ArticlePublisher preview available

Narrow gap deviation detection in Keyhole TIG welding using image processing method based on Mask-RCNN model

Authors:
Yunke Chen
Yunke Chen
Yunke Chen
  • This person is not on ResearchGate, or hasn't claimed this research yet.
Yonghua Shi at South China University of Technology
Yonghua Shi
Yanxin Cui at South China University of Technology
Yanxin Cui
Xiyin Chen at South China University of Technology
Xiyin Chen
Read publisher preview
Download citation
To read the full-text of this research, you can request a copy directly from the authors.

Abstract

In the process of K-TIG deep penetration welding, the workpiece does not need to be bevelled; therefore, the welding method is butt welding, and the gap to be welded is very narrow (0.2~1 mm). Because of the large welding current, the welding arc light intensity is very strong. These factors cause difficulties in the K-TIG weld seam tracking process. To realize weld seam tracking in the K-TIG welding process, it is necessary to extract the keyhole entrance centre and weld centreline accurately. To reduce the interference of strong arc light in the process of K-TIG welding, the information of keyhole entrance and weld seam in the process of K-TIG welding is obtained by using a high-dynamic-range camera. An image processing algorithm based on Mask-RCNN is proposed to extract the centre of the keyhole entrance accurately. An image processing algorithm based on Hough line fitting is used to accurately identify the weld centreline in the welding image and extract the welding deviation. In welding experiments, it is verified that the welding deviation extracted by the method proposed in this paper fluctuates within ± 0.133 mm, which meets the requirements of actual K-TIG welding seam tracking.
A preview of this full-text is provided by Springer Nature.
Learn more
Content available from The International Journal of Advanced Manufacturing Technology
This content is subject to copyright. Terms and conditions apply.
ORIGINAL ARTICLE
Narrow gap deviation detection in Keyhole TIG welding using image
processing method based on Mask-RCNN model
Yunke Chen
1,2
&Yonghua Shi
1,2
&Yanxin Cui
1,2
&Xiyin Chen
1,2
Received: 1 July 2020 / Accepted: 7 December 2020
#The Author(s), under exclusive licence to Springer-Verlag London Ltd. part of Springer Nature 2021
Abstract
In the process of K-TIG deep penetration welding, the workpiece does not need to be bevelled; therefore, the welding method is
butt welding, and the gap to be welded is very narrow (0.2~1 mm). Because of the large welding current, the welding arc light
intensity is very strong. These factors cause difficulties in the K-TIG weld seam tracking process. To realize weld seam tracking
in the K-TIG welding process, it is necessary to extract the keyhole entrance centre and weld centreline accurately. To reduce the
interference of strong arc light in the process of K-TIG welding, the information of keyhole entrance and weld seam in the process
of K-TIG welding is obtained by using a high-dynamic-range camera. An image processing algorithm based on Mask-RCNN is
proposed to extract the centre of the keyhole entrance accurately. An image processing algorithm based on Hough line fitting is
used to accurately identify the weld centreline in the welding image and extract the welding deviation. In welding experiments, it
is verified that the welding deviation extracted by the method proposed in this paper fluctuates within ± 0.133 mm, which meets
the requirements of actual K-TIG welding seam tracking.
Keywords K-TIG narrow gap welding .Image processing .Mask-RCNN .Seam tracking
1 Introduction
The seam tracking ability of a welding system is significant
for the welding process and for obtaining good welds [1]. In
the field of weld seam tracking, many researchers have studied
the process of weld seam tracking based on visual sensors.
