Real-Time Image Processing for Monitoring of Free Weld Pool Surface

1997 ◽  
Vol 119 (2) ◽  
pp. 161-169 ◽  
Author(s):  
R. Kovacevic ◽  
Y. M. Zhang
Keyword(s):  
Weld Pool ◽  
Specular Reflection ◽  
Welding Process ◽  
Ccd Camera ◽  
Weld Penetration ◽  
Reflection Pattern ◽  
Pool Surface ◽  
Edge Points ◽  
On Line ◽  
Weld Pool Surface

The arc weld pool is always deformed by plasma jet. In a previous study, a novel sensing mechanism was proposed to sense the free weld pool surface. The specular reflection of pulsed laser stripes from the mirror-like pool surface was captured by a CCD camera. The distorted laser stripes clearly depicted the 3D shape of the free pool surface. To monitor and control the welding process, the on-line acquisition of the reflection pattern is required. In this work, the captured image is analyzed to identify the torch and electrode. The weld pool edges are then detected. Because of the interference of the torch and electrode, the acquired pool boundary may be incomplete. To acquire the complete pool boundary, models have been fitted using the edge points. Finally, the stripes reflected from the weld pool are detected. Currently, the reflection pattern and pool boundary are being related to the weld penetration and used to control the weld penetration.

Study on Dynamic Development of Three-dimensional Weld Pool Surface in Stationary GTAW

2018 ◽  
Vol 37 (5) ◽  
pp. 455-462 ◽  
Author(s):  
Jiankang Huang ◽  
Jing He ◽  
Xiaoying He ◽  
Yu Shi ◽  
Ding Fan
Keyword(s):  
Weld Pool ◽  
Specular Reflection ◽  
Three Dimensional ◽  
Welding Process ◽  
Surface Shape ◽  
Weld Penetration ◽  
Dynamic Development ◽  
Pool Surface ◽  
Gtaw Process ◽  
Weld Pool Surface

AbstractThe weld pool contains abundant information about the welding process. In particular, the type of the weld pool surface shape, i. e., convex or concave, is determined by the weld penetration. To detect it, an innovative laser-vision-based sensing method is employed to observe the weld pool surface of the gas tungsten arc welding (GTAW). A low-power laser dots pattern is projected onto the entire weld pool surface. Its reflection is intercepted by a screen and captured by a camera. Then the dynamic development process of the weld pool surface can be detected. By observing and analyzing, the change of the reflected laser dots reflection pattern, for shape of the weld pool surface shape, was found to closely correlate to the penetration of weld pool in the welding process. A mathematical model was proposed to correlate the incident ray, reflected ray, screen and surface of weld pool based on structured laser specular reflection. The dynamic variation of the weld pool surface and its corresponding dots laser pattern were simulated and analyzed. By combining the experimental data and the mathematical analysis, the results show that the pattern of the reflected laser dots pattern is closely correlated to the development of weld pool, such as the weld penetration. The concavity of the pool surface was found to increase rapidly after the surface shape was changed from convex to concave during the stationary GTAW process.

scholarly journals Deep Learning-Based Detection of Penetration from Weld Pool Reflection Images

2020 ◽  
Vol 99 (9) ◽  
pp. 239s-245s
Author(s):  
CHAO LI ◽  
◽  
QIYUE WANG ◽  
WENHUA JIAO ◽  
MICHAEL JOHNSON ◽  
...  
Keyword(s):  
Weld Pool ◽  
Data Augmentation ◽  
Welding Process ◽  
Descent Method ◽  
Machine Learning Techniques ◽  
Gradient Descent Method ◽  
Neural Net ◽  
Weld Penetration ◽  
Process Data ◽  
Pool Surface

An innovative method was proposed to determine weld joint penetration using machine learning techniques. In our approach, the dot-structured laser images reflected from an oscillating weld pool surface were captured. Experienced welders typically evaluate the weld penetration status based on this reflected laser pattern. To overcome the challenges in identifying features and accurately processing the images using conventional machine vision algorithms, we proposed the use the raw images without any processing as the input to a convolutional neural network (CNN). The labels needed to train the CNN were the measured weld penetration states, obtained from the images on the backside of the workpiece as a set of discrete weld penetration categories. The raw data, images, and penetration state were generated from extensive experiments using an automated robotic gas tungsten arc welding process. Data augmentation was performed to enhance the robustness of the trained network, which led to 270,000 training examples, 45,000 validation examples, and 45,000 test examples. A six-layer convolutional neural net-work trained with a modified mini-batch gradient descent method led to a final testing accuracy of 90.7%. A voting mechanism based on three continuous images increased the classification accuracy to 97.6%.

scholarly journals Numerical Simulation and Industrial Application of High Speed Tandem TIG Welding

2021 ◽  
Author(s):  
Dongsheng Wu ◽  
Jiuling Huang ◽  
Kong Liang ◽  
Xueming Hua ◽  
Min Wang ◽  
...  
Keyword(s):  
Molten Metal ◽  
Weld Pool ◽  
High Speed ◽  
Metal Flow ◽  
Welding Process ◽  
Tig Welding ◽  
Pool Surface ◽  
Weld Pool Surface ◽  
Welding Processes ◽  
Molten Metal Flow

