Expert for pre insulated duct CNC cutting machine.

wavelet based feature extraction method for quantitative characterization of porosity in gas tungsten arc welding by infrared thermography in aisi 316 stainless steel for on-line monitoring and control.

by:VENTECH     2020-03-07
Introduction to electric welding of gas tungsten (GTAW)
Commonly referred to as tungsten inert gas (TIG)
Welding is most suitable for precision welding in atomic energy, aircraft, chemical and instrument industries.
This is an arc welding process in which the knot is generated by heating the workpiece with an arc between the tungsten electrode and the workpiece.
The ashelding gas is used to prevent atmospheric contamination of moltenweld pools.
TIG welding is a widely used method of metal connection.
Although many advances have been made in welding science and technology, there has indeed been a failure and welding is still considered the weakest part.
This is because the formation of the weld is affected by multiple process parameters, making the quality of the weld difficult to guarantee.
Generally, the quality of the weld can only be determined after the welding is completed by using nondestructive testing (NDT)
Such as ultrasound or radiography.
Since these technologies are all applied after the welding is completed, a lot of time, material resources and manpower are wasted before understanding the soundness of the weld.
The inherent limitations in the traditional welding process can beovercome if the welding continuously monitors the application of real-time defects and its automatic elimination
Online control of welding parameters.
In addition, defective welds can be repaired immediately without further processing.
This strategy of controlling, controlling, and maintaining weld quality is often referred to as adaptive welding or intelligent welding.
As it implies, intelligent welding will weld equipment with intelligent sensing and control, knowledge of human experts and artificial intelligence (AI)
Improve connection efficiency and reduce weld non-uniformity and defects.
Sensors are the key to the success of intelligent welding. Non-
Destructive Testing (NDT)
Sensor foron that has been considered-
Line monitoring includes optics, radiography, and infrared (IR).
The information provided by these three sensors is limited to the surface, such as the width of the beads;
Dislocation, etc.
The real-time radiography system based on the image booster has been used for the process control of arcwelding.
However, the dangers involved in the use of radiation sources limit potential applications.
On the other hand, the advantage of infrared is that it can reveal both surface and near surface disturbances.
After obtaining the thermal image, the feature corresponding to the welding defect is extracted.
These feature vectors are related to the response bias of the physical parameters that cause the defect.
Through this mapping, the corresponding physical parameters are controlled to produce defect-free welds.
This paper presents an algorithm for automatic recognition and quantification.
The paper is organized as follows.
The second section briefly reviews the infrared thermal image in nondestructive testing.
The experimental settings are provided in section 3.
The feature extraction algorithm and wavelet transform are reviewed in sections 6 and 5.
The topic is discussed in section 6.
Results are provided in Section VII, and work is presented in Section VIII.
These functions are implemented in Mat lab.
Infrared Thermal Imaging for related work is not a new online technology
Line monitoring.
Many groups around the world have used infrared research techniques in the inspection of underground defects and features, thermal physical properties, coating thickness and hidden structures.
A heat map is used to control problems in the welding process, such as arc dislocation [1].
Infrared sensors are best suited for welding quality detection, because the disturbances generated due to changes in arc positioning, thermal input, and the presence of contaminants differ in spatial and temporal surface temperature distributions.
Therefore, image analysis technology can be developed to quantify the change of temperature distribution, so as to realize the adaptive welding technology of automatic welding control [2].
Infrared thermal image for on-
On-line control of Torch Path for robot tungsten gas protection welding.
Temperature changes around the welding torch were recorded using an infrared camera.
These images are transmitted to the central computer, an image processing algorithm has been developed on the central computer to determine the torch from the joint and to transmit corrective measures to control the torch path.
However, the method developed only applies to a single V-
Slot configuration [3].
Infrared sensing and computer image processing technology can be used as a feasible method to improve the welding process by dynamic control of the penetration parameters of the joint.
During the welding process, the infrared camera is installed at the front of the weld, and the surface temperature distribution around the weld is measured.
In order to obtain different penetration depth, the welding parameters were changed and the corresponding temperature distribution was recorded.
How to consider the relative temperature relative to the specific selected temperature, not the absolute temperature [4].
Infrared sensing technology is used to track the curved profiles of gap injection reactor welded joints.
It is found that the gap in the welded joint can cause a significant decrease in the measured infrared strength or temperature, which can be used to determine the size and position of the joint gap [5].
Infrared thermal images are ideal for detecting changes in bead width and penetration depth due to changes in plate thickness, shielding gas composition, and trace element content in GTAW.
Macro temperature gradient determined from peak temperature and solid temperature
Liquid metal surface for implementing weld penetration control6].
In recent years, the application of infrared thermal imaging in nondestructive testing and evaluation has been extensively studied.
Provide non
Wide area detection of contact and underground defects.
Different passive and active thermal imaging techniques are used for defect detection.
Effective technologies include pulse thermal imaging, local thermal imaging, pulse phase thermal imaging, and vibration thermal imaging [7-9].
In order to study the welding process, the infrared thermal image is used to determine the transient thermal field accompanying the welding processof-
Flat distortion.
Heat maps reveal important features of thermal conditions that cannot be theoretically modeled in practical sense [10].
Detection and quantification of the lack of penetration and tungsten inclusions from thermal maps.
Through histogram equalization, image segmentation and morphological image processing, the quantification of heat map is realized through image processing algorithm.
Then, extract these features from the algorithm for on-
On-line weld monitoring for production of defect-free welds [11].