The machine vision system is still the main method to realize
robot welding automation [2]. There are two methods of seam
tracking based on vision sensors: active vision and passive
vision. In the scheme of active vision, laser or other light
sources are used to irradiate the workpiece to obtain the char-
acteristic information of the weld seam. Gong et al. [3]useda
structured-light vision sensing system to yield the profile of
the deep grooves of the joint and proposed an effective
algorithm to recognize the laser stripe. Fan et al. [4]presented
a seam tracking initial point alignment method based on a
laser vision system. Xu et al. [5] used a circular laser three-
dimensional scanner to construct a seam tracking system to
realize seam tracking. Kiddee et al. [6] presented a weld seam
tracking system using cross mark structured light. However,
in order to avoid the interference of very high–intensity arc
light, the detection area of active vision is usually in front of
the actual welding area. The thermal deformation and time lag
during the welding process and other factors have a greater
impact on the accuracy of the weld deviation detected by these
methods. In the passive vision scheme, a CCD camera is typ-
ically used to obtain the welding arc and weld seam informa-
tion directly, and process this information to obtain the rele-
vant features. Guo et al. [7] analysed the imaging characteris-
tics during the metal active gas (MAG) welding process and
proposed a method for determining the V-groove centreline.
Du et al. [8] proposed an algorithm to accurately identify the
feature of atypical weld seam in robotic gas metal arc welding
(GMAW). Shen et al. [9] presented a welding robot system to
measure the offset of the torch to the seam centre and the size
of the seam groove by passive vision. Wang et al. [10]pro-
posed a method based on template-matching to detect the
groove centre during GMAW. These methods are applicable
*Yonghua Shi
yhuashi@scut.edu.cn
1
School of Mechanical and Automotive Engineering, South China
University of Technology, Guangzhou 510640, China
2
Guangdong Provincial Engineering Research Center for Special
Welding Technology and Equipment, South China University of
Technology, Guangzhou 510640, China
https://doi.org/10.1007/s00170-020-06466-5
/ Published online: 6 January 2021
The International Journal of Advanced Manufacturing Technology (2021) 112:2015–2025
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
... Weld tracking image processing is the process of processing the image captured by the vision sensor, suppressing the non-essential information, improving the image quality, and extracting the weld feature information [11]. Literature [12] through the laser sensor to measure the initial height while using a voltage converter to detect the tungsten electrode and tube plate voltage, compared with the experimental analysis of the more optimal solution, the study shows that the calculation of the pipe center error of 0.2mm, and TIG robots to make the center of the welded pipe and the center of the welding torch is very accurate. ...
... The cosine of the angle between vector ( ) 11 , A x y and vector ( ) 22 , B x y in two dimensions: (15) In this paper, the similarity metric is adopted to solve the problem of weld tracking, by calculating the similarity between the template region and each position of the search region, the response map is obtained, and the peak position in the response map indicates that the region is the most similar to the template region, and the weld tracking results are obtained. ...
Research Progress of Weld Tracking Image Processing Technology Based on Deep Learning Theory
Article
Full-text available
  • Dec 2023
In this paper, a convolutional neural network is used to localize the weld seam feature points with noise interference in complex welding environments. A priori frames are introduced into the feature point extraction network, combined with position prediction and confidence prediction, to improve the accuracy and anti-interference ability of the weld tracking system. To improve welding efficiency by utilizing the continuity of weld tracking, the weld tracking network is designed based on the twin structure. The weld detection network designs the first frame to locate the key position of the bevel and inputs into the weld tracking network as a template, and the weld tracking network completes the automatic tracking of the subsequent welds. At the same time, the network introduces a hybrid domain attention mechanism, which makes full use of the weld feature channel dependence and spatial location relationship and puts more attention near the inflection point of the weld laser line to ensure the accuracy of weld tracking. The research results show that the extraction error of weld seam feature points based on the convolutional neural network is within 17, which is much lower than that of the grayscale center of gravity method and Steger's algorithm. In the weld tracking experiments under the workpiece tilting state, the average value of the absolute error of the tracking trajectory in the X-axis direction is not more than 0.7 mm, and the maximum value is less than 1.15 mm. The absolute tracking error in the Z-axis direction is relatively low, with an average of 0.638 mm and a maximum of 1.573 mm. Therefore, the weld-tracking image processing technique proposed in this paper has strong anti-noise interference capabilities and high localization accuracy. And high accuracy in localization.
