltage, welding gun linear velocity, gas flow and welding gun walking angle;
s102, standardizing the process parameters;
s103, inputting the standardized process parameters into a weld shape prediction model to obtain the weld penetration and the weld penetration of the weld.
Optionally, the weld morphology prediction model in step S103 includes two multi-layer perceptron neural network models respectively used for predicting fusion width and fusion depth, the multi-layer perceptron neural network model includes an input layer, a hidden layer and an output layer, the input layer is used for inputting each standardized process parameter, and a function expression of the hidden layer is as follows:
in the above formula, the first and second carbon atoms are,h j is the output of the hidden layer(s),ω ij for input layer neuronsiTo hidden layer neuronsjThe connection weight of (a) is set,b j being hidden layer neuronsjThe bias of (a) is such that,x i as input layer neuronsiThe process parameters are input into the device and then,nis the number of input layer neurons; hidden layer neuron adopts sigmoid function as excitation function to train and test, and outputsThe function expression of the layer is as follows:
in the above formula, the first and second carbon atoms are,yis the output of the output layer or layers,mthe number of hidden layer neurons is the number of neurons,h j is the output of the hidden layer(s),ω j the connection weights for hidden layer neuron j to the output layer,b1 is the bias of the output layer.
Optionally, step S103 is preceded by the step of training the multi-layer perceptron neural network model:
s201, designing an orthogonal experiment containing multiple horizontal grade values aiming at each process parameter, forming a welding seam sample, and respectively collecting the process parameters of gas shielded welding of the welding seam sample and the fusion width and fusion depth of the corresponding welding seam;
s202, standardizing the process parameters, and adding corresponding weld width and weld depth labels as training samples;
s203, training the two multilayer perceptron neural network models respectively by adopting the training samples until the prediction accuracy of the multilayer perceptron neural network models on the weld width or the weld penetration is larger than a set value or the training times are equal to the set maximum times.
In addition, the invention also provides a system for modeling and virtually simulating the appearance of the welding seam, which comprises a microprocessor and a memory which are connected with each other, wherein the microprocessor is programmed or configured to execute the appearance prediction method of the welding seam.
Furthermore, the present invention also provides a computer-readable storage medium having stored therein a computer program for being programmed or configured by a microprocessor to perform a method of topography prediction of the weld bead.
In addition, the invention also provides a virtual simulation method of the appearance of the welding seam, which comprises the following steps:
s301, predicting the weld width of the weld by adopting the shape prediction method of the welding weldY w And penetration depthY d ；
S302, according to geometric structural parameters of a welding bead and the weld width of the welding beadY w And penetration depth of fusionY d Generating a virtual simulation model diagram of the welding bead section appearance according to the geometric structure parameters of the welding bead and the fusion width of the welding beadY w And penetration depth of fusionY d The resulting fitted curve represents the top profile curve ADB of the weld bead.
Optionally, in step S302, the weld width is determined according to the geometric parameters of the weld and the weld widthY w And penetration depthY d The generated fitting curve is a cosine curve, and the function expression is as follows:
in the above-mentioned formula, the compound has the following structure,ybeing corresponding points on the cosine curveyThe coordinates of the axes are set to be,hthe thickness of the blunt edge in the geometric parameters of the weld,ras a butt gap in the geometric parameters of the weld,θis the groove angle in the geometric parameters of the welding seam,Y w the width of the weld is the width of the weld,Y d the penetration of the welding line is the penetration of the welding line,xbeing corresponding points on the cosine curvexAxis coordinates.
Optionally, step S302 further includes determining the geometric parameters of the weld bead and the weld widthY w Generating a virtual simulation model diagram of the surface appearance of the welding bead, wherein the virtual simulation model diagram of the surface appearance of the welding bead adopts ellipses stacked along the welding direction to represent molten pool units stacked along the welding direction, and the major axis of the ellipses is the melt widthY w Double, minor axis size equals fusion widthY w 。
In addition, the invention also provides a system for modeling and virtually simulating the appearance of the welding seam, which comprises a microprocessor and a memory which are connected with each other, wherein the microprocessor is programmed or configured to execute the method for virtually simulating the appearance of the welding seam.
