Intelligent control system and method for digital pulse consumable electrode welding machine

Abstract

The invention provides an intelligent control system and method for a digital pulse consumable electrode welding machine, and relates to the field of welding machine control. Wire feeding speed values at multiple time points are associated with actual welding current values of the welding machine to obtain welding current-wire feeding speed matching vectors; meanwhile, the wire feeding speed values at the multiple time points are associated with the actual welding voltage value of the welding machine to obtain a welding voltage-wire feeding speed matching vector; and then multi-scale analysis is conducted on a welding process parameter-wire feeding speed matching vector obtained by fusing the welding current-wire feeding speed matching vector and the welding voltage-wire feeding speed matching vector, and finally the welding machine current and voltage adjusting direction needing to be controlled is output through a classifier. Thus, accurate control over the welding current and voltage can be achieved, the complex control requirement in the arc striking stage can be met, and then the welding stability and quality are improved.

Description

Technical Field

[0001] The present application relates to the field of welding machine control, and more specifically, to an intelligent control system and method for a digital pulse consumable electrode welding machine. Background Art

[0002] Pulse welding is an advanced welding process. The welding wire is heated by a pulse and melted to form a molten droplet, which falls off onto the base material to form a molten pool. The metal in the molten pool stays at high temperature for a short time, and the metal condenses quickly, reducing the tendency of heat-sensitive materials to crack.

[0003] The arc starting stage of welding refers to the moment when the arc begins to be established during the welding process, that is, the transition stage from no arc to the formation of a stable arc. The arc starting stage is a complex control process, involving the matching of various aspects such as wire feeding speed, current command, and voltage command. The arc starting stage of pulse gas shielded welding will affect the weld quality and welding efficiency, and the arc starting stage needs to be strictly controlled.

[0004] Therefore, a digital pulse consumable electrode welding machine control solution is needed. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present application provides an intelligent control system and method for a digital pulse consumable electrode welding machine.

[0006] According to one aspect of the present application, a digital pulse consumable electrode welding machine intelligent control system is provided, which includes:

[0007] The welding process related parameter data acquisition module is used to obtain the wire feeding speed value, the actual welding current value of the welding machine, and the actual welding voltage value of the welding machine at multiple predetermined time points within a predetermined time period;

[0008] A welding process related parameter structuring module, used for performing data regularization on the wire feeding speed values, actual welding current values ​​of the welding machine, and actual welding voltage values ​​of the welding machine at the plurality of predetermined time points to obtain a wire feeding speed timing input vector, a welding current timing input vector of the welding machine, and a welding voltage timing input vector of the welding machine;

[0009] A welding process parameter-wire feeding speed matching module is used to perform vector matching analysis on the wire feeding speed timing input vector, the welding current timing input vector of the welding machine and the welding voltage timing input vector of the welding machine to obtain a welding process parameter-wire feeding speed matching vector;

[0010] A welding process parameter-wire feeding speed matching feature encoding module, used for performing feature encoding on the welding process parameter-wire feeding speed matching vector to obtain a welding process parameter-wire feeding speed matching feature vector;

[0011] The welding process parameter adjustment result generation module is used to obtain the adjustment direction of the actual welding current and the actual welding voltage of the spot welder at the current time based on the welding process parameter-wire feeding speed matching feature vector.

[0012] According to another aspect of the present application, there is provided an intelligent control method for a digital pulse consumable electrode welding machine, which includes:

[0013] Obtain wire feeding speed values, actual welding current values, and actual welding voltage values ​​of the welding machine at multiple predetermined time points within a predetermined time period;

[0014] Performing data regularization on the wire feeding speed values, actual welding current values ​​of the welding machine, and actual welding voltage values ​​of the welding machine at the plurality of predetermined time points to obtain a wire feeding speed timing input vector, a welding current timing input vector, and a welding voltage timing input vector of the welding machine;

[0015] Performing vector matching analysis on the wire feeding speed timing input vector, the welding current timing input vector of the welding machine, and the welding voltage timing input vector of the welding machine to obtain a welding process parameter-wire feeding speed matching vector;

[0016] Characteristic encoding is performed on the welding process parameter-wire feeding speed matching vector to obtain a welding process parameter-wire feeding speed matching characteristic vector;

[0017] Based on the welding process parameter-wire feeding speed matching feature vector, the adjustment direction of the actual welding current of the spot welder and the actual welding voltage of the welder at the current time is obtained.

[0018] This application has significant technical effects due to the adoption of the above technical solutions:

[0019] The digital pulse consumable electrode welding machine intelligent control system and method provided by the present application first associates the wire feeding speed values ​​at multiple time points with the actual welding current value of the welding machine to obtain a welding current-wire feeding speed matching vector, and at the same time associates the wire feeding speed values ​​at multiple time points with the actual welding voltage value of the welding machine to obtain a welding voltage-wire feeding speed matching vector, and then performs a multi-scale analysis on the welding process parameter-wire feeding speed matching vector obtained by merging the welding current-wire feeding speed matching vector and the welding voltage-wire feeding speed matching vector, and finally outputs the welding machine current and voltage adjustment direction that need to be controlled through a classifier. In this way, accurate control of the welding current and voltage can be achieved, so that the complex control requirements of the arc initiation stage can be met, thereby improving the stability and quality of welding. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0021] Figure 1 It is a system block diagram of an intelligent control system of a digital pulse consumable electrode welding machine according to an embodiment of the present application.