View
... Unlike the active vision method, no extra light sources are needed in the passive vision method, only arc light and natural light are used in the welding process. Chen et al. [96] preferred to utilize the deep learning-based Mask RCNN model, which is seen in Fig. 14, for the detection of the keyhole entrance point due to the high amount of arc light generated in Keyhole TIG (K-TIG) welding via passive vision, which is used for joining narrow-gap workpieces. After the keyhole entrance point is determined, processes such as Gaussian filtering, Laplacian edge detection, and Hough line fitting are used for centerline detection to obtain the center point and center points in the next time slot. ...
... In addition, robotic welding applications in which vision sensors are used are divided into active and passive. The studies carried out under the subheadings [96] Content courtesy of Springer Nature, terms of use apply. Rights reserved. ...
Recent developments in computer vision and artificial intelligence aided intelligent robotic welding applications
Article
Full-text available
  • May 2023
  • INT J ADV MANUF TECH
The welding process, which is an indispensable part of the manufacturing industry, has been in demand for years and continues to attract the attention of researchers. With the transition to Industry 4.0, the welding process got out of the control of the operators and became automated with sensors and artificial intelligence methods, and as a result, it became inevitable for industrial manipulators or robots to enter the production sector. One of the most important details in making the welding process autonomous in manufacturing is the sensors, and among the sensors are the vision sensors. In recent years, it is seen that robotic welding applications are applied very sensitively and successfully when visual sense and artificial intelligence are used together. This study comprehensively reviewed research and development for cutting-edge applications using visual sensors and artificial intelligence for robotic welding applications. The processes that are the subject of intelligent robotic welding applications such as calibration, determination of welding starting point, seam tracking, and welding quality are determined and discussed based on current studies and critical analyzes. The detection, tracking, diagnosis, classification, and prediction performances of various methods of machine learning (ML), which is one of the most used areas in artificial intelligence-based applications, in welding applications are examined comparatively. This review article will help researchers about what should be considered in vision sensor aided robotic welding applications and how to contribute more to studies with artificial intelligence support.
View
... et al. (2020) for particle characterization,Pustokhina et al. (2021) for anomaly detection in pedestrian walkways,Jin et al (2021) for detection of highway guardrail,Sadhukhan et al. (2020) to estimate surface temperature from thermal imagery of building,Tran et al. (2022) for detecting and identifying cracks in pavements,Shaodan et al. (2019) for ship detection,Al-Shaibani et al. (2021) for airplane type identification,Sizkouhi et al. (2020) for boundary extraction of photovoltaic plants,Chen et al. (2021c) for narrow gap deviation detection in welding,Fan et al. (2021) for detection and segmentation of underwater objects,Rahman et al. (2022) for filler detection in Scanning Electron Microscopic images,Raoofi and Motamedi (2020) for detection of construction machinery,Htet and Sein (2021) for classification of palm trees,Prados-Privado et al. (2021) for radiographic detection of teeth,Johnson (2020) for the detection of cell nuclei in the medical images,An et al (2021) for the automatic diagnosis of tongue. ...
Prediction of Star Polygon Types in Islamic Geometric Patterns with Deep Learning
Article
  • Jun 2024
Historical buildings in the Eastern world of architecture host many Islamic geometric patterns which are known as mathematically sophisticated patterns regarding their period of creation. This study focuses on the preparation of a model that can be helpful for the analysis and restoration/maintenance of these patterns. For this, a deep learning model to detect and classify star types in Islamic geometric patterns has been proposed, and the trials were evaluated. Accordingly, this study presents a database containing 5-pointed, 6-pointed, 8-pointed and 12-pointed star types. The database consists of 600 Islamic geometric patterns. A mask RCNN algorithm was trained to detect and classify star types using the prepared database. The results of the training indicate that the loss value is 0.90 and the validation loss value is 0.85. The algorithm was tested using images that it had not seen before and the results were evaluated. This paper presents a discussion on the pros and cons of the trained algorithm.
