Welding equipment monitoring system and method based on Internet of Things technology
Through the welding equipment monitoring system of Internet of Things technology, welding data and image feature encoding are used to establish a responsive relationship between welding process parameters and weld quality, and the precise control of welding pulse current is achieved, which solves the unstable welding quality in traditional welding methods. question.
Patent Information
- Application Number
- CN202411141118.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-20
- Publication Date
- 2024-11-05
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Traditional welding methods rely on welders' experience, resulting in a lack of objectivity and consistency in the selection of welding process parameters, which affects the welding quality. It is difficult to fully consider the workpiece characteristics when welding workpieces of different thicknesses or materials.
Using the welding equipment monitoring system based on the Internet of Things technology, the welding related data acquisition module, feature encoding module and pulse current adjustment result generation module can obtain the welding machine pulse current value, wire feeding speed value and weld monitoring image, and perform feature encoding and The responsive matrix is established to achieve accurate control of welding pulse current.
The stability and consistency of welding quality are improved, and the precise control of welding parameters is achieved through artificial intelligence data analysis, reducing the subjective influence of welder experience.
Smart Images

Figure CN118893374A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of intelligent welding control, and more specifically, to a welding equipment monitoring system and method based on Internet of Things technology. Background Art
[0002] In some complex welding scenarios, such as welding workpieces of different thicknesses or materials, the traditional method is to determine the welding process parameters based on welding experience. However, the experience levels and judgment criteria of different welders are different, which will greatly affect the welding quality of the weld.
[0003] Therefore, a welding equipment monitoring solution based on Internet of Things technology is needed. Summary of the invention
[0004] In view of the shortcomings of the prior art, the present application provides a welding equipment monitoring system and method based on Internet of Things technology.
[0005] According to one aspect of the present application, a welding equipment monitoring system based on Internet of Things technology is provided, which includes:
[0006] A welding-related data acquisition module, used to obtain the pulse current value of the welding machine and the wire feeding speed value of the wire feeder at multiple predetermined time points within a predetermined time period and the weld monitoring image within the predetermined time period;
[0007] A welding-related data feature encoding module, used for feature encoding the pulse current values of the welding machine and the wire feeding speed values of the wire feeder at the plurality of predetermined time points and the weld monitoring image within the predetermined time period to obtain a weld quality responsiveness feature matrix;
[0008] The welding machine pulse current adjustment result generation module is used to determine the welding machine pulse current adjustment direction at the current time point based on the weld quality responsiveness characteristic matrix.
[0009] According to another aspect of the present application, a welding equipment monitoring method based on Internet of Things technology is provided, which includes:
[0010] Acquire the pulse current value of the welding machine and the wire feeding speed value of the wire feeder at multiple predetermined time points within a predetermined time period and the weld monitoring image within the predetermined time period;
[0011] Performing feature coding on the pulse current values of the welding machine and the wire feeding speed values of the wire feeder at the plurality of predetermined time points and the weld monitoring image within the predetermined time period to obtain a weld quality responsiveness feature matrix;
[0012] Based on the weld quality responsiveness characteristic matrix, the pulse current adjustment direction of the welding machine at the current time point is determined.
[0013] This application has significant technical effects due to the adoption of the above technical solutions:
[0014] The welding equipment monitoring system and method based on the Internet of Things technology provided by the present application adopts the data analysis technology based on artificial intelligence, and extracts the associated time series features of the pulse current value of the welding machine and the wire feeding speed value of the wire feeder at multiple time points to obtain the time series features of the welding process parameters, and extracts the time series features of the weld monitoring image in a certain period of time to obtain the weld quality features, and then establishes the responsiveness relationship between the weld quality features and the welding process parameter time series features, and finally outputs the direction of the pulse current adjustment of the welding machine through the classifier. In this way, the precise control of the welding pulse current can be achieved, which helps to improve the welding quality. BRIEF DESCRIPTION OF THE DRAWINGS
[0015] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.
[0016] Figure 1 4 is a system block diagram of a welding equipment monitoring system based on Internet of Things technology according to an embodiment of the present application.
[0017] Figure 2 It is a block diagram of a welding-related data feature encoding module in a welding equipment monitoring system based on Internet of Things technology according to an embodiment of the present application.
[0018] Figure 3 It is a block diagram of a welding process parameter timing feature extraction unit in a welding equipment monitoring system based on Internet of Things technology according to an embodiment of the present application.
[0019] Figure 4 It is a block diagram of a weld quality dynamic feature extraction unit in a welding equipment monitoring system based on Internet of Things technology according to an embodiment of the present application.
