阅读说明：本技术 一种基于数字孪生的焊接监控方法 (Welding monitoring method based on digital twinning ) 是由 徐劲力 林进权 卢杰 丁刚强 袁智军 黄丰云 潘昊 莫军昌 詹强民 李丁 于 2021-07-30 设计创作，主要内容包括：本发明提供一种基于数字孪生的焊接监控方法,包括：收集焊接过程中的产品生产过程数据,所述产品生产过程数据至少包括产品焊接数据；建立飞溅分析程序,根据产品焊接数据,分析焊接过程中产品的飞溅点；将产品焊接数据输入焊接问题判断预测模型,获取所述焊接问题判断预测模型输出的产品的焊接质量状态信息及焊接质量类型；基于产品的焊接质量类型,配置对应的工艺调整方案并设置触发所述工艺调整方案的条件,以处理产品焊接质量问题。本发明通过对车间产品的焊接过程实时监控及自适应工艺调整,能够显著提高焊接车间监控的实时性和可视化效果,提高焊接质量。(The invention provides a welding monitoring method based on digital twinning, which comprises the following steps: collecting product production process data in a welding process, wherein the product production process data at least comprises product welding data; establishing a spatter analysis program, and analyzing spatter points of the product in the welding process according to the welding data of the product; inputting product welding data into a welding problem judgment and prediction model, and acquiring welding quality state information and welding quality types of the product output by the welding problem judgment and prediction model; and configuring a corresponding process adjustment scheme and setting conditions for triggering the process adjustment scheme based on the welding quality type of the product so as to solve the problem of the welding quality of the product. The invention can obviously improve the real-time performance and the visual effect of the welding workshop monitoring and improve the welding quality by monitoring the welding process of workshop products in real time and adjusting the self-adaptive process.)

1. A digital twinning based weld monitoring method, comprising:

collecting product production process data in a welding process, wherein the product production process data at least comprises product welding data;

establishing a spatter analysis program, and analyzing spatter points of the product in the welding process according to the welding data of the product;

inputting product welding data into a welding problem judgment and prediction model, and acquiring welding quality state information and welding quality types of the product output by the welding problem judgment and prediction model;

and configuring a corresponding process adjustment scheme and setting conditions for triggering the process adjustment scheme based on the welding quality type of the product so as to solve the problem of the welding quality of the product.

2. The weld monitoring method of claim 1, wherein the product manufacturing process data comprises:

the working state information comprises pressure, temperature, vibration, rotating speed and moment;

the working condition information comprises length, working radius and rated load;

the control command comprises a control command output by the main controller to the frequency converter, the servo driver and the actuator;

fault diagnosis information, wherein the fault diagnosis information comprises fault information of the main controller and fault codes of each execution device;

the working duration information comprises the accumulated working time of welding and total output;

the environment information comprises working environment humidity and temperature;

product quality problem data, the product quality problems including weld spatter, distortion, bite, hump;

the collecting welding data of the product in the welding process further comprises the following steps:

and developing and collecting product production process data by utilizing an OPC UA framework, and uniformly expressing different generation process data of the product.

3. The welding monitoring method of claim 1, wherein the establishing a spatter analysis program for analyzing spatter points of a product during welding based on product welding data comprises:

for any welding point of each product, calculating welding resistance according to the welding current and the welding voltage of the welding point in a set time period, and scanning a resistance slope curve;

finding all resistance slope peak points on the resistance slope curve, calculating the standard deviation of all the resistance slope peak points, and judging whether a protruded resistance slope change value exists on the resistance slope curve or not based on the standard deviation, wherein the time point of the protruded resistance slope change value is a splash point.