View
... In recent years, as machine learning and deep learning are more and more widely used in the field of computer vision, various excellent deep learning algorithms have emerged, and these methods are gradually applied to the field of seam tracking. For example, Chen [21] et al. used a high dynamic range CMOS camera to acquire images and identified the keyhole entrance using the Mask-RCNN model in their study of Keyhole TIG seam tracking which extracted the coordinates of the melt pool centroid and obtained the weld deviation. Zou [22] et al. combined convolutional filters and deep reinforcement learning to locate weld feature points for seam tracking. ...
A real-time seam tracking system based on extraction and restoration model in robotic GMAW
Article
Full-text available
  • Jan 2024
  • INT J ADV MANUF TECH
Seam tracking technology is an important part of the intelligent welding field. In this paper, a laser vision-based real-time seam tracking system was built. The system consists of a self-developed laser vision sensor, a six-axis robot, a gas metal arc welding system, and an industrial computer. After building the system, the system calibration was performed. During the seam tracking, the arc light, spatter, and other welding noise have a negative impact on the image processing algorithm to extract the weld feature points, and even lead to system drift and algorithm failure. To this end, a two-stage extraction and restoration model (ERM) was proposed for processing real-time welding images to improve the robustness and accuracy of the seam tracking system. In the ERM, the region of interest was first detected and extracted by the YOLOv5s model, then the extracted images were restored by the conditional generation adversarial network. After using the ERM model, a series of image processing was performed to obtain the coordinates of the weld feature points. The total time consumed by the algorithm is 37 ms per frame on average, which meets the real-time requirement. Moreover, the experimental results show that the seam tracking system based on the ERM can achieve real-time tracking for different types of planar V-bevel welds, and the average error is 0.21 mm, which meets the requirements for actual welding.
View
... The residual network (ResNet) was developed for welding state classification based on the molten pool image, achieving an impressive recognition rate of 98% in experiments. To guarantee a good welding seam tracking effect, Chen et al. [23] proposed an image processing algorithm using Mask-RCNN to accurately identify the keyhole entrance and weld centerline. The accuracy of the algorithm was verified by experiments, and the inference speed was 1.53 frames per second. ...
Real-Time Recognition of Molten Pools Based on Improved DeepLabV3+ in Keyhole Tungsten Inert Gas Welding Applications
Article
Full-text available
  • Jan 2024
During the Keyhole Tungsten Inert Gas (K-TIG) welding process, a significant amount of information related to the weld quality can be obtained from the weld pool and the keyhole of the topside molten pool image, which provides a vital basis for the control of welding quality. However, the topside molten pool image has the unstable characteristic of strong arc light, which leads to difficulty in contour extraction. The existing image segmentation algorithms cannot satisfy the requirements for accuracy, timing, and robustness. Aiming at these problems, a real-time recognition method, based on improved DeepLabV3+, for identifying the molten pool more accurately and effectively was proposed in this paper. First, MobileNetV2 was selected as the feature extraction network with which to improve detection efficiency. Then, the atrous rates of atrous convolution layers were optimized to reduce the receptive field and balance the sensitivity of the model to molten pools of different scales. Finally, the convolutional block attention module (CBAM) was introduced to improve the segmentation accuracy of the model. The experimental results verified that the proposed model had a fast segmentation speed and higher segmentation accuracy, with an average intersection ratio of 89.89% and an inference speed of 103 frames per second. Furthermore, the trained model was deployed in a real-time system and achieved a real-time performance of up to 28 frames per second, thus meeting the real-time and accuracy requirements of the K-TIG molten pool monitoring system.
View
... Applications of Mask R-CNN in the manufacturing industry are vast and include automated machine inspection when combined with augmented reality, where accuracies of 70% and 100% were achieved depending on which machines were inspected [24], automated defect detection during powder spreading in selective laser melting with 92.7% accuracy and 0.22 s per image [25], solder joint recognition with over 0.95 mAP [26], identification and tracking of objects in manufacturing plants in near-real time leading to automatic object misplacement identification where a precision (number of correct predictions compared to number of overall predictions) of 0.99 and a recall (number of correct predictions compared to number of ground truth instances) of 0.98 were achieved [27], classification and localisation of semiconductor wafer map defect patterns with 97.7% accuracy [28], detection and segmentation of aircraft cable brackets with an AP of 0.998, recall of 99.5% and mean intersection-over-union (mIoU, an average of the IoU metric explained in Sect. 3) of 84.5% with a time of 1.02 s per bracket compared to the traditional method which took ten seconds [29], surface defect detection of automotive engine parts with an mAP of 0.85 component assembly inspection with a classification accuracy of 86.6% [30], welding deviation detection in keyhole TIG deep penetration welding with a satisfactory outcome of ± 0.133 mm and variance of 0.0056mm 2 [31], automated pointer meter reading with an AP of 0.71, automated defect detection of industrial filter cloth with an accuracy of 87.3% [32] and finally, wind turbine blade defect detection and classification with an AP, using a 0.5 IoU threshold, of 82.6% [33]. ...
