A fast and adaptive method for automatic weld defect detection in various real-time X-ray imaging systems

A first and effective method is proposed to detect weld deject adaptively in various Dypes of real-time X-ray images obtained in different conditions. After weld extraction and noise reduction, a proper template of median filter is used to estimate the weld background. After the weld background is subtracted from the original image, an adaptite threshold segmentation algorithm is proposed to obtain the binary image, and then the morphological close and open operation, labeling algorithm and fids＇e alarm eliminating algorithm are applied to pracess the binary image to obtain the defect, ct detection result. At last, a fast realization procedure jbr proposed method is developed. The proposed method is tested in real-time X-ray image,s obtairted in different X-ray imaging sutems. Experiment results show that the proposed method is effective to detect low contrast weld dejects with few .false alarms and is adaptive to various types of real-time X-ray imaging systems.

Multiaxial Fatigue Analysis on Reeled Deepwater Steel Catenary Risers with Girth Weld Defects

In the present study, we simulated the reel-lay installation process of deepwater steel catenary risers（SCRs） using the finite element method and proposed multiaxial fatigue analysis for reeled SCRs. The reel-lay method is one of the most efficient and economical pipeline installation methods. However, material properties of reeled risers may change, especially in the weld zone, which can affect the fatigue performance. Applying finite element analysis（FEA）, we simulated an installation load history through the reel, aligner, and straightener and analyzed the property variations. The impact of weld defects during the installation process, lack of penetration and lack of fusion, was also discussed. Based on the FEA results, we used the Brown-Miller criterion combined with the critical plane approach to predict the fatigue life of reeled and non-reeled models. The results indicated that a weld defect has a significant influence on the material properties of a riser, and the reel-lay method can significantly reduce the fatigue life of SCRs. The analysis conclusion can help designers understand the mechanical performance of welds during reel-lay installation.

Application of Support Vector Machine in Weld Defect Detection and Recognition of X-ray Images

Support vector machines（SVM） received wide attention for its excellent ability to learn, it has been applied in many fields. A review of the application of SVM in weld defect detection and recognition of X-ray image is been presented. We will show some commonly used methods of weld defect detection and recognition using SVM, and the advantages and disadvantages of each method will be discussed. SVM appears to be promising in weld defect detection and recognition, but future research is needed before it fully mature in this filed.

Contrast Correction Using Hybrid Statistical Enhancement on Weld Defect Images

Luminosity and contrast variation problems are among the most challenging tasks in the image processing field,significantly improving image quality.Enhancement is implemented by adjusting the dark or bright intensity to improve the quality of the images and increase the segmentation performance.Recently,numerous methods had been proposed to normalise the luminosity and contrast variation.A new approach based on a direct technique using statistical data known as Hybrid Statistical Enhancement(HSE)is presented in this study.TheHSE method uses themean and standard deviation of a local and global neighbourhood and classified the pixel into three groups;the foreground,border,and problematic region(contrast&luminosity).The datasets,namely weld defect images,were utilised to demonstrate the effectiveness of the HSE method.The results from the visual and objective aspects showed that the HSE method could normalise the luminosity and enhance the contrast variation problem effectively.The proposed method was compared to the two(2)populor enhancement methods which is Homomorphic Filter(HF)and Difference of Gaussian(DoG).To prove the HSE effectiveness,a few image quality assessments were presented,and the results were discussed.The HSE method achieved a better result compared to the other methods,which are Signal Noise Ratio(8.920),Standard Deviation(18.588)and Absolute Mean Brightness Error(9.356).In conclusion,implementing the HSE method has produced an effective and efficient result for background correction and quality images improvement.

Studies on the effect of weld defect on the fatigue behavior of welded structures

Structural integrity isstated as the science and technology of margin between safety and disaster. Systematic prediction of structural integrity of critical structures such ascombustion chambers,pressure vessels,nuclear reactor components,boilers etc.,ensures the human safety,environmental protection,and the economical considerations.The present work aims at prediction of fatigue behaviour of symmetric structures like pressure vessels in the presence of common welding defects such as lack of fusion( LOF),lack of penetration( LOP) and porosity.A ring type specimen which replicates the stress pattern in thepressure vessel is considered for the study of severity of weld imperfections. Initial dimensions of weld defects are arrived by performing NDT inspection.Crack growth analysis is carried out to determine the remaining life of the welded joint with defects.