[0020] Figure 5 The figure is a flow chart of a welding equipment monitoring method based on Internet of Things technology according to an embodiment of the present application. DETAILED DESCRIPTION
[0021] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.
[0022] As mentioned in the above background technology, in complex welding scenarios, such as welding workpieces of different thicknesses or materials, traditional methods rely on the welder's experience to determine the welding process parameters. However, the welder's subjective judgment is easily affected by personal experience, skill level and subjective preference, resulting in a lack of objectivity and consistency in the selection of welding process parameters, which can lead to instability and uncertainty in welding quality. And when welding workpieces of different thicknesses or materials, traditional methods often fail to fully consider the impact of different workpiece characteristics on welding process parameters, which can lead to inappropriate selection of process parameters and thus affect weld quality. Therefore, a welding equipment monitoring solution based on Internet of Things technology is expected.
[0023] In recent years, deep learning and neural networks have been widely used in computer vision, natural language processing, text signal processing and other fields. In addition, deep learning and neural networks have also shown a level close to or even beyond that of humans in image classification, object detection, semantic segmentation, text translation and other fields. The development of deep learning and neural networks provides new solutions and solutions for welding equipment monitoring based on Internet of Things technology.
[0024] Figure 1 FIG. 1 is a system block diagram of a welding equipment monitoring system based on Internet of Things technology according to an embodiment of the present application. Figure 1 As shown, in the welding equipment monitoring system 100 based on the Internet of Things technology, it includes: a welding-related data acquisition module 110, which is used to obtain the welding machine pulse current value and the wire feeding speed value of the wire feeder at multiple predetermined time points within a predetermined time period and the weld monitoring image within the predetermined time period; a welding-related data feature encoding module 120, which is used to feature encode the welding machine pulse current value and the wire feeding speed value of the wire feeder at the multiple predetermined time points and the weld monitoring image within the predetermined time period to obtain a weld quality responsiveness feature matrix; a welding machine pulse current adjustment result generation module 130, which is used to determine the welding machine pulse current adjustment direction at the current time point based on the weld quality responsiveness feature matrix.
[0025] Specifically, in the technical solution of the present application, first, the pulse current value of the welding machine and the wire feeding speed value of the wire feeder at multiple predetermined time points within a predetermined time period and the weld monitoring image within the predetermined time period are obtained. It should be understood that the pulse current value is one of the key factors determining the welding heat input, which directly affects the penetration depth and width of the weld. Too high a pulse current value of the welding machine will lead to overheating, burn-through and excessive welding deformation, while too low a current value will lead to unfused welds. Appropriate pulse current values help to form good weld morphology, including the width, height and surface finish of the weld. The wire feeding speed determines the speed at which the welding wire enters the molten pool, thereby affecting the filling amount of the weld. The appropriate wire feeding speed can ensure the continuity and uniformity of the weld and avoid the weld being too wide or too narrow. Welding is a dynamic process. The two welding parameters of the welding machine pulse current and wire feeding speed will change with the passage of welding time. Collecting welding parameter index data at multiple time points can fully reflect the dynamic changes of the welding process and provide rich data support for subsequent modeling and analysis. The weld monitoring image can intuitively display the morphological characteristics of the weld, such as the width, height, surface finish, etc. of the weld, which are important bases for evaluating the quality of the weld. And through the weld monitoring image, various defects generated during the welding process, such as pores, cracks, and lack of fusion, can be identified. In other words, the weld monitoring image reflects the state of the welding process. By analyzing the collected data such as the welding machine pulse current, wire feeding speed, and weld monitoring image, a correlation model between the weld monitoring image and the welding process parameters can be established. This model can provide a scientific basis for the optimization of the welding process and guide the welding equipment monitoring system to select the best welding parameters to achieve a steady improvement in welding quality.
[0026] In an embodiment of the present application, a possible implementation method for obtaining the welding machine pulse current value and the wire feeding speed value of the wire feeder at multiple predetermined time points within a predetermined time period and the weld monitoring image within the predetermined time period may be: the welding machine, wire feeder and other welding equipment in the present application support Internet of Things communication and can upload working status data in real time. High-precision sensors are installed on the welding machine and wire feeder for collecting real-time pulse current values and wire feeding speed values. At the same time, a high-definition camera is configured as a weld monitoring device to collect weld monitoring images in real time.