4. The welding monitoring method of claim 1, wherein the welding problem determination prediction model is trained by:

collecting a plurality of welding parameters of each welding spot of a product, respectively constructing a deep learning model based on a single parameter and a deep learning model based on a plurality of combined parameters by taking the plurality of welding parameters as input parameters, and respectively training to obtain each deep learning model after training;

fusing each deep learning model based on a DS (Dempster-Shafer) evidence theory, and determining the optimal input combination parameters of the fused deep learning model by comparing the accuracy of the deep learning model before fusion with the accuracy of the fused deep learning model after fusion;

screening excellent hyper-parameters through a Hyperband optimization algorithm;

and acquiring a fused deep learning model based on the optimal input combination parameters, and judging a prediction model for the welding problem.

5. The weld monitoring method of claim 4, wherein the plurality of welding parameters includes a welding speed, a welding current, a welding pressure, and a welding time.

6. The welding monitoring method according to claim 1, 4 or 5, wherein the welding quality state information of the product comprises welding quality problems and no welding quality problems, and the welding quality defects comprise excessive plate gaps and electrode rod misalignment.

7. The welding monitoring method of claim 1, wherein configuring a corresponding process adjustment scheme and setting conditions for triggering the process adjustment scheme to address product weld quality issues based on the weld quality type of the product comprises:

in the process of welding a product, if welding spatter occurs or welding resistance exceeds the range of a set standard curve, adjusting welding current, welding voltage or welding time;

and when judging that a plurality of same quality problems occur in a short time after welding, selecting a proper scheme from preset corresponding schemes according to the corresponding reasons and applying the scheme.

8. The weld monitoring method of claim 1, further comprising:

in the welding process, performing three-dimensional modeling on equipment and a field of a production workshop by using the Unity, and creating a corresponding motion control script for each piece of motion equipment; and the number of the first and second groups,

the product welding data is converted into graphics by Open CV and is subjected to mapping rendering on Unity.

9. The weld monitoring method of claim 8, wherein the three-dimensional modeling of the facilities and the field of the production floor using Unity and creating a corresponding motion control script for each of the motion devices during the welding job comprises:

three-dimensional modeling of equipment is carried out by utilizing three-dimensional modeling software UG, and then a model in the UG is imported into 3D Max in an STL format;

after the model is adjusted by using 3D Max, exporting the model to a Unity3D engine in an FBX file format;

developing a model driving script in Visual Studio;

the real-time motion data of each joint of the robot is collected and processed, then transmitted to the Unity and driven according to the corresponding script, wherein the automatic welding operation is carried out on the product by controlling the motion of the robot.

10. The weld monitoring method of claim 8, wherein converting the product weld data to graphics and rendering the same on Unity using Open CV comprises:

analyzing the width of a welding seam, the welding temperature and the welding quality type according to the welding data of the product;

and converting the width of the welding seam, the welding temperature and the welding quality type into pictures for rendering based on Open CV.

Technical Field

The invention relates to the field of workshop welding, in particular to a welding monitoring method based on digital twinning.

Background

With the development of information technology and the influence of times trend, the automation and informatization levels of the manufacturing industry are continuously improved, and more manufacturing enterprises adopt an automatic production line mode to carry out production and manufacturing activities. A large amount of automatic production lines are put into use, so that the production efficiency can be greatly improved, and the economic benefit of enterprises is remarkably improved. However, one of the difficulties faced by the existing enterprises is that the production line informatization management level is not high, the enterprises lack a quick and effective means for managing and monitoring the operation state of the workshop, and the production process is difficult to realize in a transparent manner due to the lack of an effective information interaction means between the management systems and the control system. The workshop is monitored by manual recording, two-dimensional reports, configuration monitoring and other modes on the traditional production line, and the real-time performance and the visualization effect are poor.

Although industrial robots have been widely applied in the field of welding, factors such as thermal deformation, machining errors and assembly errors of workpieces in the actual welding process cause changes in the position and size of a weld joint, and the welding robots weld the workpieces according to established programs, lack of real-time monitoring and adaptive adjustment of welding quality, and cause quality problems in the welding production process of the workpieces.