Furthermore, the invention also provides a computer-readable storage medium, in which a computer program is stored, the computer program being programmed or configured by a microprocessor to perform a method of topographical virtual simulation of the weld bead.
Compared with the prior art, the invention mainly has the following advantages:
1. the shape prediction method of the welding seam can realize accurate prediction of the shape of the welding seam of the gas shielded welding and provide a data basis for virtual simulation of the shape of the welding seam.
2. The shape virtual simulation method of the welding seam can realize dynamic shape virtual simulation of the welding seam based on the shape prediction result of the welding seam, realize real-time prediction and dynamic interaction of welding seam formation, and solve the problem that the existing welding virtual system is static simulation or not shown in the dynamic behavior representation of a molten pool.
Drawings
FIG. 1 is a flowchart of a weld profile prediction method according to an embodiment of the present invention.
Fig. 2 is a schematic diagram of an application principle of a multi-layer perceptron neural network model in the embodiment of the present invention.
Fig. 3 is a schematic structural diagram of a multi-layer perceptron neural network model according to an embodiment of the present invention.
FIG. 4 is a fitting curve of the experimental value and the predicted value of the melt width in the example of the present invention.
FIG. 5 is a curve fitting the penetration-test value to the predicted value in the example of the present invention.
Fig. 6 is a virtual simulation model diagram of the weld bead cross-sectional profile obtained in the embodiment of the present invention.
Fig. 7 is a virtual simulation model diagram of the bead surface topography obtained in the embodiment of the present invention.
Detailed Description
As shown in fig. 1, the method for predicting the shape of the weld seam in the embodiment includes:
s101, collecting technological parameters of gas shielded welding, including welding voltage, welding gun linear velocity, gas flow and welding gun walking angle;
s102, standardizing the process parameters;
s103, inputting the standardized process parameters into a weld shape prediction neural network model to obtain the weld penetration and the weld penetration.
As an optional implementation manner, in this embodiment, the material of the weldment selected for gas shielded welding is Q235 steel, the thickness of a single piece is 6mm, the width is 40mm, the length is 100mm, the bevel angle is 30 degrees, the truncated edge is 0.5mm, and the butt joint gap of the root is 1mm. The welding wire is H08Mn2SiA and the diameter is 1mm. A V-shaped groove butt joint experiment of gas shielded welding is carried out by adopting a Foniss welding machine (model number TransPlus Synergic 5000).
It should be noted that the gas shielded welding herein refers to gas shielded metal arc welding, and mainly includes CO 2 Gas shielded welding and argon shielded welding, in this example with CO 2 Gas shielded welding is described as an example. CO 2 2 The welding process of gas shielded welding involves many factors, such as welding parameters, welding process, welding materials, etc., and all factors are mutually coupled under the combined action to form a more complex influence relationship. Under the condition that a welding process and welding materials are determined, welding process parameters are important factors influencing the forming quality and the geometric appearance of a welding seam, and the process parameters mainly influencing the geometric appearance of the welding seam are respectively welding current, welding voltage, linear velocity of a welding gun, gas flow, walking angle of the welding gun, the extending length of the welding wire, wire feeding speed and the like. To explore CO 2 According to the law of influence of welding process parameters of gas shielded welding on welding seam forming and welding quality, a pre-test is firstly carried out, and the fact that a better welding seam forming effect and welding quality can be obtained when the wire feeding speed is 3.5mm/s is found through the pre-test. In order to ensure good weld forming effect and welding quality, the wire feeding speed is set to be 3.5mm/s when a welding test is carried out. According to the welding characteristics of the Fonness welding machine, under the condition that the wire feeding speed is not changed, the equipment can automatically match the adaptive welding current; meanwhile, the extension length of the welding wire is mainly controlled by the wire feeding speed, so that the extension length of the welding wire is kept unchanged under the condition of a certain wire feeding speed. Therefore, in the present embodiment, mainly the welding voltage, the linear velocity of the welding torch, the gas flow rate, and the travel of the welding torch are consideredThe four process parameters of the angle have regular influence on the formation and welding quality of the welding seam, and other process parameters such as welding current, wire feeding speed, welding wire extension length and the like are kept unchanged. Accordingly, the process parameters in step S101 include welding voltage, torch linear velocity, gas flow rate, and torch travel angle.