Ladle pouring process parameter and quality estimation using Mask R-CNN and contrast-limited adaptive histogram equalisation
Article
Full-text available
  • Mar 2023
  • INT J ADV MANUF TECH
Deep learning in computer vision is becoming increasingly popular and useful for tracking object movement in many application areas, due to data collection burgeoning from the rise of the Internet of Things (IoT) and Big Data. So far, computer vision has been used in industry predominantly for quality inspection purposes such as surface defect detection; however, an emergent research area is the application for process monitoring involving tracking moving machinery in real time. In steelmaking, the deployment of computer vision for process monitoring is hindered by harsh environments, poor lighting conditions and fume presence. Therefore, application of computer vision remains unplumbed. This paper proposes a novel method for tracking hot metal ladles during pouring in poor lighting. The proposed method uses contrast-limited adaptive histogram equalisation (CLAHE) for contrast enhancement, Mask R-CNN for segmentation prediction and Kalman filters for improving predictions. Pixel-level tracking enables pouring height and rotation angle estimation which are controllable parameters. Flame severity is also estimated to indicate process quality. The method has been validated with real data collected from ladle pours. Currently, no publications presenting a method for tracking ladle pours exist. The model achieved a mean average precision (mAP) of 0.61 by the Microsoft Common Objects in Context (MSCOCO) standard. It measures key process parameters and process quality in processes with high variability, which significantly contributes to process enhancement through root-cause analysis, process optimisation and predictive maintenance. With real-time tracking, predictions could automate ladle controls for closed-loop control to minimise emissions and eliminate variability from human error.
View
Monitoring the forming dimensions of components produced by arc-directed energy deposition based on a molten pool’s geometric characteristics
Article
  • Dec 2023
A deposition size monitoring method based on a molten pool’s geometric characteristics, including image calibration, contour extraction, contour coordinate calculation, and deposition size modeling, is established in this study to observe the real-time size of components formed by arc-directed energy deposition (ADED). Image calibration was first performed using the molten pool’s imaging transformation matrix to accurately match the image’s pixels to the actual spatial position. Next, an adaptive threshold segmentation algorithm was applied to extract the contours of the calibrated molten pool’s images. The contour was further completed using breakpoint coordinates and their gradient information. The minimum bounding rectangle algorithm was used to fit the contour coordinates for determining the width and height of a single-pass deposition. A mathematical multi-pass and multi-layer depositional size model was created via the parabolic model and the equal volume algorithm for realizing the deposition size monitoring based on the molten pool images. Finally, the rocket engine shell was manufactured using ADED, and the depositional forming size was observed in real time. The overall average size deviation of the component was within ±2.41 mm, suggesting a high forming accuracy.
View
View
View
View
Image splicing detection using mask-RCNN Image splicing detection using mask-RCNN
Article
Full-text available
  • Jul 2020
Digital images have become a dominant source of information and means of communication in our society. However, they can easily be altered using readily available image editing tools. In this paper, we propose a new blind image forgery detection technique which employs a new backbone architecture for deep learning which is called ResNet-conv. ResNet-conv is obtained by replacing the feature pyramid network in ResNet-FPN with a set of convolutional layers. This new backbone is used to generate the initial feature map which is then to train the Mask-RCNN to generate masks for spliced regions in forged images. The proposed network is specifically designed to learn discriminative artifacts from tampered regions. Two different ResNet architectures are considered, namely ResNet-50 and ResNet-101. The ImageNet, He_normal, and Xavier_normal initialization techniques are employed and compared based on convergence. To train a robust model for this architecture, several post-processing techniques are applied to the input images. The proposed network is trained and evaluated using a computer-generated image splicing dataset and found to be more efficient than other techniques.