[0022] Figure 2 It is a block diagram of a welding process parameter-wire feeding speed matching module in an intelligent control system of a digital pulse consumable electrode welding machine according to an embodiment of the present application.

[0023] Figure 3 It is a block diagram of a welding process parameter adjustment result generation module in an intelligent control system of a digital pulse consumable electrode welding machine according to an embodiment of the present application.

[0024] Figure 4 It is a flow chart of an intelligent control method of a digital pulse consumable electrode welding machine according to an embodiment of the present application. DETAILED DESCRIPTION

[0025] 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.

[0026] Pulse welding is an advanced welding process. In this process, the welding wire is melted by pulse heating to form molten droplets, which fall off and form a molten pool on the parent material. Since the molten pool metal stays at high temperature for a short time and condenses quickly, it can reduce the possibility of cracks in heat-sensitive materials.

[0027] The arc striking stage of welding refers to the moment when the arc begins to be established during the welding process, that is, the stage of transitioning from the no-arc state to the formation of a stable arc. The control of the arc striking stage is relatively complex, and it is necessary to match the wire feeding speed, current command, voltage command and other aspects. For pulse gas shielded welding, the arc striking stage will affect the weld quality and welding efficiency, and this stage needs to be strictly controlled. Therefore, a digital pulse consumable electrode welding machine control solution is expected.

[0028] 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 the intelligent control of digital pulse consumable welding machines.

[0029] Figure 1 FIG. 1 is a system block diagram of an intelligent control system for a digital pulse consumable electrode welding machine according to an embodiment of the present application. Figure 1 As shown, in the digital pulse consumable electrode welding machine intelligent control system 100, it includes: a welding process related parameter data acquisition module 110, which is used to obtain the wire feeding speed value, the actual welding current value of the welding machine, and the actual welding voltage value of the welding machine at multiple predetermined time points within a predetermined time period; a welding process related parameter structuring module 120, which is used to regularize the wire feeding speed value, the actual welding current value of the welding machine, and the actual welding voltage value of the welding machine at the multiple predetermined time points to obtain the wire feeding speed timing input vector, the welding current timing input vector of the welding machine, and the welding voltage timing input vector of the welding machine; a welding process parameter-wire feeding speed matching module 130, which is used to The module 140 is used to perform vector matching analysis on the wire feeding speed timing input vector, the welding current timing input vector and the welding voltage timing input vector to obtain a welding process parameter-wire feeding speed matching vector; the welding process parameter-wire feeding speed matching feature encoding module 140 is used to perform feature encoding on the welding process parameter-wire feeding speed matching vector to obtain a welding process parameter-wire feeding speed matching feature vector; the welding process parameter adjustment result generation module 150 is used to obtain the adjustment direction of the actual welding current and the actual welding voltage of the welding machine at the current time point based on the welding process parameter-wire feeding speed matching feature vector.

[0030] In an embodiment of the present application, the welding process related parameter data acquisition module 110 is used to obtain the wire feeding speed value, the actual welding current value of the welding machine, and the actual welding voltage value of the welding machine at multiple predetermined time points within a predetermined time period. It should be understood that the welding process is a dynamically changing process, and parameters such as wire feeding speed, welding current and welding voltage will continue to fluctuate as welding proceeds. In order to accurately capture this dynamic change and effectively control it, it is necessary to obtain data at multiple time points, which can reflect the real-time state of the welding process and provide a basis for subsequent parameter adjustment. And in the actual welding process, there is a complex interaction relationship between wire feeding speed, welding current and welding voltage. By obtaining data at multiple time points, a matching relationship between these parameters can be established, and they can be optimized and controlled, thereby improving the stability and quality of welding.

[0031] In the embodiment of the present application, the welding process related parameter structuring module 120 is used to perform data regularization on the wire feeding speed values, the actual welding current value of the welding machine, and the actual welding voltage value of the welding machine at the multiple predetermined time points to obtain the wire feeding speed timing input vector, the welding current timing input vector of the welding machine, and the welding voltage timing input vector of the welding machine. Specifically, in the embodiment of the present application, the welding process related parameter structuring module is used to: arrange the wire feeding speed values, the actual welding current value of the welding machine, and the actual welding voltage value of the welding machine at the multiple predetermined time points according to the time dimension to obtain the wire feeding speed timing input vector, the welding current timing input vector of the welding machine, and the welding voltage timing input vector of the welding machine. Accordingly, considering that each parameter (such as wire feeding speed, welding current, welding voltage) will change over time during the welding process, in order to clearly capture the changing trend and dynamic characteristics of these parameters during the welding process, in the technical solution of the present application, the wire feeding speed values, the actual welding current value of the welding machine, and the actual welding voltage value of the welding machine at multiple predetermined time points need to be arranged according to the time dimension. By arranging them according to the time dimension, the time correlation between the various parameters can be established in the subsequent data analysis, which helps to understand the interaction between the various parameters in the welding process.