In the V-groove butt welding, in order to control deformation and weld seam forming quality and prevent poor conditions such as penetration or incomplete penetration, it is often necessary to perform single pass backing welding first and then perform filling welding. The penetration and the penetration are not only important indexes for measuring the backing welding quality, but also important geometric parameters for influencing the section appearance of the welding seam, and the size of the penetration and the penetration directly influences the section appearance of the welding seam and the subsequent welding quality. Therefore, the present embodiment mainly examines two kinds of morphology features of weld width and weld depth.
When the sample data has a large difference in magnitude, the prediction error of the network becomes large, and in order to avoid this, the sample data is usually preprocessed, i.e., normalized, before the network training. Through data standardization, sample data is converted according to a certain proportion, so that the sample data falls into a small specific interval, such as an interval of 0~1 or-1~1. In this embodiment, a range standardization method is adopted to perform data standardization on a training sample (welding process parameter), and a function expression for standardizing the process parameter in step S102 in this embodiment is as follows:
in the above formula, the first and second carbon atoms are,x' is a process parameterxThe corresponding process parameters after the standardization are carried out,x min andx max are respectively the process parametersxCorresponding minimum and maximum values; in addition, other normalization methods such as standard deviation normalization, Z-Score, etc. may be used as required, and are not described in detail herein.
Due to CO 2 The forming appearance of the welding seam of the gas shielded welding is subjected to welding attitude, welding voltage, welding speed, gas flow and the likeThe influence of a plurality of process parameters and the interaction among the process parameters also influence the forming appearance of the welding seam, so that the welding process has the characteristics of high nonlinearity, uncertainty and the like, and the performance prediction of the welding process is extremely difficult to realize by adopting the traditional mathematical modeling method. In recent years, with the deep and development of artificial neural network research, the artificial neural network has great advantages in solving the system problems of high nonlinearity and serious uncertainty, and opens up a new path for modeling of the welding process. Therefore, in this embodiment, a machine learning intelligent algorithm is applied, a multi-layer perceptron neural network model between welding process parameters and weld morphology is established to serve as a weld morphology prediction neural network model, as shown in fig. 2, experimental data is adopted in advance to perform optimization training on the multi-layer perceptron neural network model, and after training is completed, test data is input into the trained multi-layer perceptron neural network model to perform weld morphology prediction. As shown in fig. 3, the neural network model for predicting weld seam topography in step S103 in this embodiment includes two neural network models of a multi-layer perceptron, which are respectively used for predicting weld width and weld depth, the neural network model of the multi-layer perceptron includes an input layer, a hidden layer and an output layer, the input layer is used for inputting each standardized process parameter (welding voltage, welding gun linear velocity, gas flow and welding gun walking angle), and a function expression of the hidden layer is as follows:
in the above formula, the first and second carbon atoms are,h j is the output of the hidden layer(s),ω ij for input layer neuronsiNeurons reaching the hidden layerjThe connection weight of (a) is set,b j being hidden layer neuronjThe bias of (a) is such that,x i for input layer neuronsiThe process parameters are input into the device and then are processed,nis the number of input layer neurons; the hidden layer neuron adopts a sigmoid function as a stimulus function to train and test, wherein the function expression of the sigmoid function is as follows:
in the above formula, the first and second carbon atoms are,f(σ) In order to be a sigmoid function,σis the input quantity of the sigmoid function; the functional expression of the output layer is:
in the above formula, the first and second carbon atoms are,yis the output of the output layer or layers,mthe number of hidden layer neurons is the number of neurons,h j is the output of the hidden layer(s),ω j the connection weights for hidden layer neuron j to the output layer,b1 is the bias of the output layer. In addition, a neural network model of two output neurons can be adopted according to needs, and the weld penetration of the weld can be obtained through one-step output.