Disclosure of Invention

Aiming at the problems of poor welding workshop real-time performance and visualization effect and slow process adjustment reaction in the prior art, the invention provides a welding monitoring method based on digital twinning, which aims to realize the purposes of improving the product quality and reducing the maintenance cost, and comprises the following steps: collecting product production process data in a welding process, wherein the product production process data at least comprises product welding data; establishing a spatter analysis program, and analyzing spatter points of the product in the welding process according to the welding data of the product; inputting product welding data into a welding problem judgment and prediction model, and acquiring welding quality state information and welding quality types of the product output by the welding problem judgment and prediction model; and configuring a corresponding process adjustment scheme and setting conditions for triggering the process adjustment scheme based on the welding quality type of the product so as to solve the problem of the welding quality of the product.

On the basis of the technical scheme, the invention can be improved as follows.

Optionally, the product production process data includes: the working state information comprises pressure, temperature, vibration, rotating speed and moment; the working condition information comprises length, working radius and rated load; the control command comprises a control command output by the main controller to the frequency converter, the servo driver and the actuator; fault diagnosis information, wherein the fault diagnosis information comprises fault information of the main controller and fault codes of each execution device; the working duration information comprises the accumulated working time of welding and total output; the environment information comprises working environment humidity and temperature; product quality problem data, the product quality problems including weld spatter, distortion, bite, hump; the collecting welding data of the product in the welding process further comprises the following steps: and developing and collecting product production process data by utilizing an OPC UA framework, and uniformly expressing different generation process data of the product.

Optionally, the establishing a spatter analysis program, analyzing spatter points of the product in the welding process according to the welding data of the product, includes: for any welding point of each product, calculating welding resistance according to the welding current and the welding voltage of the welding point in a set time period, and scanning a resistance slope curve; finding all resistance slope peak points on the resistance slope curve, calculating the standard deviation of all the resistance slope peak points, and judging whether a protruded resistance slope change value exists on the resistance slope curve or not based on the standard deviation, wherein the time point of the protruded resistance slope change value is a splash point.

Optionally, the welding problem judgment and prediction model is trained in the following manner: collecting a plurality of welding parameters of each welding spot of a product, respectively constructing a deep learning model based on a single parameter and a deep learning model based on a plurality of combined parameters by taking the plurality of welding parameters as input parameters, and respectively training to obtain each deep learning model after training; fusing each deep learning model based on a DS (Dempster-Shafer) evidence theory, and determining the optimal input combination parameters of the fused deep learning model by comparing the accuracy of the deep learning model before fusion with the accuracy of the fused deep learning model after fusion; screening excellent hyper-parameters through a Hyperband optimization algorithm; and acquiring a fused deep learning model based on the optimal input combination parameters, and judging a prediction model for the welding problem.

Optionally, the plurality of welding parameters includes welding speed, welding current, welding pressure and welding time.

Optionally, the welding quality state information of the product includes a welding quality problem and a non-welding quality problem, and the welding quality type includes an excessively large plate gap and an electrode rod misalignment.

Optionally, the configuring a corresponding process adjustment scheme and setting a condition for triggering the process adjustment scheme based on the welding quality type of the product to solve the problem of the welding quality of the product includes: in the process of welding a product, if welding spatter occurs or welding resistance exceeds the range of a set standard curve, adjusting welding current, welding voltage or welding time; and when judging that a plurality of same quality problems occur in a short time after welding, selecting a proper scheme from preset corresponding schemes according to the corresponding reasons and applying the scheme.

Optionally, the method further includes: in the welding process, performing three-dimensional modeling on equipment and a field of a production workshop by using the Unity, and creating a corresponding motion control script for each piece of motion equipment; and converting the product welding data into a graph by utilizing Open CV and performing mapping rendering on Unity.

Optionally, in the welding work process, three-dimensional modeling is performed on the equipment and the site of the production plant by using Unity, and a corresponding motion control script is created for each piece of motion equipment, where the three-dimensional modeling includes: three-dimensional modeling of equipment is carried out by utilizing three-dimensional modeling software UG, and then a model in the UG is imported into 3D Max in an STL format; after the model is adjusted by using 3D Max, exporting the model to a Unity3D engine in an FBX file format; developing a model driving script in Visual Studio; the real-time motion data of each joint of the robot is collected and processed, then transmitted to the Unity and driven according to the corresponding script, wherein the automatic welding operation is carried out on the product by controlling the motion of the robot.