[0027] Figure 2 FIG. 1 is a block diagram of a welding-related data feature encoding module in a welding equipment monitoring system based on Internet of Things technology according to an embodiment of the present application. Figure 2As shown, the welding-related data feature encoding module 120 includes: a welding process parameter time series feature extraction unit 121, which is used to extract the time series features of the welding machine pulse current value and the wire feeding speed value of the wire feeder at the plurality of predetermined time points to obtain a welding process parameter associated time series feature vector; a weld quality dynamic feature extraction unit 122, which is used to extract the features of the weld monitoring image within the predetermined time period to obtain a weld quality dynamic feature vector; a welding process parameter time series feature optimization unit 123, which is used to enhance the causal dependency of the weld process parameter associated time series feature vector based on the multi-angle feature association metric based on the weld quality dynamic feature vector to obtain an optimized weld process parameter associated time series feature vector; and a responsiveness estimation unit 124, which is used to calculate the responsiveness estimation of the weld quality dynamic feature vector relative to the optimized weld process parameter associated time series feature vector to obtain the weld quality responsiveness feature matrix.
[0028] Figure 3 FIG. 1 is a block diagram of a welding process parameter timing feature extraction unit in a welding equipment monitoring system based on Internet of Things technology according to an embodiment of the present application. Figure 3 As shown, the welding process parameter timing feature extraction unit 121 includes: a welding process parameter data structuring subunit 1211, which is used to arrange the welding machine pulse current values and the wire feeding speed values of the wire feeder at the multiple predetermined time points according to the time dimension to obtain a pulse current timing input vector and a wire feeding speed timing input vector; a welding process parameter vector association subunit 1212, which is used to associate and encode the pulse current timing input vector and the wire feeding speed timing input vector to obtain a welding process parameter associated timing input matrix; a welding process parameter timing feature generation subunit 1213, which is used to pass the welding process parameter associated timing input matrix through a welding process parameter timing feature encoder to obtain the welding process parameter associated timing feature vector.
[0029] Welding is a dynamic and continuous process, and welding parameters will fluctuate over time. In order to retain the dynamic information of welding parameters changing over time, and thus more accurately reflect the actual state of welding parameters during the welding process, the technical solution of the present application needs to arrange the welding machine pulse current values and wire feeding speed values of the wire feeder at the multiple predetermined time points respectively according to the time dimension to obtain the pulse current timing input vector and the wire feeding speed timing input vector.
[0030] The regulation of pulse current directly affects the size of welding heat input, and then affects the melting speed of welding wire. When the pulse current increases, the welding heat input increases, and the melting speed of welding wire accelerates. At this time, the wire feeding speed needs to be increased accordingly to match the amount of melted welding wire, otherwise it will lead to unstable molten pool and poor weld formation. The speed of wire feeding also directly affects the stability of the welding process and the quality of the weld. If the wire feeding speed is too fast and the pulse current remains unchanged, the welding wire will not be completely melted, resulting in defects such as unfused and slag inclusion; on the contrary, if the wire feeding speed is too slow, it may cause insufficient weld filling or form a too deep molten pool, affecting the mechanical properties of the weld. That is, the pulse current and wire feeding speed of the welding machine will affect each other during the welding process, and this influence is crucial to the welding quality. In order to enable the model to understand the dynamic relationship between the pulse current and the wire feeding speed, it is necessary to associate the pulse current timing input vector and the wire feeding speed timing input vector to obtain the welding process parameter associated timing input matrix. By analyzing the timing input matrix associated with welding process parameters, we can understand the dynamic change characteristics of pulse current and wire feeding speed during welding, as well as their influence on welding quality.
[0031] In an embodiment of the present application, an implementation method of associating and encoding the pulse current timing input vector and the wire feeding speed timing input vector to obtain a welding process parameter associated timing input matrix may be: calculating the welding process parameter associated timing input matrix between the pulse current timing input vector and the wire feeding speed timing input vector using the following associated coding formula; wherein the associated coding formula is:
[0032]
[0033] Where V a represents the pulse current timing input vector, represents the transposed vector of the pulse current timing input vector, V b represents the wire feeding speed timing input vector, M represents the welding process parameter associated timing input matrix, Represents vector multiplication.
[0034] In order to deeply understand the interaction relationship between various welding parameters in the welding process and provide an important basis for the subsequent precise control of welding parameters in the welding process, it is necessary to pass the welding process parameter associated timing input matrix through the welding process parameter timing feature encoder to obtain the welding process parameter associated timing feature vector. The essence of the welding process parameter timing feature encoder in this application is a convolutional neural network model using a two-dimensional convolution kernel. Those of ordinary skill in the art should know that a convolutional neural network (CNN) is a neural network specifically used to process data with a grid structure. In the processing of the welding process parameter associated timing input matrix, since the matrix itself has a two-dimensional structure, it is extremely appropriate to use a CNN model for feature extraction. And the local perception and weight sharing characteristics of CNN enable it to maintain a high computational efficiency when processing large-scale data. In the welding process parameter associated timing input matrix, these characteristics help capture the local correlation between parameters and reduce the number of model parameters.