In this embodiment, before step S103, the method further includes a step of training a multi-layer perceptron neural network model:
s201, designing an orthogonal experiment containing multiple horizontal grade values aiming at each process parameter, forming a welding seam sample, and respectively collecting CO of the welding seam sample 2 The technological parameters of the gas shielded welding and the fusion width and the fusion depth of the corresponding welding line;
s202, standardizing the process parameters, and adding corresponding weld width and weld depth labels as training samples;
s203, training the two multilayer perceptron neural network models respectively by using the training samples until the prediction accuracy of the multilayer perceptron neural network models on the weld width or the weld depth is more than a set value or the training times are equal to the set maximum times.
In this embodiment, for the weld sample in step S201, 25 sets of orthogonal experiments with 4 factors (welding voltage, welding torch linear velocity, gas flow, and welding torch traveling angle) and 5 levels are designed, as shown in table 1:
table 1: and (4) an orthogonal experimental table.
In order to investigate the weld width and the weld depth, a weld bead section of an area with a good weld shape is cut, polished to 3000 meshes by using sand paper, then corroded by using aqua regia solution, and finally the weld bead section shape is observed by using a microscope, and the weld width Yw and the weld depth Yd are measured for multiple times to obtain an average value, and the result is retained with 3 decimal places as shown in Table 2.
Table 2: and (5) orthogonal experimental results.
The acquired test sample is used for training and testing the multi-layer perceptron neural network, and the influence primary and secondary and the mutual relation of all factors are not researched, so that the sample data is not analyzed in detail. And obtaining a sample of the neural network training of the multilayer perceptron through an orthogonal test, and using the sample to construct a weld geometrical morphology prediction model and a weld morphology virtual simulation model.
In this embodiment, the trained multi-layer perceptron neural network model is used to predict 25 sets of process parameters, and fitting curves of the experimental values and the predicted values are drawn as shown in fig. 4 and 5, and it can be visually seen from fig. 4 and 5 that the predicted values and the experimental values of the model are almost completely fitted (the two curves are basically overlapped, and only a few data points have slight differences). The maximum deviation of the melt width prediction (deviation = predicted value — experimental value) was calculated to be 0.097, and the model goodness of fit (variance score of the explained regression model) was 0.999269; the maximum deviation of the penetration prediction is 0.051, and the model fitting goodness is 0.999567. Therefore, the multi-layer perceptron neural network model is high in fitting goodness and accurate in model prediction, and can be used for weld morphology prediction and virtual simulation model establishment.
In addition, the embodiment also provides a system for modeling and virtually simulating the appearance of the welding seam, which comprises a microprocessor and a memory which are connected with each other, wherein the microprocessor is programmed or configured to execute the appearance prediction method of the welding seam. In addition, the present embodiment also provides a computer readable storage medium, in which a computer program is stored, the computer program being programmed or configured by a microprocessor to execute the above-mentioned method for predicting the profile of the weld seam.
In addition, the embodiment also provides a virtual simulation method for the appearance of the welding seam, which includes:
s301, predicting the weld width of the weld by adopting the shape prediction method of the welding weldY w And penetration depthY d ；
S302, according to the geometric structure parameters of the welding bead and the fusion width of the welding seamY w And penetration depthY d Generating a virtual simulation model diagram of the welding bead cross-sectional shape according to the geometric structure parameters of the welding bead and the fusion width of the welding beadY w And penetration depth of fusionY d The resulting fitted curve represents the top profile curve ADB of the weld bead.