Optionally, the converting the product welding data into a graph and performing mapping rendering on Unity by using Open CV includes: analyzing the width of a welding seam, the welding temperature and the welding quality type according to the welding data of the product; and converting the width of the welding seam, the welding temperature and the welding quality type into pictures for rendering based on Open CV.

According to the welding monitoring method based on the digital twin, provided by the invention, the product welding process in a welding workshop is monitored, the welding defects of the product can be found in time, the welding quality can be judged in the welding process, the welding process problem can be analyzed, the welding spatter is reduced by properly adjusting the welding parameters, the welding quality is improved, the visual management of the workshop is greatly improved, the product welding quality is improved, the production cost is reduced, and the economic benefit is improved.

Drawings

FIG. 1 is a flow chart of a digital twinning based weld monitoring method provided by the present invention;

FIG. 2 is a flow chart of a training process for a weld problem determination prediction model;

FIG. 3 is a flow chart for three-dimensional modeling of equipment and sites in a production plant;

FIG. 4 is a flow diagram of rendering a map on Unity.

Detailed Description

The following detailed description of embodiments of the present invention is provided in connection with the accompanying drawings and examples. The following examples are intended to illustrate the invention but are not intended to limit the scope of the invention.

FIG. 1 is a welding monitoring method based on digital twinning, which comprises the following steps: and S1, collecting product production process data in the welding process, wherein the product production process data at least comprises product welding data.

It can be understood that, in the process of welding and processing products in a workshop, a large amount of multi-source heterogeneous data exists, such as state data sensor data of a welding robot, data of transmission equipment and the like, and the multi-source heterogeneous data needs to be collected, wherein in the welding and processing process of the products, welding data of each product is included in product generation process data. The product generation process data mainly comprises the following data:

operating state information: such as pressure, temperature, vibration, rotational speed, torque, etc.; working condition information: such as length, working radius, rated load, etc.; controlling command: for example, the main controller outputs control instructions to devices such as a frequency converter, a servo driver, an actuator and the like; fourthly, fault diagnosis information: such as failure information of the main controller, failure codes of the respective execution devices, and the like; the working duration information: such as the accumulated working time of welding, total output and the like; environment information: such as working environment humidity, temperature; quality information of the product: welding spatter, deformation, bite, hump, etc.

Due to the variety of plant field devices, sensors, and communication specifications, specifying parsing rules for each specification individually is not only labor intensive, but also less versatile. And OPC UA has the advantages of open platform, safe communication, good expandability and the like. Therefore, OPC UA is used to develop data collection.

And when the data are collected, the data are classified and uniformly expressed. The data is divided into structured data and unstructured data. The Data is expressed as Data { type, access address, metadata } and encapsulated in a unified mode to form Json integrated Data, so that the performance of the Data in the aspects of transmission, access, expansion and the like can meet the requirements of a virtual monitoring system.

And S2, establishing a spatter analysis program, and analyzing the spatter point of the product in the welding process according to the welding data of the product.

It can be understood that, after the welding data of the product in the welding process of the product is collected in step S1, the spatter analysis program in the welding process is established in step S2, and the collected welding data of the product is analyzed to analyze the spatter point in the welding process, where the spatter point belongs to a place where the welding quality is problematic.

Wherein, establish splash analysis program, according to product welding data, the splash point of analysis welding in-process product includes: for any welding point of each product, calculating welding resistance according to the welding current and the welding voltage of any welding point in a set time period, and scanning a resistance slope curve; finding all resistance slope peak points on the resistance slope curve, calculating the standard deviation of all the resistance slope peak points, and judging whether a prominent resistance slope change value exists on the resistance slope curve or not based on the standard deviation, wherein the time point of the prominent resistance slope change value is a splash point.