View
An initial point alignment method of narrow weld using laser vision sensor
Article
Full-text available
  • May 2019
  • INT J ADV MANUF TECH
In this paper, an initial point alignment method of narrow weld using laser vision sensor is presented on the basis of the relationship between the feature point of laser stripe and initial point. The whole initial point alignment process contains two stages. At the first stage, the initial point image is captured, and the image coordinates of the feature point of laser stripe and initial point are obtained. At the second stage, according to the relationship between the feature point of laser stripe and initial point, the three-dimensional (3D) coordinates of initial point could be determined to achieve initial point alignment. The initial point alignment method mainly includes vision sensing and motion control two parts. Firstly, a new laser vision sensor with a uniform LED surface light source is developed to capture the high signal-to-noise ratio (SNR) image including narrow weld, and the feature point of laser stripe and initial point are detected using the image processing method. Secondly, initial point alignment control system including feature verification and controller is designed to achieve initial point alignment control. Finally, a series of initial point alignment experiments of straight and curve narrow weld are conducted to test the performance of the proposed method. Experimental results indicate the alignment error is less than previous methods, which could be used in automatic welding process.
View
Strong noise image processing for vision-based seam tracking in robotic gas metal arc welding
Article
Full-text available
  • Apr 2019
  • INT J ADV MANUF TECH
The robustness of the image processing algorithm is very important based on vision sensor in robotic seam tracking, which will directly affect the accuracy of weld seam shaping quality. Especially in GMAW (Gas Metal Arc Welding), there is a lot of strong noise image. This paper studies an algorithm for the several weld seam images with strong noise in robotic GMAW, such as the atypical weld seam, the strong arc light and the large spatter. Based on a purpose-built visual sensing system, the fast image segmentation, the feature area recognition of the convolutional neural network (CNN), and the feature search technique are used to identify the weld seam features accurately in the algorithm. The selection range of the threshold is increased from 0.5 × 10⁷ to 0.9 × 10⁷ by using the proposed algorithm, which reduces the difficulty of parameter adjustment and increases the stability of seam tracking system. And, the accuracy of the CNN model was 98.0% for the atypical weld seam identification. To evaluate the robustness of the proposed algorithm, the accuracy is verified using experiments on two typical strong noise images. The experiments show that the average error of feature extraction accuracy is 0.26 mm and 0.29 mm. The results show that the proposed algorithm can extract the feature of weld seam image with strong noise accurately and effectively.
View
Narrow-seam identification and deviation detection in keyhole deep-penetration TIG welding
Article
Full-text available
  • Apr 2019
  • INT J ADV MANUF TECH
A narrow-seam identification algorithm is developed to achieve seam tracking in keyhole deep-penetration tungsten inert gas welding (TIG). The welding images are captured by a high-dynamic-range camera and denoised by a bilateral filter based on a noise model analysis. The arc area is extracted as a fixed region of interest. Then, an improved Otsu algorithm and a parabolic fitting algorithm are used to obtain the centerline of the arc. The seam area is extracted as an adaptive region of interest based on a proposed HOG+LBP algorithm. Thereafter, a continuous single-pixel edge contour is extracted by the canny algorithm, and a proposed contour curvature evaluation method is used to obtain the corresponding pixel coordinates. After testing and analysis, the deviation can be reliably detected with an average measurement error within ± 0.04 mm. As a result, the algorithm proposed in this study can accurately identify the deviation during keyhole deep-penetration TIG welding, and has application prospects in the narrow-seam welding field.