In the welding process of a V-shaped groove butt joint experiment, molten drops fall into a welding area under the action of gravity, a welding bead is formed after cooling, and it can be seen from FIG. 4 that the geometric shape of the cross section of the welding bead mainly depends on the contour shapes of the side wall, the bottom and the top, wherein the geometric shapes of the side wall and the bottom of the welding bead can be determined according to the size of a substrate, the angle of a groove and a butt joint gap, and the top of the welding bead presents an upwards-convex arc shape or a downwards-concave arc shape, so that a proper mathematical model needs to be established to fit the contour of the top of the welding bead. In step S302, the weld is determined according to the geometric parameters and the weld widthY w And penetration depthY d The generated fitting curve is a cosine curve, and the function expression is as follows:
in the above-mentioned formula, the compound has the following structure,ybeing corresponding points on the cosine curveyThe coordinates of the axes are set to be,hthe thickness of the blunt edge in the geometric parameters of the weld,ras a butt gap in the geometric parameters of the weld,θis the groove angle in the geometric parameters of the welding seam,Y w in order to obtain the weld width of the weld,Y d in order to realize the penetration of the welding seam,xbeing corresponding points on the cosine curvexThe axis coordinate and the cosine curve can simultaneously meet two section shapes of 'convex arc' and 'concave arc'.
The derivation of the above equation is as follows:
firstly, establishing a cosine model of a fitting curve:
in the above-mentioned formula, the compound has the following structure,ybeing corresponding points on the cosine curveyThe coordinates of the axes are set to be,a,b,cas a function of the number of the coefficients,xbeing corresponding points on the cosine curvexAxis coordinates; then, according to the geometric parameters of the welding seam and the fusion width of the welding seamY w And penetration depthY d The geometric relationship between the generated fitting curves can be used to obtain the coefficientsa,b,cComprises the following steps:
substitution coefficienta,b,cThe weld width according to the welding seam can be obtainedY w And penetration depth of fusionY d And (4) generating a functional expression of a fitting curve (cosine curve). During virtual simulation modeling, the contour equation of the welding seam needs to be deduced by combining the parameters of penetration, fusion width and welding bead geometric structure to obtain a contour model. In this embodiment, the geometric parameters of the weld bead are V-shaped groove structures, and the dimension and angle of the groove structures are fixed, and the geometric parameters of the weld bead measured according to the size of the weldment are: truncated edge thickness h =0.5mm, butt joint gapr=1mm, bevel angleDegree of rotationθ=30 °. In addition, the penetration and the fusion width are also measured under the fixed groove, for example, the penetration is the distance from the upper surface of the welding seam to the bottom of the groove, and the fusion width is the width of the upper surface of the welding seam measured in the V-shaped groove. The virtual simulation model diagram of the welding bead cross-sectional profile finally generated in the embodiment is shown in fig. 6, in which a-Y w /2,0)、B(Y w /2,0) is the curve end point of the top profile curve ADB of the weld bead, D is the midpoint of the top profile curve ADB of the weld bead, and fig. 6 notes the weldment thicknessH1Is 6mm, the bevel angleθIs 30 degrees and has a blunt edge thicknessh0.5mm, root butt gaprIs 1mm, and therefore:
in this embodiment, step S302 further includes determining the geometric parameters of the weld bead and the weld widthY w Generating a virtual simulation model diagram of the surface appearance of the welding bead, wherein the virtual simulation model diagram of the surface appearance of the welding bead adopts ellipses stacked along the welding direction to represent a molten pool unit stacked along the welding direction, and the major axis of each ellipse is fusion widthY w Double, minor axis size equals fusion widthY w . The texture characteristics of the upper surface of the welding bead and the width of the welding bead can be represented by the upper surface of the welding bead, and in order to establish a virtual simulation model of the surface of the welding bead, the texture characteristics of the upper surface of the welding bead and the width of the welding bead are extracted to represent the geometric appearance of the surface of the welding bead. In the welding process, the weld surface topography forming process can be understood as the stacking of a plurality of molten pool units along the welding direction, and as the texture of the molten pool is similar to an elliptic curve, a virtual simulation model diagram of the weld bead surface topography as shown in fig. 7 is finally obtained, wherein d is the butt joint edge of the weld, e is the texture of the molten pool, f is the edge of the groove, g is the surface groove gap, and h is the root butt joint gap.