It can be understood that a certain time is needed for welding each welding point of each product, the welding resistance at different time points is in a certain change rule in the welding process, and the problem of welding defects can be judged by searching the abrupt change constant point of the welding resistance, wherein the welding defect problem is mainly identified by the spattering point in the welding process.

Specifically, according to each different time point, the welding resistance is calculated according to the welding current and the welding voltage, the resistance slope curve is scanned, all resistance slope peak values in the resistance slope curve are found out, and the standard deviation of all peak value points is calculated to judge whether a prominent slope change value exists. When the standard deviation of the peak sample point is greater than the preset value, there is an abnormal point, and conversely, when the standard deviation is less than the preset value, there is no abnormal protruding point, i.e., no splash point, on the resistance slope curve. If the specific position judgment formula of the splash point is shown as the formula (1), if the ith peak value point meets the formula, the point is regarded as the splash point.

Wherein N is the number of resistance slope peak points on the resistance slope curve, y_iThe value of the resistivity slope at the ith point is 0.65 as empirical data, and D is distance.

D (distance) is defined as in formula (2):

this distance D represents the maximum value of the found samples considered to be the peak point minus the average of the samples.

And judging the spattering point in the product welding process according to the formulas (1) and (2) and the judgment method, and providing data support for subsequently adjusting the welding process scheme.

And S3, inputting the welding data of the product into the welding problem judgment and prediction model, and acquiring the welding quality state information and the welding quality type of the product output by the welding problem judgment and prediction model.

Referring to fig. 2, the welding problem judgment prediction model is trained in the following manner: collecting a plurality of welding parameters of each welding spot of a product, respectively constructing a deep learning model based on a single parameter and a deep learning model based on a plurality of combined parameters by taking the plurality of welding parameters as input parameters, and respectively training to obtain each deep learning model after training; fusing each deep learning model based on a DS (Dempster-Shafer) evidence theory, and determining the optimal input combination parameters of the fused deep learning model by comparing the accuracy of the deep learning model before fusion with the accuracy of the fused deep learning model after fusion; excellent hyper-parameters are screened out through a Hyperband optimization algorithm, a fused deep learning model based on optimal input combination parameters is obtained, and a prediction model is judged for a welding problem.

It will be appreciated that during welding of a product, various welding parameters may be collected, such as, for example, the number of spatters, welding speed, welding current, welding pressure, and welding time. A welding problem judgment and prediction model based on deep learning is established based on the welding parameters, specifically, factors such as the number of spatters, welding speed, welding current, welding pressure and welding time are used as input parameters, a simple full-link neural network model is made for each parameter or a plurality of combined parameters, the number of layers is 256x128x64x10, an activation function is relu, and a loss function is a cross entropy function. And training to obtain each deep learning model with well adjusted built-in parameters, and then obtaining the accuracy of each deep learning model through a verification set. And changing the output of each deep learning model into probabilities of different classifications, fusing different models according to a DS evidence theory, verifying different fused models, judging whether the accuracy of the fused models is improved to the accuracy of the models before fusion, and finally selecting the parameter combination with the highest accuracy. Wherein, the DS evidence theory formula is as follows:

where m () is the basic probability distribution function of the model. A. B, C are various classifications of models.

Setting different hyper-parameters and screening, wherein the number of nerve layers is 32-256, the step pitch is 8, the activation functions are relu, sigmoid and the like, the learning rate range is 0.01-0.0001, the step pitch is 0.0005, screening excellent hyper-parameters from the hyper-parameters through a Hyperband algorithm, setting the model hyper-parameters as excellent hyper-parameters by taking the optimal parameter as input, and training the model by taking the problems that the plate gap is too large and the electrode rod does not meet the medium welding quality as output to obtain a well-trained deep learning model.