[0032] In the embodiment of the present application, the welding process parameter-wire feeding speed matching module 130 is used to perform vector matching analysis on the wire feeding speed timing input vector, the welding current timing input vector of the welding machine, and the welding voltage timing input vector of the welding machine to obtain a welding process parameter-wire feeding speed matching vector. Specifically, Figure 2 FIG. 1 is a block diagram of a welding process parameter-wire feeding speed matching module in an intelligent control system of a digital pulse consumable electrode welding machine according to an embodiment of the present application. Figure 2As shown, the welding process parameter-wire feeding speed matching module 130 includes: a wire feeding speed timing feature extraction unit 131, which is used to pass the wire feeding speed timing input vector through a wire feeding speed timing encoder to obtain a wire feeding speed timing feature vector; a welding machine welding current timing feature extraction unit 132, which is used to pass the welding machine welding current timing input vector through a welding current timing encoder to obtain a welding machine welding current timing feature vector; a welding machine welding voltage timing feature extraction unit 133, which is used to pass the welding machine welding voltage timing input vector through a welding voltage timing encoder to obtain a welding machine welding voltage timing feature vector; a welding current-wire feeding speed matching unit 134, used for multiplying the wire feeding speed timing characteristic vector and the welding machine welding current timing characteristic vector by position point to obtain a welding current-wire feeding speed matching vector; a welding current-wire feeding speed matching unit 135, used for multiplying the wire feeding speed timing characteristic vector and the welding machine welding voltage timing characteristic vector by position point to obtain a welding voltage-wire feeding speed matching vector; a welding process parameter-wire feeding speed matching fusion generation unit 136, used for performing intrinsic structure-guided feature matching on the welding current-wire feeding speed matching vector and the welding voltage-wire feeding speed matching vector to obtain the welding process parameter-wire feeding speed matching vector.

[0033] Considering that the wire feeding speed timing input vector arranged according to the time dimension contains all the information about the wire feeding speed changing over time, this information is often redundant. In order to automatically extract the characteristic information useful for welding process control from the original data and better reflect the dynamic characteristics of the wire feeding speed in the welding process, it is necessary to pass the wire feeding speed timing input vector through a wire feeding speed timing encoder to obtain a wire feeding speed timing characteristic vector. The wire feeding speed timing encoder obtained through training can learn the general laws and characteristics of the wire feeding speed changes during the welding process. In this way, when facing a new welding task, the encoder can quickly and accurately extract features from the new wire feeding speed timing input vector based on existing knowledge and experience, which is conducive to realizing intelligent control of the welding process. In the present application, the wire feeding speed timing encoder is a convolutional neural network model including a fully connected layer and a one-dimensional convolutional layer. It should be understandable that since the changes in wire feeding speed during the welding process often have certain local correlations, the one-dimensional convolutional layer can effectively extract the local features in the wire feeding speed timing input vector through the local convolutional encoding operation of the convolution kernel, and the fully connected layer can integrate the local features extracted by the one-dimensional convolutional layer to obtain the final wire feeding speed timing feature vector.

[0034] The welding current timing input vector contains a large number of raw data welding current data points. In order to extract the timing information of the welding current change from these raw current data points, it is necessary to pass the welding machine welding current timing input vector through the welding current timing encoder to obtain the welding machine welding current timing feature vector. After being processed by the welding current timing encoder, the original data with redundant timing information is converted into a concise and representative welding current timing feature vector. In this application, the welding current timing encoder is a convolutional neural network model including a fully connected layer and a one-dimensional convolutional layer. It should be understood that the one-dimensional convolutional layer can capture the changing trend and characteristics of the welding current in a short time by sliding the one-dimensional convolution kernel in the original timing input vector, and the fully connected layer can integrate the local features in the convolution layer, and the activation function (such as ReLU, etc.) in the fully connected layer can introduce nonlinear factors, so that the current timing encoder can capture the nonlinear relationship in the welding current timing input vector, thereby improving the expression ability of the generated feature vector.

[0035] The welding voltage timing data contains a large amount of original information, which includes both features useful for welding process control and redundancy and noise. In order to extract key features from the original voltage timing input vector and realize data dimensionality reduction, so that subsequent processing is more efficient, it is necessary to pass the welding voltage timing input vector of the welder through the welding voltage timing encoder to obtain the welding voltage timing feature vector of the welder. The welding voltage timing feature vector extracted by the welding voltage timing encoder can more accurately reflect the dynamic changes of voltage during the welding process, thereby providing more accurate information input for the control system, which helps to improve welding quality and efficiency. In this application, the welding voltage timing encoder is a convolutional neural network model including a fully connected layer and a one-dimensional convolutional layer. This structure combines the local feature extraction capability of the convolutional neural network and the global feature integration capability of the fully connected layer, and is suitable for processing welding voltage timing data.

[0036] Welding current and wire feeding speed are two important parameters that are interrelated and influence each other in the welding process. In order to reveal the interaction between these two parameters, in the technical solution of the present application, it is necessary to multiply the wire feeding speed timing characteristic vector and the welding current timing characteristic vector of the welding machine by position points to obtain the welding current-wire feeding speed matching vector. It should be well understood that the welding current-wire feeding speed matching vector not only reflects the change trend of the respective parameters of the welding current and the wire feeding speed, but also reveals the matching degree and coordination between them.