The embodiment further develops a welding seam shape prediction and virtual simulation system through Python programming, the system integrates the main functions of data retrieval, modification and deletion, new data entry, experimental data modeling, welding seam shape prediction and virtual simulation demonstration, user information maintenance and management and the like, and a theoretical model and a virtual simulation platform can be provided for welding practical training virtual simulation. The test data information and the prediction results corresponding to the set of test data can be displayed in real time in the simulation data demonstration module. The virtual simulation demonstration module for the weld appearance can dynamically demonstrate the weld appearance of virtual simulation based on test parameters, at the moment, the weld section appearance can change along with the test parameters, the weld surface appearance can synchronously change, and meanwhile, the position of the weld surface appearance can also change constantly. In the embodiment, the corresponding weld appearance geometric dimension prediction result and virtual simulation result are obtained through testing when the test number is 60, the welding gun walking angle is 71 degrees, the welding voltage is 22V, the welding gun linear velocity is 7.0 mm/s and the gas flow is 7L/min, and the corresponding weld appearance geometric dimension prediction result and virtual simulation result are obtained through testing when the test number is 87 degrees, the welding gun walking angle is 67 degrees, the welding voltage is 21V degrees, the welding gun linear velocity is 3.0 mm/s and the gas flow is 8L/min, and the weld appearance geometric dimension prediction results and virtual simulation results of the two test working conditions are compared, so that the change of the weld appearance under different test parameters can be found. The appearance prediction and virtual simulation system of the welding seam developed through Python programming in the embodiment integrates the main functions of experimental data modeling, seam appearance prediction, virtual simulation demonstration and the like, realizes real-time prediction and dynamic interaction of seam formation, solves the problem that the existing welding virtual system is static simulation or not displayed in the dynamic behavior representation of a molten pool, and provides a theoretical model and a virtual simulation platform for virtual simulation training of welding.
In addition, the embodiment also provides a system for modeling and virtually simulating the appearance of the welding seam, which comprises a microprocessor and a memory which are connected with each other, wherein the microprocessor is programmed or configured to execute the method for virtually simulating the appearance of the welding seam. In addition, the present embodiment also provides a computer-readable storage medium, in which a computer program is stored, where the computer program is programmed or configured by a microprocessor to execute the above-mentioned virtual simulation method for the profile of the welding seam.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-readable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein. The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks. These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks. These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The above description is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments, and all technical solutions belonging to the idea of the present invention belong to the protection scope of the present invention. It should be noted that modifications and embellishments within the scope of the invention may occur to those skilled in the art without departing from the principle of the invention, and are considered to be within the scope of the invention.
Claims (10)
1. A method for predicting the appearance of a welding seam is characterized by comprising the following steps:
s101, collecting technological parameters of gas shielded welding, including welding voltage, welding gun linear velocity, gas flow and welding gun walking angle;
s102, standardizing the process parameters;
s103, inputting the standardized process parameters into a weld shape prediction neural network model to obtain the weld penetration and the weld penetration.