In short, the deep learning models can be trained according to a single welding parameter or a plurality of combined welding parameters, and the accuracy of each deep learning model is calculated by using the verification set. And fusing the plurality of deep learning models to obtain a fused deep learning model, similarly, calculating the accuracy of the fused deep learning model by using a verification set, wherein when the fused deep learning model has the highest accuracy, the parameter combination at the moment is the optimal parameter combination, and the fused deep learning model is also the optimal deep learning model, namely the finally obtained welding problem judgment and prediction model.

And when the optimal welding problem judgment and prediction model is obtained, inputting the acquired welding data into the welding problem judgment and prediction model, and outputting the welding quality state information and the welding quality type of the product. The welding quality state information of the product comprises welding quality and non-welding quality, and the welding quality category comprises different types of welding quality problems of qualified welding, overlarge plate gap and unequal electrode rods.

And S4, configuring a corresponding process adjustment scheme based on the welding quality type of the product and setting conditions for triggering the process adjustment scheme so as to deal with the welding quality problem of the product.

Specifically, in the process of welding a product, if welding spatter occurs or welding resistance exceeds a set standard curve range, welding current, welding voltage or welding time is adjusted. And when judging that a plurality of same quality problems occur in a short time after welding, selecting a proper scheme from preset corresponding schemes according to the corresponding reasons and applying the scheme.

It can be understood that, after analyzing and judging the spattering point and the welding quality in the welding process through the above embodiments, this step needs to take corresponding measures to deal with the quality problem in the welding process. Specifically, for different welding problems, corresponding process adjustment schemes are designed and trigger process adjustment conditions are set so as to rapidly process the welding problems.

In the implementation process, when welding spatter is judged to occur in the welding process or the welding resistance exceeds the range of the set standard curve band, the welding current, the welding voltage or the welding time are adjusted. If the spatter occurs in 200ms, the welding time is prolonged by reducing the welding small current, so that the nugget can be slowly increased, a smooth resistance curve is obtained, and the further problem of welding is prevented.

And when judging that a plurality of same quality problems appear in a short time after welding, selecting a proper scheme from preset corresponding schemes according to the corresponding reasons and applying the scheme. If the problem of matching gaps between the plates is found, welding parameters can be reasonably adjusted, pressure is properly increased, welding operation is controlled, and the attachment of the static electrode arm of the welding tongs is ensured as much as possible; the welding sequence is reasonably adjusted, and the matching gap of the plates is reduced; and informing a manager to optimize the part size and the tool clamp, and ensuring the matching clearance.

In a possible implementation manner, the method further includes: s5, in the welding process, performing three-dimensional modeling on equipment and a site of a production workshop by using Unity, and creating a corresponding motion control script for each piece of motion equipment; and S6, converting the product welding data into graphics by utilizing Open CV and performing mapping rendering on Unity.

It can be understood that in the embodiment of the invention, for the welding of the product, the traditional manual welding process is abandoned, and the robot is adopted to automatically weld the product, so that the full-automatic welding process of a workshop is realized. Specifically, the welding process of each welding point of the product is realized by controlling the movement of a mechanical arm of the robot and the like, so that the movement condition of the robot plays a very critical role in the welding quality of the product.

Therefore, in the process of welding a product, the motion process of the robot needs to be monitored and simulated, and in the embodiment of the present invention, step S5 simulates the motion model of each motion device of the workshop by three-dimensionally modeling the devices (e.g., the robot) and the field of the production workshop using Unity. And in the welding process, the welding data of the product is monitored in the whole process, and step S6 uses Open CV to convert the welding data of the product into graphics and render the graphics on Unity.

Note that, the steps S5 and S6 are not performed in the order of S5 and S6, and the steps S5 and S6 may be performed in parallel. In addition, the sequence of steps S5, S6 and other steps S1 to S4 may be performed in parallel.