[0037] The wire feeding speed directly affects the feeding rate of the welding material, while the welding voltage determines the stability and energy input of the arc. There is a close interdependence between these two parameters. The change of one parameter often affects the other parameter. In order to capture the correlation between the two parameters, it is necessary to multiply the wire feeding speed timing characteristic vector and the welding voltage timing characteristic vector of the welding machine by position to obtain the welding voltage-wire feeding speed matching vector. By multiplying the timing characteristic vectors of these two parameters by position, this correlation can be captured, thereby providing more accurate information for subsequent control strategies.

[0038] The welding process is a complex process with multiple parameters, multiple variables and nonlinearity. Welding current, welding voltage and wire feeding speed are the three most critical parameters, and there is a close connection and mutual influence between them. Considering any one parameter alone cannot fully reflect the true state of the welding process. In order to comprehensively consider the interaction and matching relationship between welding current, welding voltage and wire feeding speed, it is necessary to fuse the welding current-wire feeding speed matching vector and the welding voltage-wire feeding speed matching vector in the technical solution of the present application to obtain the welding process parameter-wire feeding speed matching vector. By fusing these two characteristic vectors, the control system can more accurately identify the unstable factors in the welding process and take corresponding measures to make adjustments, thereby improving the stability and control accuracy of the welding process.

[0039] In particular, in the technical solution of the present application, it is considered that the welding current-wire feeding speed matching vector (reflecting the relationship between current and wire feeding speed) and the welding voltage-wire feeding speed matching vector (reflecting the relationship between voltage and wire feeding speed) may have different numerical ranges. For example, the units and magnitudes of welding current and welding voltage are different, and direct merging may cause one feature to have a much greater impact on the result than the other, even if this impact does not reflect the true situation of the data. This will cause the model training to favor certain features and ignore other equally important features. The welding current-wire feeding speed matching vector captures the interaction between current and wire feeding speed, reflecting the energy supply during the welding process; while the welding voltage-wire feeding speed matching vector focuses more on the relationship between voltage and wire feeding speed, reflecting arc stability and heat input. These two matching vectors represent different aspects of information in the welding process. Direct merging without considering their semantic differences may confuse the model's understanding of this information and fail to accurately distinguish the respective effects of current and voltage and their complex relationship with wire feeding speed. The data of welding current and welding voltage are both time series data, but their change patterns and frequencies may be different. The welding current may fluctuate more frequently in response to changes in wire feeding speed, while the change in welding voltage may be relatively gentle. Directly merging data with different time characteristics may destroy the original time series relationship, cause information distortion, and affect the correct understanding of the dynamic characteristics of the welding process. In order to solve this problem, in the technical solution of this application, the welding current-wire feeding speed matching vector and the welding voltage-wire feeding speed matching vector are subjected to intrinsic structure-guided feature matching.

[0040] Specifically, in an embodiment of the present application, the welding process parameter-wire feeding speed matching fusion generation unit includes: a vector modulus normalization feature coordination subunit, which is used to perform modulus normalization feature coordination on the welding current-wire feeding speed matching vector and the welding voltage-wire feeding speed matching vector to obtain a pre-aligned welding current-wire feeding speed matching vector and a pre-aligned welding voltage-wire feeding speed matching vector; a vector essential information extraction subunit, which is used to perform essential information extraction on the pre-aligned welding current-wire feeding speed matching vector and the pre-aligned welding voltage-wire feeding speed matching vector to obtain a first principal component pre-aligned welding current-wire feeding speed matching vector and a second principal component pre-aligned welding voltage-wire feeding speed matching vector; a vector feature detail precise matching subunit, which is used to perform feature detail precise matching on the first principal component pre-aligned welding current-wire feeding speed matching vector and the second principal component pre-aligned welding voltage-wire feeding speed matching vector to obtain the welding process parameter-wire feeding speed matching vector.

[0041] More specifically, in an embodiment of the present application, the vector modulus normalized feature coordination subunit is used to: calculate the square of the second norm of the welding current-wire feeding speed matching vector to obtain the welding current-wire feeding speed feature intensity value; calculate the natural exponential function value for each eigenvalue of the welding current-wire feeding speed matching vector to obtain an indexed welding current-wire feeding speed matching vector; divide each eigenvalue of the indexed welding current-wire feeding speed matching vector by the sum of the exponential welding current-wire feeding speed matching vector to obtain a normalized welding current-wire feeding speed matching vector; multiply each eigenvalue of the normalized welding current-wire feeding speed matching vector by the welding current-wire feeding speed feature intensity value to obtain the pre-aligned welding current-wire feeding speed matching vector.

[0042] More specifically, in an embodiment of the present application, the vector essential information extraction subunit is used to: perform eigenvalue-based decomposition of the pre-aligned welding current-wire feeding speed matching vector to obtain a first intrinsic decomposition pre-aligned welding current-wire feeding speed matching vector sequence and a first diagonal matrix; cascade the first eigenvector and the second eigenvector in the first intrinsic decomposition pre-aligned welding current-wire feeding speed matching vector sequence to obtain the first principal component pre-aligned welding current-wire feeding speed matching vector.