[0035] In an embodiment of the present application, a possible implementation method of passing the welding process parameter associated timing input matrix through a welding process parameter timing feature encoder to obtain a welding process parameter associated timing feature vector may be: using each layer of the welding process parameter timing feature encoder to perform the following operations on the input data in the forward transfer of the layer: using the convolution units of each layer of the welding process parameter timing feature encoder to perform convolution processing based on a two-dimensional convolution kernel on the input data to obtain a convolution feature map; using the pooling units of each layer of the welding process parameter timing feature encoder to perform pooling processing on each feature matrix along the channel dimension of the convolution feature map to obtain a pooling feature map; using the activation units of each layer of the welding process parameter timing feature encoder to perform nonlinear activation on the feature values of each position in the pooling feature map to obtain an activation feature map; wherein the output of the last layer of the welding process parameter timing feature encoder is the welding process parameter associated timing feature vector.
[0036] Figure 4 FIG. 1 is a block diagram of a weld quality dynamic feature extraction unit in a welding equipment monitoring system based on Internet of Things technology according to an embodiment of the present application. Figure 4 As shown, the weld quality dynamic feature extraction unit 122 includes: a video key frame sampling subunit 1221, used to sample the weld monitoring video of the predetermined time period to obtain a plurality of weld monitoring key frames; a weld quality dynamic feature generation subunit 1222, used to pass the plurality of weld monitoring key frames through a weld quality feature encoder to obtain a weld quality dynamic feature map; a weld quality dynamic feature dimensionality reduction subunit 1223, used to pool the weld quality dynamic feature map along the channel dimension of each feature matrix to obtain the weld quality dynamic feature vector.
[0037] A welding monitoring video of a period of time usually contains a large amount of redundant information because there is often a high degree of similarity between consecutive frames. In order to reduce the amount of data that needs to be stored and processed, improve the data compression ratio, thereby saving storage space and reducing storage costs, key frame extraction can be performed on the weld monitoring video of a predetermined time period. That is, the weld monitoring video of the predetermined time period is sampled to obtain multiple weld monitoring key frames. The weld monitoring key frames are representative video frames in the welding process. By processing and analyzing these key frames containing important information in the welding process, the welding quality can be judged more accurately and the data processing efficiency can be improved, which is particularly important for welding parameters that need to be controlled in real time. Specifically, in an embodiment of the present application, the weld monitoring video of the predetermined time period is sampled at a predetermined sampling frequency to obtain the multiple weld monitoring key frames. It is worth mentioning that the predetermined sampling frequency described here is not fixed and can be adaptively adjusted depending on the specific usage scenario.
[0038] The quality of the weld changes dynamically during the welding process, and the key frames capture the key moments of these changes. By processing these key frames, the weld quality features can be extracted. In order to integrate the weld quality information at different time points and conduct a comprehensive analysis, the multiple weld monitoring key frames can be passed through the weld quality feature encoder to obtain a dynamic feature map of the weld quality. The essence of the weld quality feature encoder in this application is a convolutional neural network model using a time attention mechanism. The time attention mechanism is a technology widely used in sequence data processing. It allows the model to assign different weights to different time steps when processing sequence data, so as to pay more attention to those time steps that are more important to the task. In the weld quality feature encoder, the time attention mechanism can help the model pay more attention to key frames containing important weld quality information. Guided by the time attention mechanism, the model can more accurately capture the dynamic change characteristics of the weld quality, thereby improving the detection accuracy. It is also worth mentioning that there are various interference factors (such as noise, light changes, etc.) in the welding process, and traditional convolutional neural network models may find it difficult to accurately extract weld quality features. The time attention mechanism can help the model ignore these interference factors and enhance the robustness of the model.
[0039] In an embodiment of the present application, an implementation method of passing the multiple weld monitoring key frames through a weld quality feature encoder to obtain a weld quality dynamic feature map may be: extracting adjacent first weld monitoring key frames and second weld monitoring key frames from the multiple weld monitoring key frames; passing the first weld monitoring key frame and the second weld monitoring key frame through the first convolution layer and the second convolution layer of the weld quality feature encoder respectively to obtain a first convolution feature map and a second convolution feature map; calculating the position point multiplication between the first convolution feature map and the second convolution feature map to obtain a time attention map; inputting the time attention map into a Softmax activation function to obtain a time attention feature map; passing the second weld monitoring key frame through the third convolution layer of the weld quality feature encoder to obtain a third convolution feature map; calculating the position point multiplication between the third convolution feature map and the time attention feature map to obtain a time attention feature map corresponding to the second weld monitoring key frame.