In one possible embodiment, during the welding operation, referring to fig. 3, in S5, the facilities and the field of the production plant are modeled in three dimensions by Unity, and a corresponding motion control script is created for each motion facility, including: s51, performing three-dimensional modeling on the equipment by using three-dimensional modeling software UG, and then importing the model in the UG into 3D Max in an STL format; s52, adjusting the model by using 3D Max, and exporting the model to a Unity3D engine in an FBX file format; s53, developing a model driving script in Visual Studio; and S54, acquiring and processing real-time motion data of each joint of the robot, transmitting the real-time motion data to Unity, and driving according to a corresponding script, wherein the robot is controlled to move to perform automatic welding operation on a product.

Specifically, production equipment models such as a welding robot model, a welding gun model, a conveying equipment model and a product model are established in UG and are exported to STL format. And adjusting the established various equipment models to prevent the loss of point and plane, adding corresponding materials and rendering to the models according to the real equipment of the models, and exporting the models in an FBX format.

For the motion simulation of the equipment model, whether the equipment model is a motion model or not is distinguished, and for the model which can move, a driving behavior mode of the model is set by calling methods such as transform. The data call function is set so that the script can obtain the drive data from the database to make the model move.

It should be noted that, in the embodiment of the present invention, the welding process of the workshop is comprehensively monitored, and a statistical chart program is designed and packaged for various statistical data. Specifically, a chart interface program such as a broken line chart or a pie chart is developed on the UGUI. The programs are packaged and modularized, and then a chart module is called and relevant attributes (such as a data interface, a chart title and an axis) of the statistical chart are configured according to the monitoring requirement of the workshop.

In one possible embodiment, referring to fig. 4, the converting the product welding data into graphics and rendering the maps on Unity by using Open CV in step S6 includes: s61, analyzing the width of the welding seam, the welding temperature and the welding quality type according to the welding data of the product; and S62, converting the weld width, the welding temperature and the welding quality type into pictures for rendering based on Open CV.

It can be understood that, for the welding data of the product in the welding process, the simulation can be performed for graphical display, and specifically, the width of the welding seam, the welding temperature and different types of quality defects generated in the welding process can be analyzed according to the welding data.

The specific implementation mode of analyzing the width of the welding seam according to the welding data is that the width of the welding seam is in a descending trend along with the increase of the welding speed under the same current condition. Therefore, under the same material, through data analysis of welding speed and current, a two-factor coupling curve function can be established to simulate the welding seam width, as shown in formula (5).

w＝af(x)+bθ(y)； (5)

w is the width of the weld, f (x) is a function of the welding current and the welding width, theta (x) is a function of the welding speed and the welding width, and a and b are coupling influence factors of the welding current and the welding speed on the welding width respectively.

Similarly, the specific implementation of analyzing the welding temperature according to the welding data is to determine the temperature field at the welding position according to the data of the temperature sensor or the welding current and the welding speed.

Similarly, the welding defect is analyzed according to the welding data by analyzing the resistance data and determining the welding defect and the welding position by adding the time and speed data through the spatter analysis program of S3.

Welding data in the welding process are analyzed, the Open CV is used for converting the welding data into picture information, dynamic pictures are rendered on the Unity, the rendered dynamic pictures can show different welding qualities in the welding process, and the whole welding process can be monitored in the whole process.

According to the welding monitoring method based on the digital twin, provided by the embodiment of the invention, the operation condition of each device can be visually observed through digital simulation of a welding workshop, and the defects of products can be timely discovered. Welding parameters can be properly adjusted in the welding process, so that welding spatter is reduced to improve welding quality, the problem of the process can be analyzed and found after a certain number of workpieces are processed, the process is adaptively adjusted, visual management of a workshop is greatly improved, the welding quality of products is improved, the production cost is reduced, and the economic benefit is improved.

It should be noted that, in the foregoing embodiments, the descriptions of the respective embodiments have respective emphasis, and reference may be made to relevant descriptions of other embodiments for parts that are not described in detail in a certain embodiment.

As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable 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 invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams 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 computer, 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.

While preferred embodiments of the present invention have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. Therefore, it is intended that the appended claims be interpreted as including preferred embodiments and all such alterations and modifications as fall within the scope of the invention.

It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.