[0043] In the embodiment of the present application, specifically, the formula expression of the welding process parameter-wire feeding speed matching fusion generation unit can be: the welding current-wire feeding speed matching vector and the welding voltage-wire feeding speed matching vector are processed by the following formula to obtain the welding process parameter-wire feeding speed matching vector; wherein the formula is:

[0044]

[0045] Among them, V1 represents the welding current-wire feeding speed matching vector, V2 represents the welding voltage-wire feeding speed matching vector, represents the square of the vector norm, v 1i represents the eigenvalue of the ith position of the welding current-wire speed matching vector, v 2i represents the eigenvalue of the ith position of the welding voltage-wire speed matching vector, exp represents the natural exponential function, V1′ represents the pre-alignment welding current-wire speed matching vector, V2′ represents the pre-alignment welding voltage-wire speed matching vector, PCA(·) represents the essential information extraction, U1 represents the sequence of the first eigendecomposition pre-alignment welding current-wire speed matching vector, Λ1 represents the first diagonal matrix, λ 11 and λ 1mdenote the eigenvalues ​​of the first and mth positions of the first diagonal matrix, respectively. T denotes the transpose of the vector. concat{·;·} denotes vector concatenation. v 11 、v 12 、v 1m denote the first, second and mth eigenvectors of the sequence of the first eigendecomposition pre-aligned welding current-wire speed matching vector, V k represents the first principal component pre-alignment welding current-wire feeding speed matching vector, U2 represents the sequence of the second eigendecomposition pre-alignment welding voltage-wire feeding speed matching vector, Λ2 represents the second diagonal matrix, λ 21 and λ 2m denote the eigenvalues ​​of the first and mth positions of the second diagonal matrix, respectively. concat{·;·} denotes vector concatenation. v 21 、v 22 、v 2m denote the first, second and mth eigenvectors of the sequence of the second eigendecomposition pre-aligned welding voltage-wire speed matching vector, V s represents the second principal component pre-alignment welding voltage-wire feeding speed matching vector, represents positional subtraction, ⊙ represents positional multiplication, Indicates positional addition, Conv 1×1 represents point convolution, f vp (V k ,V s ) represents the welding process parameter-wire feeding speed matching vector.

[0046] That is, in the technical solution of the present application, the welding current-wire speed matching vector and the welding voltage-wire speed matching vector are subjected to intrinsic structure-guided feature matching to obtain the welding process parameter-wire speed matching vector, which first performs modulus normalization feature coordination on the welding current-wire speed matching vector and the welding voltage-wire speed matching vector to obtain the pre-aligned welding current-wire speed matching vector and the pre-aligned welding voltage-wire speed matching vector. This process is not just for simply adjusting the scale, but an important technology in feature engineering, which aims to solve the problem of inconsistent scales between features in the data set. By calculating the modulus of the feature vector and dividing each vector by its own modulus, it can be ensured that all feature vectors have the same length, thereby eliminating the influence of scale differences between different feature vectors. Normalization processing not only contributes to numerical stability in subsequent processing, but also makes the comparison between feature vectors more fair and reasonable, laying the foundation for subsequent feature alignment.

[0047] Next, the pre-aligned welding current-wire speed matching vector and the pre-aligned welding voltage-wire speed matching vector are subjected to essential information extraction to obtain a first principal component pre-aligned welding current-wire speed matching vector and a second principal component pre-aligned welding voltage-wire speed matching vector. That is, after completing the normalized modulation, the pre-aligned welding current-wire speed matching vector and the welding voltage-wire speed matching vector are subjected to essential information extraction. Essential information extraction, as an unsupervised linear dimensionality reduction method, is widely used in data compression and feature extraction. This process determines the principal components by finding the direction of the maximum variance of the data. These principal components are new coordinate axes in the original feature space, which can capture the structure of the data with minimal information loss. It is worth mentioning that by selecting the eigenvectors corresponding to the first few largest eigenvalues, the dimension of the feature space can be effectively reduced while maintaining the key information of the data. This dimensionality reduction can not only accelerate the model training process, but also help the model to generalize better, reduce the impact of noise, and thus improve the performance of the model.

[0048] Finally, the first principal component pre-aligned welding current-wire feeding speed matching vector and the second principal component pre-aligned welding voltage-wire feeding speed matching vector are accurately matched in feature details to obtain the welding process parameter-wire feeding speed matching vector. The goal of this step is to find the best matching point or area between the two principal component feature vectors to achieve alignment between the two. Fine-grained alignment goes beyond the traditional rough fusion methods such as simple weighted average or splicing, and needs to consider the time series characteristics or spatial distribution characteristics of the feature vector. The fusion after alignment is not a simple numerical operation, but requires the design of a specific mechanism to determine how to combine the information of the two feature vectors. This content-based intelligent fusion method can more effectively utilize multi-source information and improve the model's characterization ability and decision-making quality.