[0040] The dynamic feature map of weld quality usually contains a large amount of feature information. Directly processing these high-dimensional feature maps requires huge computing resources and may lead to slow model training. In order to reduce the feature dimension and reduce the amount of calculation, the dynamic feature map of weld quality can be pooled along the channel dimension to obtain the dynamic feature vector of weld quality. The pooling operation downsamples the dynamic feature map of weld quality to reduce its spatial resolution, thereby reducing the feature dimension. While reducing the feature dimension, it can also retain important information in the feature map, namely the key features.
[0041] In particular, in the technical solution of the present application, it is considered that the weld process parameter associated time series feature vector is obtained based on the time series data of process parameters such as the current value and wire feeding speed value during the welding process, while the weld quality dynamic feature vector is extracted based on the key frames of the weld monitoring image. The differences in these two data types and sources may lead to inconsistencies in their feature expressions. The welding process parameter time series feature encoder and the weld quality feature encoder may use different feature extraction methods, which may lead to differences in the information captured in the feature vector, and further lead to the lack of posterior-prior causal association between the weld quality dynamic feature vector and the weld process parameter associated time series feature vector. In order to solve this problem, in the technical solution of the present application, based on the weld quality dynamic feature vector, the weld process parameter associated time series feature vector is enhanced by causal dependence based on multi-angle feature association measurement to obtain an optimized weld process parameter associated time series feature vector.
[0042] Specifically, in an embodiment of the present application, based on the dynamic feature vector of weld quality, the causal dependency enhancement based on the multi-angle feature association measurement is performed on the weld process parameter associated time series feature vector to obtain an optimized weld process parameter associated time series feature vector, including: calculating the position difference between the weld process parameter associated time series feature vector and the weld quality dynamic feature vector to obtain a first viewpoint measurement representation vector; calculating the position point multiplication between the weld process parameter associated time series feature vector and the weld quality dynamic feature vector to obtain a second viewpoint measurement representation vector; calculating the position sum between the weld process parameter associated time series feature vector and the weld quality dynamic feature vector to obtain to a third viewpoint metric representation vector; calculate the gradient amplitude of each position in the first viewpoint metric representation vector, the second viewpoint metric representation vector and the third viewpoint metric representation vector to obtain a first viewpoint metric perception vector, a second viewpoint metric perception vector and a third viewpoint metric perception vector; input the first viewpoint metric perception vector, the second viewpoint metric perception vector and the third viewpoint metric perception vector into a backward machine learning inverse multi-viewpoint perception optimization module to obtain a reverse multi-viewpoint perception optimization representation vector; calculate the position point multiplication between the reverse multi-viewpoint perception optimization representation vector and the weld process parameter associated time series feature vector to obtain the optimized weld process parameter associated time series feature vector.
[0043] More specifically, in the embodiment of the present application, the backward machine learning reverse multi-view perception optimization module is expressed by the formula:
[0044]
[0045] in, and represents the eigenvalues of the i-th position of the first viewpoint metric perception vector, the second viewpoint metric perception vector and the third viewpoint metric perception vector, respectively, and μ represents the three eigenvalues of the corresponding positions in the first viewpoint metric perception vector, the second viewpoint metric perception vector and the third viewpoint metric perception vector , σ is the standard deviation of the three eigenvalues at corresponding positions in the first viewpoint metric perception vector, the second viewpoint metric perception vector and the third viewpoint metric perception vector, represents the maximum value of the three eigenvalues at the corresponding positions in the first viewpoint metric perception vector, the second viewpoint metric perception vector and the third viewpoint metric perception vector, ReLU represents a linear rectification function, softmax represents a normalized exponential function, and v ri represents the eigenvalue of the i-th position of the inverse multi-view perceptual optimization representation vector, and exp(·) represents an exponential operation with the natural constant e as the base.
[0046] Specifically, in the technical solution of the present application, based on the dynamic feature vector of weld quality, the causal dependency enhancement optimization based on multi-angle feature association measurement is performed on the time series feature vector associated with weld process parameters, and the process includes firstly calculating the position difference, point multiplication and sum between the time series feature vector associated with weld process parameters and the dynamic feature vector of weld quality through multi-view analysis (Multi-View Analysis) to obtain three viewpoint measurement representation vectors, and quantifying the relationship between features from different angles. Then, the gradient amplitude of each position in these viewpoint measurement representation vectors is calculated through backward error perception to identify key change points to provide a basis for feature optimization.