2. The method for predicting the shape of the welding seam according to claim 1, wherein the neural network model for predicting the shape of the welding seam in step S103 comprises two neural network models of a multi-layer perceptron, which are respectively used for predicting the weld width and the weld depth, the neural network model of the multi-layer perceptron comprises an input layer, a hidden layer and an output layer, the input layer is used for inputting each standardized process parameter, and the function expression of the hidden layer is as follows:
in the above formula, the first and second carbon atoms are,h j is the output of the hidden layer(s),ω ij for input layer neuronsiNeurons reaching the hidden layerjThe connection weight of (a) is set,b j being hidden layer neuronjThe bias of (a) is such that,x i for input layer neuronsiThe process parameters are input into the device and then,nis the number of input layer neurons; hidden layer neurons using sigmoid function as excitation functionTraining and testing, wherein the function expression of the output layer is as follows:
in the above formula, the first and second carbon atoms are,yis the output of the output layer or layers,mthe number of hidden layer neurons is the number of neurons,h j is the output of the hidden layer(s),ω j the connection weights for hidden layer neuron j to the output layer,b1 is the bias of the output layer.
3. The method for predicting the morphology of the welding seam according to claim 2, characterized in that step S103 is preceded by the step of training a multi-layer perceptron neural network model:
s201, designing an orthogonal experiment containing multiple horizontal grade values aiming at each process parameter, forming a welding seam sample, and respectively collecting the process parameters of gas shielded welding of the welding seam sample and the fusion width and fusion depth of the corresponding welding seam;
s202, standardizing the process parameters, and adding corresponding weld width and weld depth labels as training samples;
s203, training the two multilayer perceptron neural network models respectively by using the training samples until the prediction accuracy of the multilayer perceptron neural network models on the weld width or the weld depth is more than a set value or the training times are equal to the set maximum times.
4. A system for topographical modeling and virtual simulation of a weld joint, comprising a microprocessor and a memory connected to one another, wherein said microprocessor is programmed or configured to perform a method for topographical prediction of a weld joint as claimed in any one of claims 1 to 3.
5. A computer-readable storage medium, in which a computer program is stored, wherein the computer program is adapted to be programmed or configured by a microprocessor to perform the method of topography prediction of a weld bead according to any one of claims 1 to 3.
6. A virtual simulation method for the appearance of a welding seam is characterized by comprising the following steps:
s301, predicting the weld width of the weld by using the shape prediction method of the welding weld according to any one of claims 1-3Y w And penetration depthY d ；
S302, according to geometric structural parameters of a welding bead and the weld width of the welding beadY w And penetration depthY d Generating a virtual simulation model diagram of the welding bead section appearance according to the geometric structure parameters and the weld width of the weldY w And penetration depthY d The resulting fitted curve represents the top profile curve ADB of the weld bead.
7. The method for virtual simulation of the morphology of a weld according to claim 6, wherein step S302 is performed according to the geometric parameters of the weld and the weld fusion widthY w And penetration depthY d The generated fitting curve is a cosine curve, and the function expression is as follows:
in the above formula, the first and second carbon atoms are,ybeing corresponding points on the cosine curveyThe coordinates of the axes are set to be,hthe thickness of the blunt edge in the geometric parameters of the weld,ras a butt gap in the geometric parameters of the weld,θthe bevel angle in the geometric parameters of the weld,Y w in order to obtain the weld width of the weld,Y d in order to realize the penetration of the welding seam,xbeing corresponding points on the cosine curvexAxis coordinates.
8. The method for virtually simulating the morphology of a welding bead according to claim 6, wherein step S302 further comprises referencing parameters according to the geometry of the welding beadNumber and weld bead widthY w Generating a virtual simulation model diagram of the surface appearance of the welding bead, wherein the virtual simulation model diagram of the surface appearance of the welding bead adopts ellipses stacked along the welding direction to represent molten pool units stacked along the welding direction, and the major axis of the ellipses is the melt widthY w Double, minor axis size equals fusion widthY w 。
9. A system for topographical modeling and virtual simulation of a weld joint, comprising a microprocessor and a memory connected to one another, wherein said microprocessor is programmed or configured to perform a method for topographical virtual simulation of a weld joint as claimed in any one of claims 6 to 8.