[0049] In an embodiment of the present application, the welding process parameter-wire feeding speed matching feature encoding module 140 is used to feature encode the welding process parameter-wire feeding speed matching vector to obtain the welding process parameter-wire feeding speed matching feature vector. Specifically, in an embodiment of the present application, the welding process parameter-wire feeding speed matching feature encoding module is used to: pass the welding process parameter-wire feeding speed matching vector through a multi-scale matching feature encoder to obtain the welding process parameter-wire feeding speed matching feature vector. In the present application, the multi-scale matching feature encoder is a convolutional neural network model including a first convolution layer and a second convolution layer, wherein the first convolution layer uses a one-dimensional convolution kernel of a first scale, and the second convolution layer uses a one-dimensional convolution kernel of a second scale, wherein the first scale is different from the second scale. It should be understood that the welding process parameter-wire feeding speed matching vector contains rich welding process information, but this information may be distributed in different time scales. In order to capture the multi-level information in the welding process, the welding process parameter-wire feeding speed matching vector needs to be input into the multi-scale matching feature encoder. The multi-scale matching feature encoder uses different one-dimensional convolution kernels to capture features within different time scales. The one-dimensional convolution kernel of the first scale is better at capturing the welding parameter wire feeding speed matching features within a short time range, while the one-dimensional convolution kernel of the second scale is better at capturing the welding parameter wire feeding speed matching features within a long time range. Through multi-scale convolution, the encoder can generate richer feature representations. These feature representations not only contain the basic information of the original vector, but also contain high-level features processed at different scales, thereby enhancing the expressiveness and generalization capabilities of the original time series matching data.

[0050] In the embodiment of the present application, the welding process parameter adjustment result generation module 150 is used to obtain the adjustment direction of the actual welding current and the actual welding voltage of the spot welder at the current time based on the welding process parameter-wire feeding speed matching feature vector. Specifically, Figure 3 FIG. 1 is a block diagram of a welding process parameter adjustment result generation module in an intelligent control system of a digital pulse consumable electrode welding machine according to an embodiment of the present application. Figure 3As shown, the welding process parameter adjustment result generation module 150 includes: a welding machine actual welding current adjustment result generation unit 151, which is used to pass the welding process parameter-wire feeding speed matching feature vector through a welding current adjustment classifier to obtain a welding current adjustment classification result, and the welding current adjustment classification result is used to indicate whether the actual welding current of the welding machine at the current time point should be increased, decreased, or remain unchanged; a welding machine actual welding voltage adjustment result generation unit 152, which is used to pass the welding process parameter-wire feeding speed matching feature vector through a welding voltage adjustment classifier to obtain a welding voltage adjustment classification result, and the welding voltage adjustment classification result is used to indicate whether the actual welding current of the welding machine at the current time point should be increased, decreased, or remain unchanged.

[0051] Finally, the welding process parameter-wire feeding speed matching feature vector is passed through the welding current adjustment classifier to obtain a welding current adjustment classification result for indicating that the actual welding current of the welder at the current time point should be increased, decreased or kept unchanged. At the same time, the welding process parameter-wire feeding speed matching feature vector is passed through the welding voltage adjustment classifier to obtain a welding voltage adjustment classification result for indicating that the actual welding current of the welder at the current time point should be increased, decreased or kept unchanged. That is, based on the matching feature information in the welding process parameter-wire feeding speed matching feature vector, the welding current adjustment classifier and the welding voltage adjustment classifier are automatically output to adjust the direction of the welding current and welding voltage of the current welder. The welding current adjustment classifier and the welding voltage adjustment classifier are trained machine learning models, which are trained by labeled training data to learn the mapping relationship between the original data and the output adjustment result. When a new welding process parameter-wire feeding speed matching feature vector is input, the corresponding classification result is output according to the learned mapping mode. In this way, accurate control of the welding current and voltage of the welder can be achieved, so as to cope with the complex control requirements in the arc initiation stage, thereby improving the stability and quality of welding.

[0052] In summary, the digital pulse consumable welding machine intelligent control system 100 based on the embodiment of the present application is explained, which first associates the wire feeding speed value at multiple time points with the actual welding current value of the welding machine to obtain the welding current-wire feeding speed matching vector, and at the same time associates the wire feeding speed value at multiple time points with the actual welding voltage value of the welding machine to obtain the welding voltage-wire feeding speed matching vector, and then performs multi-scale analysis on the welding process parameter-wire feeding speed matching vector obtained by fusion of the welding current-wire feeding speed matching vector and the welding voltage-wire feeding speed matching vector, and finally outputs the welding machine current and voltage adjustment direction that need to be controlled through the classifier. In this way, accurate control of the welding current and voltage can be achieved, so as to cope with the complex control requirements of the arc initiation stage, thereby improving the stability and quality of welding.

[0053] As described above, the digital pulse melting electrode welding machine intelligent control system 100 according to the embodiment of the present application can be implemented in various terminal devices, such as a server for intelligent control of the digital pulse melting electrode welding machine. In one example, the digital pulse melting electrode welding machine intelligent control system 100 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 digital pulse melting electrode welding machine intelligent control system 100 can be a software module in the operating system of the terminal device, or it can be an application developed for the terminal device; of course, the digital pulse melting electrode welding machine intelligent control system 100 can also be one of the many hardware modules of the terminal device.

[0054] Alternatively, in another example, the digital pulse consumable welding machine intelligent control system 100 and the terminal device may also be separate devices, and the digital pulse consumable welding machine intelligent control system 100 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.

