An adaptive welding control method, system, electronic device, and medium

Abstract

The application relates to the technical field of oil and gas pipeline welding, and discloses a self-adaptive welding control method, a system, an electronic device and a medium, the method comprising the following steps: acquiring first visual images and first welding parameters; establishing a mapping relationship between each first visual image and each first welding parameter; inputting each first visual image and the corresponding first welding parameter into a dense convolutional neural network as a data set for training, so as to determine a welding parameter prediction model; acquiring a second visual image of the surface of an oil and gas pipeline in real time during welding, inputting the second visual image into the welding parameter prediction model, and predicting each second welding parameter; and controlling a welding robot to perform pipeline welding according to each second welding parameter. The welding parameter prediction model can be used to predict corresponding welding parameters from visual images in real time, so that the welding parameters of the welding robot can be adjusted adaptively, and the stability of the welding process and the weld quality are guaranteed.

Description

Technical Field

[0001] This invention relates to the field of oil and gas pipeline welding technology, and in particular to an adaptive welding control method, system, electronic device and medium. Background Technology

[0002] Existing welding machines for oil and gas pipelines have some basic adjustable parameters, such as welding current, welding speed, and arc voltage, which can be directly set according to welding needs to achieve better welding results. However, in actual use, due to the limited number of adjustable motion and process parameters and poor real-time adjustment, synchronous joint adjustment cannot be achieved, which limits the flexibility and applicability of the equipment. For example, existing technology describes an automatic welding control method that generates welding parameters corresponding to the welding required for welding based on the weld morphology of the weld layer and weld bead, predicts the weld morphology based on the obtained welding parameters, compares the predicted weld morphology with the planned morphology, and adjusts the welding parameters based on the comparison results. The above-mentioned existing technology adjusts the welding parameters of the welding robot by predicting the weld morphology, and cannot adaptively control the welding robot according to the real-time weld morphology.

[0003] For example, existing technology describes a weld pool center recognition system and weld seam tracking method based on HDR images. By acquiring the deviation between the weld pool center and the weld seam, and correcting the position of the K-TIG welding torch in real time based on the deviation, weld seam tracking is achieved, which can be used for automatic tracking operations in K-TIG welding. However, the above technical solution is not suitable for welding oil and gas pipelines, and it also cannot adaptively control the welding robot.

[0004] In summary, existing pipe welding technologies cannot adaptively adjust the welding parameters of the welding robot according to the morphological state of the pipe during welding. Summary of the Invention

[0005] To overcome the problem that existing pipe welding technologies cannot adaptively adjust the welding robot parameters according to the morphology of the pipe during welding, this invention provides an adaptive welding control method, system, electronic device, and medium.

[0006] In a first aspect, to solve the above-mentioned technical problems, the present invention provides an adaptive welding control method, comprising:

[0007] Acquire the first visual image of the oil and gas pipeline surface and the first welding parameters of the welding robot during welding;

[0008] Establish a mapping relationship between each first visual image and each first welding parameter, and determine each first welding parameter corresponding to each first visual image;

[0009] Each first visual image and each first welding parameter with a mapping relationship are used as a dataset and input into a dense convolutional neural network for training to build a welding parameter prediction model.

[0010] The second visual image of the oil and gas pipeline surface is acquired in real time during welding, and the second visual image is input into the welding parameter prediction model to predict the various second welding parameters of the welding robot corresponding to the second visual image.

[0011] The welding robot is controlled to perform pipe welding based on various secondary welding parameters.

[0012] In a second aspect, the present invention provides an adaptive welding control system, comprising:

[0013] The first data acquisition module is used to acquire the first visual image of the oil and gas pipeline surface during welding and the first welding parameters of the welding robot.

[0014] The mapping relationship establishment module is used to establish the mapping relationship between each first visual image and each first welding parameter, and to determine each first welding parameter corresponding to each first visual image.

[0015] The model training module is used to take each first visual image and each first welding parameter that establishes the mapping relationship as a dataset and input them into a dense convolutional neural network for training to build a welding parameter prediction model.

[0016] The second data acquisition module is used to acquire a second visual image of the surface of the oil and gas pipeline in real time during welding, and input the second visual image into the welding parameter prediction model to predict the various second welding parameters of the welding robot corresponding to the second visual image.

[0017] The welding robot control module is used to control the welding robot to perform pipe welding according to various second welding parameters.

[0018] Thirdly, the present invention provides a computing device, including a memory, a processor, and a program stored in the memory and running on the processor, wherein the processor executes the program to implement the steps of the adaptive welding control method described above.