[0047] Then, the first viewpoint metric perception vector, the second viewpoint metric perception vector and the third viewpoint metric perception vector are input into the backward machine learning reverse multi-viewpoint perception optimization module, wherein the backward machine learning reverse multi-viewpoint perception optimization module captures the significant comprehensive ability factor of the feature importance perception (Feature Importance Perception) of the feature sub-dimension distribution of the corresponding position in the first to third viewpoint metric perception vectors by adopting a squeeze-incentive mechanism composed of -functions. Then, the position point multiplication between the reverse multi-viewpoint perception optimization representation vector and the weld process parameter associated time series feature vector is calculated to perform reverse optimization perception (Backward Optimization Perception) on the weld process parameter associated time series feature vector to adjust the weld process parameter associated time series feature vector so as to better associate with the weld quality dynamic feature vector, thereby enhancing the causal relationship between the weld process parameter associated time series feature vector and the weld quality dynamic feature vector.
[0048] Next, the responsiveness estimation of the dynamic characteristic vector of weld quality relative to the time series characteristic vector associated with the optimized weld process parameters is calculated to obtain the weld quality responsiveness characteristic matrix. It should be understood that responsiveness estimation refers to the sensitivity and dependence of the output data in a module on its input data by calculation and quantification. That is, by calculating the responsiveness estimation between the dynamic characteristic vector of weld quality and the time series characteristic vector associated with the optimized weld process parameters, it is possible to quantitatively evaluate how the weld quality changes with the change of process parameters, which helps to deeply understand the influence mechanism of process parameters on weld quality during welding. Based on the results of responsiveness estimation, welding parameters can be adjusted to improve weld quality and welding efficiency.
[0049] In an embodiment of the present application, a possible implementation method for calculating the responsiveness estimation of the dynamic characteristic vector of the weld quality relative to the time series characteristic vector associated with the optimized weld process parameters to obtain the weld quality responsiveness characteristic matrix may be: calculating the responsiveness estimation of the dynamic characteristic vector of the weld quality relative to the time series characteristic vector associated with the optimized weld process parameters using the following responsiveness estimation formula to obtain the weld quality responsiveness characteristic matrix; wherein the responsiveness estimation formula is:
[0050]
[0051] Where V c represents the dynamic characteristic vector of the weld quality, V d represents the time series feature vector associated with the optimized welding process parameters, M a represents the weld quality responsiveness characteristic matrix, Represents matrix multiplication.
[0052] Finally, the weld quality responsiveness feature matrix is passed through the current adjustment classifier to obtain a classification result, and the classification result is used to indicate whether the pulse current value of the welder at the current time point should be increased, decreased, or kept unchanged. It should be understood that the current classifier is a classification model based on machine learning, which can perform classification predictions based on the input features. In the welding equipment monitoring system, the role of the current classifier is to automatically determine whether the current welding current needs to be adjusted based on the real-time feedback of the weld quality, and give specific adjustment suggestions. The classification result directly guides the adjustment of the welding current. When the classifier determines that the current welding current needs to be increased, the system will automatically adjust the current value to deepen the weld penetration; when it is determined that it needs to be reduced, the current value is reduced to prevent overheating or burn-through; when it is determined that it should remain unchanged, the current value is maintained to ensure the stability of the welding quality. Through the precise control of the current classifier, it can be ensured that the current adjustment during the welding process is always kept within the optimal range, thereby improving the consistency and stability of the weld quality.
[0053] In summary, the welding equipment monitoring system 100 based on the Internet of Things technology according to the embodiment of the present application is explained, which adopts the data analysis technology based on artificial intelligence, extracts the associated time series features of the pulse current value of the welding machine and the wire feeding speed value of the wire feeder at multiple time points to obtain the time series features of the welding process parameters, and extracts the time series features of the weld monitoring image in a certain time period to obtain the weld quality features, and then establishes the responsiveness relationship between the weld quality features and the welding process parameter time series features, and finally outputs the direction of the pulse current adjustment of the welding machine through the classifier based on the established responsiveness features. In this way, the precise control of the welding pulse current can be achieved, which helps to improve the welding quality.
[0054] As described above, the welding equipment monitoring system 100 based on the Internet of Things technology according to the embodiment of the present application can be implemented in various terminal devices, such as a server for welding equipment monitoring based on the Internet of Things technology. In one example, the welding equipment monitoring system 100 based on the Internet of Things technology according to the embodiment of the present application can be integrated into the terminal device as a software module and / or a hardware module. For example, the welding equipment monitoring system 100 based on the Internet of Things technology can be a software module in the operating system of the terminal device, or can be an application developed for the terminal device; of course, the welding equipment monitoring system 100 based on the Internet of Things technology can also be one of the many hardware modules of the terminal device.