10. A computer-readable storage medium, in which a computer program is stored, wherein the computer program is adapted to be programmed or configured by a microprocessor to perform a method for virtual simulation of the topography of a weld bead according to any one of claims 6 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211478685.1A CN115730520A (en) | 2022-11-24 | 2022-11-24 | Shape prediction and virtual simulation method and system for welding seam |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211478685.1A CN115730520A (en) | 2022-11-24 | 2022-11-24 | Shape prediction and virtual simulation method and system for welding seam |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115730520A true CN115730520A (en) | 2023-03-03 |
Family
ID=85297818
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211478685.1A Pending CN115730520A (en) | 2022-11-24 | 2022-11-24 | Shape prediction and virtual simulation method and system for welding seam |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115730520A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117583698A (en) * | 2024-01-19 | 2024-02-23 | 中建材（合肥）粉体科技装备有限公司 | Automatic surfacing device and surfacing control method |
-
2022
- 2022-11-24 CN CN202211478685.1A patent/CN115730520A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117583698A (en) * | 2024-01-19 | 2024-02-23 | 中建材（合肥）粉体科技装备有限公司 | Automatic surfacing device and surfacing control method |
| CN117583698B (en) * | 2024-01-19 | 2024-04-26 | 中建材（合肥）粉体科技装备有限公司 | Automatic surfacing device and surfacing control method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Xiong et al. | Bead geometry prediction for robotic GMAW-based rapid manufacturing through a neural network and a second-order regression analysis | |
| Kim et al. | An investigation into an intelligent system for predicting bead geometry in GMA welding process | |
| CN102905848B (en) | Automated determination of jet orientation parameters in three-dimensional fluid jet cutting | |
| Chang et al. | Penetration quality prediction of asymmetrical fillet root welding based on optimized BP neural network | |
| KR20140128361A (en) | Virtual welding system | |
| CN109420819B (en) | Robot additive manufacturing method based on MIG/MAG welding process | |
| Yang et al. | Combining the Taguchi method with artificial neural network to construct a prediction model of a CO 2 laser cutting experiment | |
| CN115730520A (en) | Shape prediction and virtual simulation method and system for welding seam | |
| CN106994684A (en) | The method of control machine people's instrument | |
| Wang et al. | Intelligent modelling of back-side weld bead geometry using weld pool surface characteristic parameters | |
| Yang et al. | An automatic detection and identification method of welded joints based on deep neural network | |
| WO2021009963A1 (en) | Welding work data storage device, welding work assistance system, and welding robot control device | |
| CN106994545A (en) | Device for assisting in setting and setting householder method | |
| CN108549759B (en) | Part coating thickness calculation and simulation method for automatic spraying process | |
| CN113732558A (en) | Online detection method and device for typical quality problems of robot welding | |
| Wang | Three-dimensional vision applications in GTAW process modeling and control | |
| CN107598401A (en) | Plasma arc welding (PAW) electric current molten wide control system based on Fuzzy Reasoning Neural Network | |
| CN110153582A (en) | Welding scheme generation method, device and welding system | |
| Steele et al. | Development of closed‐loop control of robotic welding processes | |
| KR101253542B1 (en) | Method of visual representation of a virtual welding training | |
| Ghaderi et al. | The application of imperialist competitive algorithm for optimization of deposition rate in submerged arc welding process using TiO 2 nano particle | |
| CN111814245A (en) | Ceiling joint layout acquisition method and device, electronic equipment and storage medium | |
| KR101874824B1 (en) | Apparatus for generating beads of virtual welding trainning simulator | |
| Carrino et al. | A neuro-fuzzy approach for increasing productivity in gas metal arc welding processes | |
| Sharma et al. | Multi-objective optimization of bead geometry for submerged arc welding of pipeline steel using RSM-fuzzy approach |
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 |