Search

Copyright of Metal Working (1674-165X) is the property of Metal Working Editorial Office and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Wire feeding speed has an important effect on the welding quality. In order to realize robot intelligent welding, it is necessary to study the accurate detection method of the wire feeding speed. First, the working principle of the wire feeding speed detection is studied, thus the wire feeding speed online detection can be realized. Then, a kind of wire feeding speed detection system is designed, which wirelessly transmits the sensing signal of the welding wire to the welding robot. Finally, the detection method of the wire feeding speed based on the filtering algorithm of distortion sensing signal is studied, including the principle of no mutation of adjacent wire feeding speed sensing signal, the interference signal elimination algorithm for adjacent detection signal of multiple sensing signal loss without abrupt change, and the detection method of the wire feeding speed. The experimental results show that the main noise in the original wire feeding speed sensing signal can be eliminated by using the designed algorithm and system, and the accuracy of the wire feeding speed detection can be improved. In addition, the width of weld pass after robot welding can not be affected by the change of the welding current.

With the wide application of multi-layer and multi-pass welding in industry, the traditional manual welding method is difficult to meet the needs of manufacture. Welding Robot has the advantages of stable productivity, ensuring welding quality even in special environment, so the welding robots are used at a growing trend in manufacturing fields to complete different welding tasks. In this paper, an intelligence learning method for welding robot is designed, aiming at the prediction of welding process parameters and bead geometry parameters in the welding process, deep and machine learning algorithms are used for realization. It provides an instruction for the design of process parameters to realize the intellectualization and automation of welding robot. The experimental results show that automatic parameters learning based on machine learning are effective and different learning methods should be selected for different process parameter prediction tasks in order to achieve the best prediction effect.

Robotic welding requires a higher weld image resolution for easy weld identification; however, the higher the resolution, the higher the cost. Therefore, this paper proposes an improved CLAHE algorithm, which can not only effectively denoise and retain edge information but also improve the contrast of images. First, an improved bilateral filtering algorithm is used to process high-resolution images to remove noise while preserving edge details. Then, the CLAHE (Contrast Limited Adaptive Histogram Equalization) algorithm and Gaussian masking algorithm are used to enhance the enhanced image, and then differential processing is used to reduce the noise in the two images, while preserving the details of the image, enhancing the image contrast, and obtaining the final enhanced image. Finally, the effectiveness of the algorithm is verified by comparing the peak signal-to-noise ratio and structural similarity with other algorithms. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Shanghai Jiao Tong University (1006-2467) is the property of Journal of Shanghai Jiao Tong University Editorial Office and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

The image noise generated by the welding process, such as arc light, splash, and smoke, brings significant challenges for the laser vision sensor-based welding robot to locate the weld seam and accurately conduct automatic welding. Currently, deep learning-based approaches surpass traditional methods in flexibility and robustness. However, their significant computational cost leads to a mismatch with the real-time requirement of automated welding. In this paper, we propose an efficient hybrid architecture of Convolutional Neural Network (CNN) and transformer, referred to as Dynamic Squeeze Network (DSNet), for real-time weld seam segmentation. More precisely, a lightweight segmentation framework is developed to fully leverage the advantages of the transformer structure without significantly increasing computational overhead. In this respect, an efficient encoder, which aims to increase its features diversity, has been designed and resulted in substantial improvement of encoding performance. Moreover, we propose a plug-and-play lightweight attention module that generates more effective attention weights by exploiting statistical information of weld seam data and introducing linear priors. Extensive experiments on weld seam images using NVIDIA GTX 1050Ti show that our approach reduces the number of parameters by 54x, decreases computational complexity by 34x, and improves inference speed by 33x compared to the baseline method TransUNet. DSNet achieves superior accuracy (78.01% IoU, 87.64% Dice) and speed performance (100 FPS) with lower model complexity and computational burden than most state-of-the-art methods. The code is available at https://github.com/hackerschen/DSNet. [ABSTRACT FROM AUTHOR]