View
Automatic Nucleus Segmentation with Mask-RCNN
Chapter
  • Jan 2020
Mask-RCNN is a recently proposed state-of-the-art algorithm for object detection, object localization, and object instance segmentation of natural images. In this paper, it is demonstrated that Mask-RCNN can be used to perform highly effective and efficient automatic segmentations of a wide range of microscopy images of cell nuclei, for a variety of cells acquired under a variety of conditions. In addition, it is shown that a cyclic learning rate regime allows effective training of a Mask-RCNN model without any need to finetune the learning rate, thereby eliminating a manual and time-consuming aspect of the training procedure. The results presented here will be of interest to those in the medical imaging field and to computer vision researchers more generally.
View
View
Mask R-CNN
Article
  • Jun 2018
We present a conceptually simple, flexible, and general framework for object instance segmentation. Our approach efficiently detects objects in an image while simultaneously generating a high-quality segmentation mask for each instance. The method, called Mask R-CNN, extends Faster R-CNN by adding a branch for predicting an object mask in parallel with the existing branch for bounding box recognition. Mask R-CNN is simple to train and adds only a small overhead to Faster R-CNN, running at 5 fps. Moreover, Mask R-CNN is easy to generalize to other tasks, \eg, allowing us to estimate human poses in the same framework. We show top results in all three tracks of the COCO suite of challenges, including instance segmentation, bounding-box object detection, and person keypoint detection. Without bells and whistles, Mask R-CNN outperforms all existing, single-model entries on every task, including the COCO 2016 challenge winners. We hope our simple and effective approach will serve as a solid baseline and help ease future research in instance-level recognition. Code has been made available at: https://github.com/facebookresearch/Detectron.
View
Microstructure evolution and mechanical properties of keyhole deep penetration TIG welds of S32101 duplex stainless steel
Article
  • Jan 2018
  • MAT SCI ENG A-STRUCT
In this research, workpieces of 10.8 mm thickness S32101 duplex stainless steel (DSS) were welded using a novel keyhole deep penetration TIG (DP-TIG) welding system in a single pass without groove preparation and filler metals. The weld geometry profiles conducted with different welding currents and welding speeds were analyzed. The microstructure evolution of the welded joints was carefully observed by means of optical microscope (OM) and scanning electron microscopy (SEM). Microhardness and mechanical properties of the weldments were tested and investigated in detail. The results show that the misorientation angle distribution of ferrite and austenite gradually migrated from high angle grain boundaries (HAGBs, θ ≥ 15°) to low angle grain boundaries (LAGBs, 2° ≤ θ ≤ 5°) during the welding process. In addition, a large number of Σ3 CSL boundaries gradually disappeared and the texture of austenite evolved from the original cubic texture into copper texture. The phase fractions of the weld metal (WM) are close to that of the base metal (BM). The impact property of the WM was lower than that of the BM. However, the tensile strength and microhardness of the WM were superior in comparison to the BM. This indicates that the one-sided welding with double-sided shaping of 10.8 mm thickness S32101 DSS can be welded by using the high productivity, low-cost and defect-free DP-TIG welding equipment. It also implies that DP-TIG welding can be widely applied and has many advantages in industrial applications.
View
View
Stable keyhole welding process with K-TIG
Article
  • Jul 2016
  • J MATER PROCESS TECH
Keyhole welding process was successfully achieved by using a self-made K-TIG torch. High quality welds were produced in 8mm-thick stainless steel plates. To evaluate the keyhole stability in the novel manufacturing process, keyhole behavior was observed by using a vision system from backside of the workpiece. Efflux plasma interference was eliminated by using a filter glass. Keyhole exit behavior was imaged in real time during the welding process, keyhole size and position were extracted from the keyhole image sequence. Keyhole behavior parameters, including keyhole length, width, length/width, and deviation distance were measured. It was found that keyhole exit was deviated away from the torch axis even though the arc current was very high in K-TIG welding process, keyhole exit had oval shape. Continuous open keyhole welding were easily achieved in K-TIG process, without unstable keyhole stage when the keyhole firstly opens. Keyhole size, shape and position were all related to welding current. The observation results lay solid foundation to understand the thermal-physical behavior in K-TIG, and for the further optimizing of the welding torch.
View