[0019] Fourthly, the present invention provides a computer-readable storage medium storing instructions that, when executed on a terminal device, cause the terminal device to perform the steps of the adaptive welding control method described above.

[0020] The beneficial effects of this invention are as follows: A dense convolutional neural network is trained using a first visual image of the oil and gas pipeline during welding and the corresponding first welding parameters to obtain a welding parameter prediction model. A second visual image of the oil and gas pipeline during welding, acquired in real time, is then input into the welding parameter prediction model to obtain the corresponding second welding parameters. These second welding parameters are then used to control the welding robot in real time. This application uses a welding parameter prediction model to predict the corresponding welding parameters from the visual image in real time, adaptively adjusting various welding parameters of the welding robot to ensure the stability of the welding process and the quality of the weld. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0022] Figure 1 This is a flowchart illustrating an adaptive welding control method according to an embodiment of the present invention.

[0023] Figure 2 This is a schematic diagram of the structure of the channel attention mechanism module and the spatial attention mechanism module of the present invention;

[0024] Figure 3 This is a schematic diagram of the structure of an adaptive welding control system according to an embodiment of the present invention. Detailed Implementation

[0025] The following embodiments are further explanations and supplements to the present invention and do not constitute any limitation on the present invention.

[0026] The following describes an adaptive welding control method, system, electronic device, and medium according to embodiments of the present invention, with reference to the accompanying drawings.

[0027] like Figure 1 As shown, the present invention provides an adaptive welding control method, comprising:

[0028] S1. Obtain the first visual image of the oil and gas pipeline surface during welding and the first welding parameters of the welding robot.

[0029] S2. Establish the mapping relationship between each first visual image and each first welding parameter, and determine each first welding parameter corresponding to each first visual image.

[0030] S3. Take the first visual images and the first welding parameters that establish the mapping relationship as a dataset and input them into a dense convolutional neural network for training to build a welding parameter prediction model.

[0031] S4. Acquire a second visual image of the oil and gas pipeline surface during welding in real time, and input the second visual image into the welding parameter prediction model to predict the various second welding parameters of the welding robot corresponding to the second visual image.

[0032] S5. Control the welding robot to perform pipe welding according to each of the second welding parameters.

[0033] In this embodiment, a dense convolutional neural network is trained using a first visual image of the oil and gas pipeline during welding and the corresponding first welding parameters to obtain a welding parameter prediction model. A second visual image of the oil and gas pipeline during welding, acquired in real time, is then input into the welding parameter prediction model to obtain the corresponding second welding parameters. These second welding parameters are then used to control the welding robot in real time. This application uses the welding parameter prediction model to predict the corresponding welding parameters from the visual image in real time, adaptively adjusting various welding parameters of the welding robot to ensure the stability of the welding process and the quality of the weld.

[0034] Optionally, a mapping relationship is established between each first visual image and each first welding parameter, and each first welding parameter corresponding to each first visual image is determined, including:

[0035] When a first visual image is acquired, the first welding parameters of the welding robot corresponding to the first visual image are obtained, and the first welding parameters are marked as marking information in the first visual image.

[0036] In this embodiment, the different first-view images are all associated with specific welding parameters of the welding robot, for example:

[0037] The first visual image of the first frame is an image of the weld surface morphology;

[0038] The marking information corresponding to the first visual image of the first frame is: welding current: 180A, welding voltage: 24V, welding speed: 12cm / min, welding torch oscillation frequency: 3Hz, and welding torch attitude angle: 20°.

[0039] The first visual image of the Nth frame is the molten pool topography image;

[0040] The marking information corresponding to the first visual image of the Nth frame is: welding current: 175A, welding voltage: 22V, welding speed: 11.5cm / min, welding torch oscillation frequency: 2Hz, and welding torch attitude angle: 10°.

[0041] Optionally, the welding parameter prediction model is a prediction model formed by sequentially adding a channel attention mechanism module and a spatial attention mechanism module between the convolutional layer and the fully connected layer of a dense convolutional neural network.

[0042] The second visual image is input into the welding parameter prediction model to predict various second welding parameters of the welding robot corresponding to the second visual image, including:

[0043] Multiple convolution operations are performed on the second visual image through convolutional layers to extract local features of the second visual image and determine multiple feature maps.

[0044] Based on the global information of each feature map, the target feature map is selected from each feature map through the channel attention mechanism module;

[0045] The spatial attention mechanism module is used to filter out target regions from the target feature map.

[0046] Features of the target region are extracted by using fully connected layers in a dense convolutional neural network to determine high-level abstract global features;

[0047] The normalized exponential function in a dense convolutional neural network maps each second welding parameter to a high-level abstract global feature map.