Seam tracking with structured light vision has been widely applied into the robot welding. And the precise laser stripe extraction is the premise of automatic laser vision seam tracking. However, conventional laser stripe extraction methods based on image processing have the shortcomings of poor flexibility and robustness, which are easily affected by considerable image noises in the welding processing, such as arc light, smoke, and splash. To address this issue, inspired by image segmentation, with the strong contextual feature expression ability of deep convolutional neural network (DCNN), a novel image denoising method of seam images is proposed in this paper for automatic laser stripe extraction to serve intelligent robot welding applications, such as seam tracking, seam type detection, weld bead detection, etc. With the deep encoder-decoder network framework, aimed at the information loss issue by multiple convolutional and pooling operations in DCNNs, an attention dense convolutional block is proposed to extract and accumulate multi-scale feature maps. Meanwhile, a residual bi-directional ConvLSTM block (BiConvLSTM) is proposed to effectively learn multi-scale and long-range spatial contexts from local feature maps. Finally, a weighted loss function is proposed for model training to address the class unbalanced issue. Combined with the seam image set, the experimental results show that the proposed image denoising network could correctly extract the laser stripes from seam images which could demonstrate that the proposed method shows a high detection precision and good robustness against the strong image noise interference from welding process. [ABSTRACT FROM AUTHOR]

The extremely rapid development of the welding technology has forced the automation - and then the robotization of manufacturing processes - from the 70s to the 80s. In the beginning one have chosen those applications with simpler and more precise parts. Therefore, the robots (nonadaptive at that time) were able to weld good enough the structures. Afterwards, the components required to be welded with a robot became increasingly complex and imprecise. It became necessary to equip the robotic systems with adaptive functions, first with the geometric adaptivity only, then with process adaptivity too. At present, robots with an expanded adaptability are becoming more and more applied to common parts, whose preparation for welding costs less obviously. The paper presents the structure and functioning of the adaptive robotic systems for electric arc welding, especially the sensory device and their adaptive regulators, as well as the specific developments of the welding equipment necessary for them to be successfully integrated into adaptive cells and systems. After introducing systems implemented in major European factories, solutions that work in the Romanian energy industry are described.

Seam-tracking technology based on vision sensing is the key technology to realize robot intelligent welding. Because of changes in the weld gap, the diversity of workpiece material and surface conditions, and the interference of welding noises, it is extremely difficult for the robot to realize real-time seam-tracking in gas metal arc welding (GMAW) of large structures. So, a method of position detection and weld gap feature extraction is proposed for GMAW seam-tracking system of fillet weld with variable gap. This method established a highly robust weld feature extraction model based on a low-cost structured-light vision sensor. In this model, using the reflection characteristic of the stripes on fillet weld, a feature extraction method for variable gap weld based on column gray difference operator was proposed. To overcome the interference of welding noises, this model adopted the AND logic operation method between adjacent sampled images. To overcome the interference of uneven and strong reflection conditions, an optimized random sampling consistency (RANSAC) algorithm was proposed and used for weld feature extraction. The optimized RANSAC algorithm used the linear slope of the stripes to optimize the random sampling process of the algorithm. The experimental results showed that the average detection errors of the proposed method in the Y-axis and Z-axis directions are 0.19 mm and 0.23 mm, respectively, at a welding speed of 1000 mm/min. The proposed method could realize seam tracking of the variable gap fillet weld and still had good robustness under the uneven and strong reflection of the workpiece surfaces. [ABSTRACT FROM AUTHOR]

Start point guiding and seam tracking are important parts of the robot’s intelligent welding process. However, some start point guiding methods have a large amount of calculation and low accuracy. In addition, due to the large curvature of the curved weld, the problem of visual advance in seam tracking should be solved. Therefore, an efficient and accurate start point guiding and seam tracking method for curve weld is proposed in this article. First, according to the characteristics of structural light of different welds, more efficient and accurate start point detection algorithms are proposed for curve v-groove welds and curve lap welds. The extraction of key characteristic points and the judgment of the relationship between them could achieve efficient weld start point detection, and the reasonable robot speed planning during the start point searching process could improve the accuracy of the start point guiding. Besides, a sliding data queue method with cubic B-spline fitting is proposed to overcome the visual advance problem and achieve accurate seam tracking of curve welds. The sliding data queue method can plan the data queue automatically, and improve the computational efficiency by reducing the amount of data to be processed. Cubic B-spline fitting can reduce the measurement error and eliminate the data fluctuation, so as to obtain a smooth welding trajectory. The experimental results show that the proposed method could achieve efficient and accurate start point guiding and seam tracking of curve v-groove weld and lap weld. [ABSTRACT FROM AUTHOR]