[0055] Figure 4 FIG. 1 is a flow chart of an intelligent control method for a digital pulse consumable electrode welding machine according to an embodiment of the present application. Figure 4 As shown, in the intelligent control method of the digital pulse consumable electrode welding machine, the following steps are included: S110, obtaining the wire feeding speed value, the actual welding current value of the welding machine, and the actual welding voltage value of the welding machine at multiple predetermined time points within a predetermined time period; S120, performing data regularization on the wire feeding speed values, the actual welding current value of the welding machine, and the actual welding voltage value of the welding machine at the multiple predetermined time points to obtain a wire feeding speed timing input vector, a welding current timing input vector, and a welding voltage timing input vector; S130, performing vector matching analysis on the wire feeding speed timing input vector, the welding current timing input vector, and the welding voltage timing input vector to obtain a welding process parameter-wire feeding speed matching vector; S140, performing feature encoding on the welding process parameter-wire feeding speed matching vector to obtain a welding process parameter-wire feeding speed matching feature vector; S150, based on the welding process parameter-wire feeding speed matching feature vector, obtaining the adjustment direction of the actual welding current of the welding machine and the actual welding voltage of the welding machine at the current time point.

[0056] Here, those skilled in the art can understand that the specific operations of each step in the above-mentioned digital pulse melting electrode welding machine intelligent control method have been described in the above reference. Figures 1 to 3 The description of the digital pulse melting electrode welding machine intelligent control system has been introduced in detail, and therefore, its repeated description will be omitted.

[0057] In summary, the intelligent control method of the digital pulse consumable electrode welding machine based on the embodiment of the present application is explained, which first associates the wire feeding speed values ​​at multiple time points with the actual welding current value of the welding machine to obtain the welding current-wire feeding speed matching vector, and at the same time associates the wire feeding speed values ​​at multiple time points with the actual welding voltage value of the welding machine to obtain the welding voltage-wire feeding speed matching vector, and then performs a multi-scale analysis on the welding process parameter-wire feeding speed matching vector obtained by merging the welding current-wire feeding speed matching vector and the welding voltage-wire feeding speed matching vector, and finally outputs the welding machine current and voltage adjustment direction that need to be controlled through the classifier. In this way, accurate control of the welding current and voltage can be achieved, so as to cope with the complex control requirements of the arc initiation stage, thereby improving the stability and quality of welding.

Claims

1. An intelligent control system for a digital pulse consumable welding machine, characterized in that: include: The welding process related parameter data acquisition module is used to obtain the wire feeding speed value, the actual welding current value of the welding machine, and the actual welding voltage value of the welding machine at multiple predetermined time points within a predetermined time period; A welding process related parameter structuring module, used for performing data regularization on the wire feeding speed values, actual welding current values ​​of the welding machine, and actual welding voltage values ​​of the welding machine at the plurality of predetermined time points to obtain a wire feeding speed timing input vector, a welding current timing input vector of the welding machine, and a welding voltage timing input vector of the welding machine; A welding process parameter-wire feeding speed matching module is used to perform vector matching analysis on the wire feeding speed timing input vector, the welding current timing input vector of the welding machine and the welding voltage timing input vector of the welding machine to obtain a welding process parameter-wire feeding speed matching vector; A welding process parameter-wire feeding speed matching feature encoding module, used for performing feature encoding on the welding process parameter-wire feeding speed matching vector to obtain a welding process parameter-wire feeding speed matching feature vector; The welding process parameter adjustment result generation module is used to obtain the adjustment direction of the actual welding current and the actual welding voltage of the spot welder at the current time based on the welding process parameter-wire feeding speed matching feature vector.

2. The intelligent control system of the digital pulse consumable electrode welding machine according to claim 1 is characterized in that: The welding process related parameter structuring module is used to: arrange the wire feeding speed values, the actual welding current values ​​of the welding machine, and the actual welding voltage values ​​of the welding machine at the multiple predetermined time points according to the time dimension to obtain the wire feeding speed timing input vector, the welding current timing input vector of the welding machine, and the welding voltage timing input vector of the welding machine.

3. The intelligent control system of the digital pulse consumable electrode welding machine according to claim 2 is characterized in that: The welding process parameter-wire feeding speed matching module includes: A wire feeding speed timing feature extraction unit, used for passing the wire feeding speed timing input vector through a wire feeding speed timing encoder to obtain a wire feeding speed timing feature vector; A welding current timing feature extraction unit for a welding machine, used for passing the welding current timing input vector of the welding machine through a welding current timing encoder to obtain a welding current timing feature vector of the welding machine; A welding voltage timing feature extraction unit for a welding machine, used for passing the welding voltage timing input vector of the welding machine through a welding voltage timing encoder to obtain a welding voltage timing feature vector of the welding machine; A welding current-wire feeding speed matching unit, used for multiplying the wire feeding speed timing characteristic vector and the welding current timing characteristic vector of the welding machine by position points to obtain a welding current-wire feeding speed matching vector; A welding current-wire feeding speed matching unit, used for multiplying the wire feeding speed timing characteristic vector and the welding voltage timing characteristic vector of the welding machine by position points to obtain a welding voltage-wire feeding speed matching vector; The welding process parameter-wire feeding speed matching fusion generation unit is used to perform intrinsic structure-guided feature matching on the welding current-wire feeding speed matching vector and the welding voltage-wire feeding speed matching vector to obtain the welding process parameter-wire feeding speed matching vector.