[0055] Alternatively, in another example, the welding equipment monitoring system 100 based on the Internet of Things technology and the terminal device may also be separate devices, and the welding equipment monitoring system 100 based on the Internet of Things technology may be connected to the terminal device via a wired and / or wireless network, and transmit interactive information in accordance with an agreed data format.
[0056] Figure 5 FIG. 1 is a flow chart of a welding equipment monitoring method based on Internet of Things technology according to an embodiment of the present application. Figure 5 As shown, in the welding equipment monitoring method based on the Internet of Things technology, the method includes: S110, obtaining the welding machine pulse current value and the wire feeding speed value of the wire feeder at multiple predetermined time points within a predetermined time period and the weld monitoring image within the predetermined time period; S120, feature encoding the welding machine pulse current value and the wire feeding speed value of the wire feeder at the multiple predetermined time points and the weld monitoring image within the predetermined time period to obtain a weld quality responsiveness feature matrix; S130, judging the welding machine pulse current adjustment direction at the current time point based on the weld quality responsiveness feature matrix.
[0057] Here, those skilled in the art can understand that the specific operations of each step in the above-mentioned welding equipment monitoring method based on the Internet of Things technology have been referred to above. Figures 1 to 4 The description of the welding equipment monitoring system based on the Internet of Things technology has been introduced in detail, and therefore, its repeated description will be omitted.
[0058] In summary, the welding equipment monitoring method based on the Internet of Things technology according to the embodiment of the present application is explained, which adopts the data analysis technology based on artificial intelligence, extracts the associated time series features of the pulse current value of the welding machine and the wire feeding speed value of the wire feeder at multiple time points to obtain the time series features of the welding process parameters, and extracts the time series features of the weld monitoring image in a certain time period to obtain the weld quality features, and then establishes the responsiveness relationship between the weld quality features and the welding process parameter time series features, and finally outputs the direction of the pulse current adjustment of the welding machine through the classifier based on the established responsiveness features. In this way, the precise control of the welding pulse current can be achieved, which helps to improve the welding quality.
Claims
1. A welding equipment monitoring system based on Internet of Things technology, characterized in that: include: A welding-related data acquisition module, used to obtain the pulse current value of the welding machine and the wire feeding speed value of the wire feeder at multiple predetermined time points within a predetermined time period and the weld monitoring image within the predetermined time period; A welding-related data feature encoding module, used for feature encoding the pulse current values of the welding machine and the wire feeding speed values of the wire feeder at the plurality of predetermined time points and the weld monitoring image within the predetermined time period to obtain a weld quality responsiveness feature matrix; The welding machine pulse current adjustment result generation module is used to determine the welding machine pulse current adjustment direction at the current time point based on the weld quality responsiveness characteristic matrix.
2. The welding equipment monitoring system based on Internet of Things technology according to claim 1 is characterized in that: The welding-related data feature coding module comprises: A welding process parameter time series feature extraction unit, used for extracting time series features of the welding machine pulse current values and the wire feeding speed values of the wire feeder at the plurality of predetermined time points to obtain a welding process parameter associated time series feature vector; A weld quality dynamic feature extraction unit, used for extracting features from the weld monitoring image within the predetermined time period to obtain a weld quality dynamic feature vector; A welding process parameter time series feature optimization unit is used to enhance the causal dependency of the welding process parameter associated time series feature vector based on the multi-angle feature association metric based on the weld quality dynamic feature vector to obtain an optimized welding process parameter associated time series feature vector; A responsiveness estimation unit is used to calculate the responsiveness estimation of the dynamic characteristic vector of the weld quality relative to the time series characteristic vector associated with the optimized weld process parameters to obtain the weld quality responsiveness characteristic matrix.
3. The welding equipment monitoring system based on Internet of Things technology according to claim 2 is characterized in that: The welding process parameter time series feature extraction unit comprises: The welding process parameter data structuring subunit is used to arrange the welding machine pulse current values and the wire feeding speed values of the wire feeder at the plurality of predetermined time points according to the time dimension to obtain a pulse current timing input vector and a wire feeding speed timing input vector; A welding process parameter vector association subunit, used for associating and encoding the pulse current timing input vector and the wire feeding speed timing input vector to obtain a welding process parameter associated timing input matrix; The welding process parameter timing feature generating subunit is used to pass the welding process parameter associated timing input matrix through the welding process parameter timing feature encoder to obtain the welding process parameter associated timing feature vector.