Multi-blade wheel structures are essential components in many industrial and mechanical systems. Substituting manual welding with robot welding can significantly enhance processing efficiency for them. However, robotic welding requires complex teaching and programming to plan paths. This is not conducive to the improvement of welding efficiency. To address this issue, this paper establishes a welding robot system based on dual 3D cameras and presents a 3D vision-driven robotic path planning approach. In this method, the eye-to-hand camera and the eye-in-hand camera are used as the shooting path planning guide and the welding path planning guide, respectively. First, this paper uses a DBSCAN-based method to segment the point cloud and plan the shooting pose according to the geometric information of the blade. Second, to extract the weld accurately, this paper proposes a concave-convex feature descriptor called flatness, and feature points are extracted using a K-means-based clustering method. The weld path is obtained by fitting the feature points. Finally, to ensure welding quality stability, a welding torch posture planning method based on the neighborhood centroid is proposed, following the principle of the dihedral angle. The comparison experiment verifies that the proposed method can realize the path planning of the welding multi-blade wheel structures without teaching and programming with high efficiency and precision. • Developed a welding robot system utilizing dual DLP cameras to enable precise welding path planning. • Proposed a novel descriptor for point cloud analysis, delineating concave-convex features and constructing feature spaces. • Implemented efficient clustering techniques for weld seam feature extraction. • Implemented welding torch posture planning based on the principle of dihedral angles. [ABSTRACT FROM AUTHOR]

This paper proposes an efficient approach for extracting feature points from weld images in noisy construction environments. Inspired by the human pose estimation, the proposed method reformulates the weld feature point extraction as a skeletal keypoint detection task. A quick object detector locates the weld region amidst complex backgrounds, followed by efficient feature point extraction via two coordinate classification tasks. This approach achieves sub-pixel accuracy at a low computational cost and confines the annotation within one bounding box and four keypoints per image, eliminating pixel-level labeling. Test results demonstrate real-time, accurate feature point extraction with superior efficiency and robustness compared to traditional methods. The proposed approach thus facilitates the quality control for automated welding in real-world construction scenarios. • A lightweight deep learning-based method is developed for weld seam feature extraction. • A novel procedure of weld seam feature extraction inspired by human pose estimation is presented. • Weld feature point coordinate location is transformed to horizontal and vertical coordinate classification. • The proposed method exhibits better performance compared to the mainstream methods in terms of accuracy and efficiency. [ABSTRACT FROM AUTHOR]

The extremely rapid development of the welding technology has forced the automation - and then the robotization of manufacturing processes - from the 70s to the 80s. In the beginning one have chosen those applications with simpler and more precise parts. Therefore, the robots (nonadaptive at that time) were able to weld good enough the structures. Afterwards, the components required to be welded with a robot became increasingly complex and imprecise. It became necessary to equip the robotic systems with adaptive functions, first with the geometric adaptivity only, then with process adaptivity too. At present, robots with an expanded adaptability are becoming more and more applied to common parts, whose preparation for welding costs less obviously. The paper presents the structure and functioning of the adaptive robotic systems for electric arc welding, especially the sensory device and their adaptive regulators, as well as the specific developments of the welding equipment necessary for them to be successfully integrated into adaptive cells and systems. After introducing systems implemented in major European factories, solutions that work in the Romanian energy industry are described. [ABSTRACT FROM AUTHOR]

Recently, laser vision sensors are widely applied in initial point alignment and seam tracking to improve the level of intelligent welding because of good characteristics. However, since the deformation of laser stripe is unobvious at the narrow weld with 0.2 mm width, these methods are not applicable for the narrow weld. Moreover, there are rare researches that could achieve initial point alignment and seam tracking of narrow weld simultaneously. Therefore, an initial point alignment and seam tracking system for narrow weld is proposed in this paper. At first, a laser vision sensor with extra light emitting diode light is used to obtain laser and weld seam image. Besides, the seam feature point is extracted and three-dimensional coordinates can be obtained with vision model. In addition, three controllers including decision controller, initial point alignment controller, and seam-tracking controller are proposed to achieve initial point alignment and seam tracking control in X- and Z-axis directions. Moreover, feature verification, Kalman filter, and output pulse verification are designed to improve the accuracy and stability of this system. Finally, many initial point alignment and seam-tracking experiments of narrow weld are conducted. Experimental results demonstrate that proposed system can well achieve initial point alignment and seam tracking of planar and curved surface narrow weld. [ABSTRACT FROM AUTHOR]