[0048] This embodiment constructs a DenseNet dense convolutional neural network based on a mixed attention module. The original DenseNet structure consists of a sequentially connected input layer, convolutional layer, fully connected layer, and softmax layer (normalization function). In this embodiment, a mixed attention module is added between the convolutional and fully connected layers. This mixed attention module comprises a channel attention module (CA module) and a spatial attention module (SA module) to capture important channel and spatial features of the weld and molten pool morphology. The softmax layer is used to predict welding parameters.

[0049] The core feature of the dense convolutional neural network based on the hybrid attention mechanism used in this embodiment is that the output of each layer is connected to all subsequent layers, thereby achieving better feature reuse and gradient propagation. This structure enables the model to capture local and global information in visual images.

[0050] The role of hybrid attention mechanisms (including spatial attention mechanisms and channel attention mechanisms) is to enhance the network's ability to express features, enabling the model to focus more accurately on important information in the welding process, thereby improving the precision of welding process control.

[0051] In this embodiment, the data needs to be preprocessed before inputting the second visual image into the welding parameter prediction model. This includes steps such as denoising and normalizing the acquired second visual image to ensure that the data format and range meet the input requirements of the dense convolutional neural network. At the same time, such preprocessing also helps to improve the generalization ability of the neural network.

[0052] Optionally, based on the global information of each feature map, the target feature map is selected from each feature map through the channel attention mechanism module, including:

[0053] Each feature map is input into a different channel of the channel attention mechanism module;

[0054] The first weight value is calculated based on the global information of each feature map and the preset weight of the corresponding channel; whereby the first weight value represents the importance of the global information in the feature map.

[0055] The feature map corresponding to the maximum value among the first weight values ​​is taken as the target feature map.

[0056] In this embodiment, the channel attention mechanism module assigns weights to each channel to highlight key feature channels and suppress unimportant or redundant feature channels. Then, based on the global information of each feature map (such as weld morphology, molten pool state, etc.), it determines which feature map channels are more important. Thus, the feature map corresponding to the most important global information can be selected from each channel.

[0057] like Figure 2 As shown, each channel in the Channel Attention (CA) module includes two branches, where:

[0058] The first branch includes Maxpool, FC (Fully Connected Layer), ReLU (Rectified Linear Unit activation function), and FC (Fully Connected Layer). It is mainly used to reduce the spatial dimension of the feature map by selecting the maximum value within the pooling window to obtain global information. Then, the global features are identified and classified through two fully connected layers. After identification and classification, the first weight value is calculated through the activation function.

[0059] The second branch includes Avgpool (average pooling layer), FC (Fully Connected Layer), ReLU (Rectified Linear Unit activation function), and FC (Fully Connected Layer). It is mainly used to reduce the spatial dimension of the feature map by averaging all values ​​within the pooling window to obtain global information. Then, the global features are identified and classified through two fully connected layers. After identification and classification, the first weight value is calculated through the activation function.

[0060] The first weight values ​​calculated from the first branch and the second branch are merged, and the result is mapped to a range of 0 to 1 using the Sigmoid function to obtain the final first weight value.

[0061] Optionally, the target region is selected from the target feature map through the spatial attention mechanism module, including:

[0062] The second weight value of each pixel in the target feature map is calculated using a spatial attention mechanism; where the second weight value represents the importance of the spatial position in the target feature map.

[0063] Pixels whose second weight value is greater than or equal to a preset value are used as target pixels;

[0064] The region corresponding to each target pixel is taken as the target region.

[0065] In this embodiment, the spatial attention mechanism module assigns weights to each spatial location (pixel) in the feature map, highlighting features at key locations and suppressing features in irrelevant or background areas. The spatial attention mechanism module focuses on extracting important regions in the image, such as weld edges, weld start and end points, and molten pool boundaries. It ensures that the model can focus on these key spatial locations during the welding process, rather than being distracted by irrelevant background information.

[0066] like Figure 2 As shown, the Spatial Attention (SA) module includes Maxpool, Avgpool, Concatenate, Conv, and the Sigmoid function.

[0067] The spatial dimension of the feature map is reduced by averaging all values ​​within the pooling window, and then by maximizing all values ​​within the pooling window. The feature maps reduced by Maxpool and Avgpool are then fused through the Concatenate layer. The features at each spatial location are extracted by Conv and the second weight value is calculated. Finally, the second weight value is mapped to a value between 0 and 1 using the Sigmoid function.

[0068] Optionally, the first visual image and the second visual image include a weld surface morphology image and a molten pool morphology image, and the first welding parameters and the second welding parameters include welding current, welding voltage, welding speed, welding torch oscillation frequency, and welding torch attitude angle.