4. The intelligent control system of the digital pulse consumable electrode welding machine according to claim 3 is characterized in that: The welding process parameter-wire feeding speed matching fusion generation unit comprises: A vector modulus normalization feature coordination subunit, used for performing modulus normalization feature coordination on the welding current-wire feeding speed matching vector and the welding voltage-wire feeding speed matching vector to obtain a pre-aligned welding current-wire feeding speed matching vector and a pre-aligned welding voltage-wire feeding speed matching vector; A vector essential information extraction subunit is used to extract the essential information of the pre-aligned welding current-wire feeding speed matching vector and the pre-aligned welding voltage-wire feeding speed matching vector to obtain a first principal component pre-aligned welding current-wire feeding speed matching vector and a second principal component pre-aligned welding voltage-wire feeding speed matching vector; The vector feature detail precise matching subunit is used to perform feature detail precise matching on the first principal component pre-aligned welding current-wire feeding speed matching vector and the second principal component pre-aligned welding voltage-wire feeding speed matching vector to obtain the welding process parameter-wire feeding speed matching vector.

5. The intelligent control system of the digital pulse consumable electrode welding machine according to claim 4 is characterized in that: The vector modulus normalization feature coordination subunit is used to: Calculating the square of the second norm of the welding current-wire feeding speed matching vector to obtain a welding current-wire feeding speed characteristic intensity value; Calculating the natural exponential function value for each characteristic value of the welding current-wire feeding speed matching vector to obtain an exponential welding current-wire feeding speed matching vector; Dividing each characteristic value of the exponential welding current-wire feeding speed matching vector by the sum of the exponential welding current-wire feeding speed matching vectors to obtain a normalized welding current-wire feeding speed matching vector; Each characteristic value of the normalized welding current-wire feeding speed matching vector is multiplied by the welding current-wire feeding speed characteristic intensity value to obtain the pre-alignment welding current-wire feeding speed matching vector.

6. The intelligent control system of the digital pulse consumable electrode welding machine according to claim 5 is characterized in that: The vector essential information extraction subunit is used to: Performing eigenvalue-based decomposition on the pre-alignment welding current-wire feeding speed matching vector to obtain a sequence of first eigen-decomposed pre-alignment welding current-wire feeding speed matching vectors and a first diagonal matrix; The first eigenvector and the second eigenvector in the sequence of the first eigendecomposition pre-aligned welding current-wire feeding speed matching vector are cascaded to obtain the first principal component pre-aligned welding current-wire feeding speed matching vector.

7. The intelligent control system of the digital pulse consumable electrode welding machine according to claim 6 is characterized in that: The welding process parameter-wire feeding speed matching feature encoding module is used to: pass the welding process parameter-wire feeding speed matching vector through a multi-scale matching feature encoder to obtain the welding process parameter-wire feeding speed matching feature vector.

8. The intelligent control system of the digital pulse consumable electrode welding machine according to claim 7 is characterized in that: The wire feeding speed timing encoder, the welding current timing encoder and the welding voltage timing encoder are convolutional neural network models including a fully connected layer and a one-dimensional convolutional layer, and the multi-scale matching feature encoder is a convolutional neural network model including a first convolutional layer and a second convolutional layer, wherein the first convolutional layer uses a one-dimensional convolutional kernel of a first scale, and the second convolutional layer uses a one-dimensional convolutional kernel of a second scale, wherein the first scale is different from the second scale.

9. The intelligent control system of the digital pulse consumable electrode welding machine according to claim 8 is characterized in that: The welding process parameter adjustment result generation module includes: A welding machine actual welding current adjustment result generating unit, used for passing the welding process parameter-wire feeding speed matching feature vector through a welding current adjustment classifier to obtain a welding current adjustment classification result, wherein the welding current adjustment classification result is used to indicate whether the actual welding current of the welding machine at the current time point should be increased, decreased or kept unchanged; The actual welding voltage adjustment result generating unit of the welding machine is used to pass the welding process parameter-wire feeding speed matching feature vector through the welding voltage adjustment classifier to obtain the welding voltage adjustment classification result, and the welding voltage adjustment classification result is used to indicate whether the actual welding voltage of the welding machine at the current time point should increase, decrease or remain unchanged.

10. An intelligent control method for a digital pulse consumable electrode welding machine, characterized in that: include: Obtain wire feeding speed values, actual welding current values, and actual welding voltage values ​​of the welding machine at multiple predetermined time points within a predetermined time period; Performing data regularization on the wire feeding speed values, actual welding current values ​​of the welding machine, and actual welding voltage values ​​of the welding machine at the plurality of predetermined time points to obtain a wire feeding speed timing input vector, a welding current timing input vector, and a welding voltage timing input vector of the welding machine; Performing vector matching analysis on the wire feeding speed timing input vector, the welding current timing input vector of the welding machine, and the welding voltage timing input vector of the welding machine to obtain a welding process parameter-wire feeding speed matching vector; Characteristic encoding is performed on the welding process parameter-wire feeding speed matching vector to obtain a welding process parameter-wire feeding speed matching characteristic vector; Based on the welding process parameter-wire feeding speed matching feature vector, the adjustment direction of the actual welding current of the spot welder and the actual welding voltage of the welder at the current time is obtained.

Images