Our paper mainly introduces a novel path planning method for non-ideal sphere-pipe intersecting curve robot welding based on laser scanning. This method integrates the generation of laser scanning trajectory, the processing of scanning data, and the path planning of non-ideal sphere-pipe joint welding. First, the paper perfects the ideal sphere-pipe intersection model and represents the parametric equation of ideal intersecting curve, which can cover all the intersection ways for sphere-pipe joints. Since the spheres and pipes applied in actual production are not standard, this paper adopts the scheme of scanning and identifying weld seam using the laser displacement sensor and gives the laser sensor scanning trajectory by analyzing the direction and attitude of space welds. In this paper, by sampling and filtering the distance data obtained from laser sensor, a novel weld point identification algorithm suitable for the above scanning trajectory is proposed. In response to the constantly changing of sphere-pipe joints' weld inclination and attitude, this paper adopts the robot-positioner welding scheme and introduces a novel algorithm for solving the position of two external axes. The ADAMS simulation experiments prove that this scheme can effectively avoid the adverse effects of the uphill and downhill welding on the welding quality. [ABSTRACT FROM AUTHOR]

Sphere-pipe joints welding is widely used in industrial applications. This paper presents a robot welding approach for the sphere-pipe joints with swing and multi-layer planning. Firstly, various coordinate systems are used to describe the geometric relationship between weld seam and robot welding torch. The sphere-pipe intersecting curve welding process is basically uphill and downhill welding. Therefore, this paper establishes a description model of the welding torch attitude, which parameterizes the attitude description and automatically adjusts the torch attitude during the welding process according to the change of weld inclination angle. To overcome the negative effects of gravity, such as deepening of the molten pool and reduction of the weld width, this paper integrates the swing welding technology into trajectory planning and gives a solving algorithm for the welding torch swing curve. The swing welding also can reduce the number of weld pass. Therefore, multi-layer single-pass swing welding is an economical and efficient way for thicker weldments. In this paper, a multi-layer single-pass swing welding planning algorithm is proposed, which can automatically determine the height and swing amplitude of each welding layer. Finally, the industrial robot Puma560 is used to carry out experimental simulation, and the simulation results are used to verify the feasibility and accuracy of this approach. [ABSTRACT FROM AUTHOR]

Initial weld point guiding is the premise of intelligentized welding process. At present, many initial weld point guiding methods based on structured light vision sensors and passive vision sensors have been presented. However, the methods based on structured light vision sensors are not suitable for micro-gap weld and the methods based on passive vision sensors have large guiding errors, which limits their application for micro-gap weld. Therefore, a precise initial weld point guiding method of micro-gap weld is proposed in this paper. This initial weld point guiding method consists of two parts. In the vision sensing part, a structured light vision sensor with a narrow-band optical filter and a LED light is designed to acquire clear image, including laser stripe and micro-gap weld seam. Then, the image processing method is designed to detect feature point and the 3-D coordinates of the feature point can be obtained based on an established vision model. In the control part, combining the 3-D coordinates of the feature point with the initial weld point guiding model, the 3-D coordinates of initial weld point can be determined. Then, the initial weld point guiding controller is adopted to make welding torch align with the initial weld point. Finally, many initial weld point guiding experiments are carried out. Experimental results show that the proposed method can achieve the initial weld point guiding of planar and curved surface micro-gap weld accurately, and the guiding accuracy is improved compared with previous methods, which can be applied in welding production. [ABSTRACT FROM AUTHOR]

• A light-weight object detection network is constructed for the extraction of weld feature points in automatic welding. Compared with the mainstream deep learning-based object detection algorithms, our method greatly improves the image processing speed of welding while retaining high accuracy and stability. • The experimental results show that the maximum mean error on different types of welding seams is 0.1940 mm, and the image processing speed reaches about 49 FPS on the CPU, meeting the accuracy and real-time requirements of weld seam tracking. • Besides, we hope that the strategy can provide novel solutions for extending the advantages of light-weight object detection method to more industrial scenarios. In the seam tracking process based on laser vision, the camera continuously collects weld seam images and locates weld feature points by utilizing image processing algorithms. It is critical to extract the feature points of the weld seam accurately and in real-time from noise interference. Deep learning-based weld seam recognition methods have high accuracy and strong robustness, but it is hard to satisfy the real-time requirements when they are deployed on devices with low computing power. To solve this problem, a lightweight object detection network based on Single Shot MultiBox Detector is proposed. Then the proposed method is in comparison with the mainstream deep learning-based algorithms, and the welding experiments are executed. The experimental results show that the average localization error is within ± 0.2 mm, and the image processing speed reaches about 49 FPS on the CPU, demonstrating that the proposed network could satisfy the requirements of accuracy and real-time. [ABSTRACT FROM AUTHOR]