[0069] In this embodiment, the weld surface morphology image can be obtained by a line profile scanner, the molten pool morphology data can be obtained by a molten pool camera, and the first welding parameters can be collected by a multi-source sensor.

[0070] Optionally, the welding robot is controlled to perform pipe welding according to various second welding parameters, including:

[0071] Welding current, welding voltage, and welding speed are input into the welding power control system of the welding robot, and welding torch oscillation frequency and welding torch attitude angle are input into the motion control system of the welding robot.

[0072] In this embodiment, the welding parameters of the welding robot can be adaptively and synchronously adjusted to achieve intelligent decision-making in welding control, thereby helping to improve welding quality and production efficiency.

[0073] Furthermore, an adaptive welding control method based on deep learning is employed, extracting the parameters as numerical process or motion parameters that the robot needs to execute. A dense convolutional neural network for adaptive welding is constructed, and the welding parameters are adaptively adjusted based on closed-loop feedback parameters to ensure the stability and consistency of welding quality. Simultaneously, the adaptive adjustment of welding parameters can be used to automatically compensate for defects in the upper surface quality, thereby effectively ensuring the welding pass rate and weld quality.

[0074] like Figure 3 As shown, the present invention provides an adaptive welding control system, comprising:

[0075] The first data acquisition module is used to acquire the first visual image of the oil and gas pipeline surface during welding and the first welding parameters of the welding robot.

[0076] The mapping relationship establishment module is used to establish the mapping relationship between each first visual image and each first welding parameter, and to determine each first welding parameter corresponding to each first visual image.

[0077] The model training module is used to take each first visual image and each first welding parameter that establishes the mapping relationship as a dataset and input them into a dense convolutional neural network for training to build a welding parameter prediction model.

[0078] The second data acquisition module is used to acquire a second visual image of the surface of the oil and gas pipeline in real time during welding, and input the second visual image into the welding parameter prediction model to predict the various second welding parameters of the welding robot corresponding to the second visual image.

[0079] The welding robot control module is used to control the welding robot to perform pipe welding according to various second welding parameters.

[0080] Optionally, the mapping relationship establishment module is specifically used for:

[0081] When a first visual image is acquired, the first welding parameters of the welding robot corresponding to the first visual image are obtained, and the first welding parameters are marked as marking information in the first visual image.

[0082] Optionally, the dense convolutional neural network includes convolutional layers, a channel attention mechanism module, a spatial attention mechanism module, and a softmax layer;

[0083] The second data acquisition module is specifically used for:

[0084] Multiple convolution operations are performed on the second visual image through convolutional layers to extract local features of the second visual image and determine multiple feature maps.

[0085] Based on the global information of each feature map, the target feature map is selected from each feature map through the channel attention mechanism module;

[0086] The spatial attention mechanism module is used to filter out target regions from the target feature map.

[0087] Features of the target region are extracted by using fully connected layers in a dense convolutional neural network to determine high-level abstract global features;

[0088] The normalized exponential function in a dense convolutional neural network maps each second welding parameter to a high-level abstract global feature map.

[0089] Optionally, the second data acquisition module is specifically used for:

[0090] Each feature map is input into a different channel of the channel attention mechanism module;

[0091] The first weight value is calculated based on the global information of each feature map and the preset weight of the corresponding channel; whereby the first weight value represents the importance of the global information in the feature map.

[0092] The feature map corresponding to the maximum value among the first weight values ​​is taken as the target feature map.

[0093] Optionally, the second data acquisition module is specifically used for:

[0094] The second weight value of each pixel in the target feature map is calculated using a spatial attention mechanism; where the second weight value represents the importance of the spatial position in the target feature map.

[0095] Pixels whose second weight value is greater than or equal to a preset value are used as target pixels;

[0096] The region corresponding to each target pixel is taken as the target region.

[0097] Optionally, the welding robot control module is specifically used for:

[0098] Welding current, welding voltage, and welding speed are input into the welding power control system of the welding robot, and welding torch oscillation frequency and welding torch attitude angle are input into the motion control system of the welding robot.

[0099] The present invention also provides a computing device, including a memory, a manager, and a program stored in the memory and running on the manager, wherein the manager executes the program to implement some or all of the steps of the above-described adaptive welding control method.

[0100] The computing device can be a computer, and the corresponding program is computer software. The parameters and steps of the computing device of the present invention can be referred to the parameters and steps in the embodiment of the adaptive welding control method above, and will not be repeated here.