4. The welding equipment monitoring system based on Internet of Things technology according to claim 3 is characterized in that: The weld quality dynamic feature extraction unit comprises: A video key frame sampling subunit, used for sampling the weld monitoring video of the predetermined time period to obtain a plurality of weld monitoring key frames; A weld quality dynamic feature generation subunit, used for passing the plurality of weld monitoring key frames through a weld quality feature encoder to obtain a weld quality dynamic feature map; The weld quality dynamic feature dimensionality reduction subunit is used to pool the feature matrices of the weld quality dynamic feature map along the channel dimension to obtain the weld quality dynamic feature vector.
5. The welding equipment monitoring system based on Internet of Things technology according to claim 4 is characterized in that: The welding process parameter temporal feature encoder is a convolutional neural network model using a two-dimensional convolution kernel, and the weld quality feature encoder is a convolutional neural network model using a temporal attention mechanism.
6. The welding equipment monitoring system based on Internet of Things technology according to claim 5 is characterized in that: The welding process parameter timing characteristic optimization unit is used to: Calculating the position difference between the weld process parameter associated time series feature vector and the weld quality dynamic feature vector to obtain a first viewpoint metric representation vector; Calculate the position point multiplication between the weld process parameter associated time series feature vector and the weld quality dynamic feature vector to obtain a second viewpoint metric representation vector; Calculating the positional sum of the weld process parameter associated time series feature vector and the weld quality dynamic feature vector to obtain a third viewpoint metric representation vector; Calculating the gradient magnitude of each position in the first viewpoint metric representation vector, the second viewpoint metric representation vector, and the third viewpoint metric representation vector to obtain a first viewpoint metric perception vector, a second viewpoint metric perception vector, and a third viewpoint metric perception vector; Inputting the first viewpoint metric perception vector, the second viewpoint metric perception vector and the third viewpoint metric perception vector into a backward machine learning inverse multi-view perception optimization module to obtain a reverse multi-view perception optimization representation vector; The optimized weld process parameter associated time series feature vector is obtained by calculating the position point multiplication between the reverse multi-view perception optimization representation vector and the weld process parameter associated time series feature vector.
7. The welding equipment monitoring system based on Internet of Things technology according to claim 6 is characterized in that: The welding machine pulse current adjustment result generation module is used to: pass the weld quality responsiveness characteristic matrix through the current adjustment classifier to obtain a classification result, and the classification result is used to indicate whether the welding machine pulse current value at the current time point should increase, decrease or remain unchanged.
8. A welding equipment monitoring method based on Internet of Things technology, characterized in that: include: Acquire the pulse current value of the welding machine and the wire feeding speed value of the wire feeder at multiple predetermined time points within a predetermined time period and the weld monitoring image within the predetermined time period; Performing feature coding on the pulse current values of the welding machine and the wire feeding speed values of the wire feeder at the plurality of predetermined time points and the weld monitoring image within the predetermined time period to obtain a weld quality responsiveness feature matrix; Based on the weld quality responsiveness characteristic matrix, the pulse current adjustment direction of the welding machine at the current time point is determined.
9. The method for monitoring welding equipment based on Internet of Things technology according to claim 8, characterized in that: The pulse current values of the welding machine and the wire feeding speed values of the wire feeder at the plurality of predetermined time points and the weld monitoring image within the predetermined time period are feature-encoded to obtain a weld quality responsiveness feature matrix, including: Extracting time series features of the pulse current values of the welding machine and the wire feeding speed values of the wire feeder at the plurality of predetermined time points to obtain a time series feature vector associated with the welding process parameters; Extracting features from the weld monitoring image within the predetermined time period to obtain a dynamic feature vector of weld quality; Based on the weld quality dynamic feature vector, the weld process parameter associated time series feature vector is enhanced by causal dependency based on multi-angle feature association measurement to obtain an optimized weld process parameter associated time series feature vector; The responsiveness estimate of the weld quality dynamic characteristic vector relative to the optimized weld process parameter associated time series characteristic vector is calculated to obtain the weld quality responsiveness characteristic matrix.
10. The method for monitoring welding equipment based on Internet of Things technology according to claim 9, characterized in that: Based on the weld quality responsiveness characteristic matrix, the adjustment direction of the welder pulse current at the current time point is judged, including: passing the weld quality responsiveness characteristic matrix through a current adjustment classifier to obtain a classification result, and the classification result is used to indicate whether the welder pulse current value at the current time point should increase, decrease or remain unchanged.
Citation Information
Cited By
Remote medical system and method based on 5G technology
CN119338743A
Intelligent control system and method for digital pulse consumable electrode welding machine
CN119525654A