Abstract Self-developed high speed tandem TIG welding equipment were adopted to manufacture titanium welded tubes with high efficiency and high quality. The joint made by this high efficient welding process met Chinese standard requirements. A coupled electrode, arc and weld pool numerical model was also developed to investigate temperature and velocity distributions, and energy propagation of this welding process. The numerical results showed that the Marangoni stress was much higher than the arc shear stress, and was mainly positive after leading and trailing arcs in the x and y directions, so the molten metal flowed backward on the top weld pool surface. Previous studies proposed that a “pull-push” flow pattern defined as a backward molten metal flow after the leading arc and a forward molten metal flow before the trailing arc existed on the top weld pool surface in tandem arc welding processes, while it was not observed in this case. The calculated arc efficiency of the high speed tandem TIG welding was about 79.8%.

scholarly journals The Reconsitution of the Weld Pool Surface in Stationary TIG Welding Process With Filler Wire

2020 ◽  
Author(s):  
Jiankang Huang ◽  
LIU Guangyin ◽  
HE Jing ◽  
YU Shurong ◽  
LIU Shien ◽  
...  
Keyword(s):  
Weld Pool ◽  
Three Dimensional ◽  
Welding Process ◽  
Tig Welding ◽  
Filler Wire ◽  
Surface Evolution ◽  
Recovery Algorithm ◽  
Pool Surface ◽  
Weld Pool Surface ◽  
Dimensional Surface

Abstract In order to study the dynamic characteristics of the weld pool surface during the TIG welding process of the filler wire, an observation test platform for the study of the three-dimensional surface behavior evolution of the TIG weld pool based on the grid structure laser was used to observe the weld pool surface and obtain the reflection grid laser image. The three-dimensional surface evolution of the fixed-point TIG welding pool is accurately restored by the three-dimensional recovery algorithm of the weld pool surface, so as to obtain the three-dimensional surface morphology of the weld pool. The difference between the obtained weld pool height and the experimental results is very small, and the results are basically the same.

scholarly journals Automated Recognition of Weld Pool Characteristics from Active Vision Sensing

2021 ◽  
Vol 100 (5) ◽  
Author(s):  
YONGCHAO CHENG ◽  
◽  
QIYUE WANG ◽  
WENHUA JIAO ◽  
JUN XIAO ◽  
...  
Keyword(s):  
Weld Pool ◽  
Engineering Application ◽  
Welding Process ◽  
Active Vision ◽  
Node Number ◽  
Automated Recognition ◽  
Gas Tungsten Arc ◽  
Pool Surface ◽  
Laser Stripe ◽  
Weld Pool Surface

While penetration occurs underneath the workpiece, the raw information used to detect it during welding must be measurable to a sensor attached to the torch. Challenges are apparent because it is difficult to find such measurable raw information that fundamentally correlates with the phenomena occurring underneath. Additional challenges arise because the welding process is extremely complex such that analytically correlating any raw information to the underneath phenomena is practically impossible; therefore, handcrafted methods to propose features from raw information are human dependent and labor extensive. In this paper, the profile of the weld pool surface was proposed as the raw information. An innovative method was proposed to acquire it by projecting a single laser stripe on the weld pool surface transversely and intercepting its reflection from the mirror-like weld pool surface. To minimize human intervention, which can affect success, a deep-learning-based method was proposed to automatically recognize features from the single-stripe active vision images by fitting a convolutional neural network (CNN). To train the CNN, spot gas tungsten arc welding experiments were designed and conducted to collect the active vision images in pairs with their actual penetration states measured by a camera that views the backside surface of the workpiece. The CNN architecture was optimized by trying different hyperparameters, including kernel number, kernel size, and node number. The accuracy of the optimized model is about 98% and the cycle time in the personal computer is ~ 0.1 s, which fully meets the required engineering application.

Sensing Free Surface of Arc Weld Pool Using Specular Reflection: Principle and Analysis

1996 ◽  
Vol 210 (6) ◽  
pp. 553-564 ◽  
Author(s):  
R Kovacevic ◽  
Y M Zhang
Keyword(s):  
Weld Pool ◽  
Specular Reflection ◽  
Three Dimensional ◽  
Surface Shape ◽  
Pool Surface ◽  
Imaging Model ◽  
Dimensional Shape ◽  
Weld Pool Surface ◽  
Welding Processes ◽  
Three Dimensional Shape

The weld pool surface provides important information for understanding arc welding processes. In this study, a novel vision sensor is proposed to measure the three-dimensional shape of the free weld pool surface. A pulsed laser is projected on to the weld pool through a specific grid. Specular reflection from the pool surface is sensed using a high shutter speed camera. The three-dimensional weld pool surface shape is clearly shown by the specular reflection. To determine the shape of the pool surface, an image processing technology has been developed to extract the skeleton of the specular reflection from the acquired image. The imaging principle is analysed to determine the correlation between the reflection and the weld pool surface. If the weld pool surface is known, the corresponding specular reflection can be directly calculated using the imaging model which is derived based on the reflection law. However, no explicit models can be obtained to determine the weld pool surface using the reflection and sensor parameters. To solve this difficulty, an iterative algorithm is proposed. The weld pool surface can now be calculated in 1 second from the specular reflection of the weld pool surface. A higher calculation speed is currently being pursued.

Sign in / Sign up

Export Citation Format

Share Document