A technique based on inversion is used to extract features from thermal maps.
The phase and amplitude of the image are obtained by fast Fourier transform, and it is found that the phase profile of the defect has a unique inflection point blind frequency, so the defect is quantified [12].
Unlike traditional contrast-based methods, the thermal graphics signal reconstruction method is used to analyze thermal graphics data to improve detection and provide automatic pass/fault handling in the pulse thermal image [1]13].
The experimental device uses the precision TIG 375 automatic TIG welding machine for the experiment.
TIG welding or GTAW use non
Consumable tungsten electrode protected by inert gas.
The electrode is made of a pure tungsten electrode and mixed with a small amount of oxide (Tolia, zirconium)
It improves the stability of the arc and is easier to hit.
Because the process uses non-
The electrode is consumed and additional filling materials are usually added.
The experiment was 3mm thick by the American Iron and Steel Association (AISI)
Type steel plate with dimensions of 316x50mm.
Edge and surface of the preparation board using standard preparation technology to facilitate dockingwelding.
The experiment was carried out without filling the material.
DC technology, the electrode is connected to the negative pole, the workpiece is connected to the positive pole.
The infrared camera captures the temperature distribution at the weld in the case of 45 degrees angle with the weld plate.
The infrared camera obtains the heat map known as the heat map from the distribution, and the customized interface transmits these images from the camera to the computer for further analysis.
Subsequent reprocessing was performed to expand the heat meters of the heat map to standardize their comparison.
The camera detects infrared radiation used to characterize the thermal distribution of the welded plate.
The infrared camera determines the temperature distribution by sampling part of the emission energy within the wavelength range of 8 to 12? m.
Every scan of the camera is treated as a size image.
The frame rate of the camera is 40 ms.
Air holes are a group of small gaps mainly generated by closed gases.
The parent metals melted under the Arc tend to absorb gases such as hydrogen, carbon monoxide, nitrogen and oxygen if they exist around the molten metal pool [14].
This is caused by electrodes or longer arcs or faster arc propagation speeds or too low, too high arc currents or incorrect welding techniques or electrodes that have damping and
Air holes are deliberately introduced by adding grease to the weld.
In the heat map, in the heat zone with the highest temperature in the heat map, the pores are shown as relatively low temperature areas.
The obtained heat map file consists of 1747 frames.
The welding arc can hold up to 547 frames and then remove the welding torch.
Air holes appear from 97 frames.
The 181, 182 and 183rd frames depict the air hole rate as shown in the figure1, Fig. 2 and Fig. 3 respectively. [
Figure 1 slightly][
Figure 2:[
Figure 3 slightly]
Feature extraction algorithm
A review Image processing algorithm will be developed to separate hole properties and quantify hole properties.
Traditional image processing algorithms include the conversion from color to gray scale, edge detection by Sobel or Kani filters, isolation of air holes by morphological image processing operations, and quantitative representation of air holes.
However, the performance of the algorithm depends on the threshold selection of edge detection, the size and shape of the structural element selection for morphological image processing.
So, even if the nature is the same, the algorithm needs to write different programs for each heat map.
In addition, edge detection, expansion, Region Growth and erosion involve applying occlusion on each pixel, a time-consuming process.
Time is also a very important parameter because this algorithm is used for on-
Online monitoring of welds.
Therefore, it is important to develop independent and time-consuming standardized feature extraction algorithms to facilitate online extraction.
Online monitoring of welds.
This independent algorithm, which takes less time, is based on regional growth.
However, the region growth technology also involves the selection of seed pixel values and thresholds.
Feature extraction technology based on European distance can also directly extract the air hole rate from color images.
However, all of the above technologies are not suitable if there are defects in different resolutions.
Therefore, a standardized algorithm suitable for extracting defects at different resolutions needs to be developed.
Wavelet transform is suitable for multi-resolution analysis.
The application of wavelet transform in image is reviewed, and the image efficiency of the four subbands is obtained.
They are approximate, horizontal detail coefficients, vertical detail coefficients, and diagonal detail coefficients. Let f(t)
Any square product function. The continuous-
Time wavelet transform of F (t)
About wavelet [PSI](t)
Defined as W (a,b)[? ? ][[SIGMA]. sub. -[infinity]]. sup. +[infinity]]f(t)(1/[Square root ()a))[PSI]*((t-b)/a)dt (1)
Where a and B are real, * represent complex coupling, a is called a scale or expansion variable, and B represents a time offset or translation [16].
Carlson wagon travel agency provides all supported W for redundant sensethat signals (a,b)
No need to restore the original signal f (t).
A new representation of non-redundant wavelet is [
Mathematical expressions that cannot be reproduced in ASCII. ](2)
This equation does not involve continuum of expansion and translation;
Instead, discrete values of these parameters are used.
Two-dimensional sequence d (k,l)
It is called discrete wavelet transform.
Discrete is only in variables a and B.
The Haar base is obtained with a multi-resolution function.
The zoom camp is [phi]= [1. sub. [0,1]]. Thefilter h[n]is given by H[n]= {[2. sup. -1/2]
If n = 0, then (3)
There are two non-equations above.
Zero coefficient equal [2. sup. -1/2]at n=0 and n=1[17].
The Haar wavelet is [PSI](t)= {-1 if 0 [
Less than or equal to]
If 1/2 [, t
Custom message
Chat Online 编辑模式下无法使用
Leave Your Message inputting...
Thank you for your enquiry. We will get back to you ASAP