[0101] Those skilled in the art will recognize that this invention can be implemented as a system, method, or computer program product. Therefore, this disclosure can be embodied in the following forms: it can be entirely hardware, entirely software (including firmware, resident software, microcode, etc.), or a combination of hardware and software, generally referred to herein as a "circuit," "module," or "system." Furthermore, in some embodiments, the invention can also be implemented as a computer program product contained in one or more computer-readable media, which contains computer-readable program code. Computer-readable storage media can be, for example, but not limited to—electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof.

[0102] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0103] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. An adaptive welding control method, characterized in that, include: Acquire the first visual image of the oil and gas pipeline surface and the first welding parameters of the welding robot during welding; Establish a mapping relationship between each first visual image and each first welding parameter, and determine each first welding parameter corresponding to each first visual image; The first visual images and the first welding parameters that establish the mapping relationship are used as a dataset to train and construct a welding parameter prediction model. A second visual image of the oil and gas pipeline surface is acquired in real time during welding, and the second visual image is input into the welding parameter prediction model to predict the various second welding parameters of the welding robot corresponding to the second visual image. The welding robot is controlled to perform pipe welding according to each of the second welding parameters; The step of establishing the mapping relationship between each of the first visual images and each of the first welding parameters, and determining each of the first welding parameters corresponding to each of the first visual images, includes: When a frame of a first visual image is acquired, the first welding parameters of the welding robot corresponding to the first visual image are obtained, and the first welding parameters are marked as marking information in the first visual image. The welding parameter prediction model is a prediction model formed by sequentially adding a channel attention mechanism module and a spatial attention mechanism module between the convolutional layer and the fully connected layer of a dense convolutional neural network. The step of inputting the second visual image into the welding parameter prediction model to predict various second welding parameters of the welding robot corresponding to the second visual image includes: The second visual image is subjected to multiple convolution operations through the convolutional layer to extract local features of the second visual image and determine multiple feature maps. Based on the global information of each feature map, the target feature map is selected from each feature map through the channel attention mechanism module; The spatial attention mechanism module is used to filter out target regions from the target feature map. The fully connected layer extracts features from the target region to determine high-level abstract global features; The normalized exponential function in the dense convolutional neural network is used to map each of the second welding parameters to the high-level abstract global feature map. The step of selecting target feature maps from each feature map based on global information of each feature map, through the channel attention mechanism module, includes: Each of the feature maps is input into a different channel of the channel attention mechanism module; A first weight value is calculated based on the global information of each feature map and the preset weight of the corresponding channel; wherein, the first weight value represents the importance of the global information in the feature map; The feature map corresponding to the maximum value among the first weight values ​​is taken as the target feature map; The step of filtering target regions from the target feature map using the spatial attention mechanism module includes: A second weight value for each pixel in the target feature map is calculated using a spatial attention mechanism; wherein, the second weight value characterizes the importance of the spatial location in the target feature map; Each pixel whose second weight value is greater than or equal to a preset value is taken as the target pixel; The region corresponding to each of the target pixels is taken as the target region.

2. The method according to claim 1, characterized in that, The first visual image and the second visual image include a weld surface morphology image and a molten pool morphology image. The first welding parameters and the second welding parameters include welding current, welding voltage, welding speed, welding torch oscillation frequency, and welding torch attitude angle.

3. The method according to claim 2, characterized in that, The step of controlling the welding robot to perform pipe welding according to each of the second welding parameters includes: The welding current, welding voltage, and welding speed are input into the welding power control system of the welding robot, and the welding torch oscillation frequency and welding torch attitude angle are input into the motion control system of the welding robot.

4. An adaptive welding control system, characterized in that, An adaptive welding control method according to any one of claims 1-3 includes: The first data acquisition module is used to acquire the first visual image of the oil and gas pipeline surface during welding and the first welding parameters of the welding robot. The mapping relationship establishment module is used to establish a mapping relationship between each first visual image and each first welding parameter, and to determine each first welding parameter corresponding to each first visual image. The model training module is used to take each of the first visual images and each of the first welding parameters that establish the mapping relationship as a dataset and input them into a dense convolutional neural network for training to determine the welding parameter prediction model. The second data acquisition module is used to acquire a second visual image of the surface of the oil and gas pipeline in real time during welding, and input the second visual image into the welding parameter prediction model to predict the various second welding parameters of the welding robot corresponding to the second visual image. The welding robot control module is used to control the welding robot to perform pipe welding according to the second welding parameters.

5. A computing device, comprising a memory, a processor, and a program stored in the memory and running on the processor, characterized in that, When the processor executes the program, it implements the steps of an adaptive welding control method as described in any one of claims 1-3.

6. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores instructions that, when executed on a terminal device, cause the terminal device to perform the steps of an adaptive welding control method as described in any one of claims 1-3.

Images