Automated identification of steel weld defects,a convolutional neural network improved machine learning approach

This paper proposes a machine-learning-based methodology to automatically classify different types of steel weld defects,including lack of the fusion,porosity,slag inclusion,and the qualified(no defects)cases.This methodology solves the shortcomings of existing detection methods,such as expensive equipment,complicated operation and inability to detect internal defects.The study first collected percussed data from welded steel members with or without weld defects.Then,three methods,the Mel frequency cepstral coefficients,short-time Fourier transform(STFT),and continuous wavelet transform were implemented and compared to explore the most appropriate features for classification of weld statuses.Classic and convolutional neural network-enhanced algorithms were used to classify,the extracted features.Furthermore,experiments were designed and performed to validate the proposed method.Results showed that STFT achieved higher accuracies(up to 96.63%on average)in the weld status classification.The convolutional neural network-enhanced support vector machine(SVM)outperformed six other algorithms with an average accuracy of 95.8%.In addition,random forest and SVM were efficient approaches with a balanced trade-off between the accuracies and the computational efforts.

Weldability of steel used in an aged bridge

A series of experiments was performed to investigate the weldability of steel used in an aged bridge.A steel material used in an aged railway bridge constructed in 1912 was extracted for this investigation.The chemical compositions of this steel were suitable for welding.However,the aged steel contained much sulfur.Cruciform welded joints were fabricated with this aged steel.Welding defects or cracks were not observed in the joints.The Vickers hardness test on the welded part did not confirm extreme hardening or softening.After yielding by the static tensile test,the cruciform joints were fractured at the welded parts.One of the specimens was fractured in the middle of the thickness of the aged steel.The Sulfur contained in the aged steel might cause this type of fracture.The results show that there may be a risk of brittle fracture not only from the welded part but also from the base metal.The chemical compositions of aged steel must be analyzed when repair welding is applied to the steel.

Automatic Detection of Weld Defects in Pressure Vessel X-Ray Image Based on CNN

The visual automatic detection method based on artificial intelligence has attracted more and more attention. In order to improve the performance of weld nondestructive defect detection,we propose DRepDet(Dilated RepPoints Detector). First, we analyze the weld defect dataset in detail and summarize the distribution characteristics of weld defect data, that is, the defect scale is very different and the aspect ratio distribution range is large. Second, according to the distribution characteristics of defect data, we design DResBlock module, and introduce dilated convolution with different dilated rates in the process of feature extraction to expand the receptive field and improve the detection performance of large-scale defects. Based on DResBlock and anchor-free detection framework RepPoints, we design DRepDet. Extensive experiments show that our proposed detector can detect 7 types of defects. When using combined dilated rate convolution network in detection, the AP50 and Recall50 of big defects are improved by 3.1% and 3.3% respectively, while the performance of small defects is not affected, almost the same or slightly improved. The final performance of the whole network is improved a large margin,with 6% AP50 and 4.2% Recall50 compared with Cascade RCNN and 1.4% AP50 and 2.9% Recall50 compared with RepPoints.

Feature extraction of welding defect based on machine vision

There are many flaws in welding images such as noise, low contrast, and blurred edges, which affect feature extraction from welding defect regions and impede classification and recognition of welding defects. To deal with the complexity of welding defect images, this paper proposes an effective method for extracting the features of welding defect regions. Firstly, image preprocessing, image segmentation and image background removal are carried out to a welding image in order to extract welding defect region; and then an 8-connected-component labeling method is used to mark defect regions. Finally, it extracts geometric characteristic parameters including perimeter, area, circularity and others. The experimental result shows that the method proposed in the paper can accurately extract the features of welding defect regions. It has good adaptability and practicability.

Noise Reduction of Welding Defect Image Based on NSCT and Anisotropic Diffusion

In order to reduce noise effectively in the welding defect image and preserve the minutiae information, a noise reduction method of welding defect image based on nonsubsampled contourlet transform(NSCT) and anisotropic diffusion is proposed. Firstly, an X-ray welding defect image is decomposed by NSCT. Then total variation(TV) model and Catte_PM model are used for the obtained low-pass component and band-pass components, respectively. Finally, the denoised image is synthesized by inverse NSCT. Experimental results show that, compared with the hybrid method of wavelet threshold shrinkage with TV diffusion, the method combining NSCT with P_Laplace diffusion, and the method combining contourlet with TV model and adaptive contrast diffusion, the proposed method has a great improvement in the aspects of subjective visual effect, peak signal-to-noise ratio(PSNR) and mean-square error(MSE). Noise is suppressed more effectively and the minutiae information is preserved better in the image.

Repair welding process of friction stir welding groove defect

The groove defect formed in the friction stir welding dramatically deteriorates weld appearances and mechanical properties of the joints owing to its larger size and penetration. Therefore, the friction stir repair welding was utilized to remove such a groove defect, and the focus was placed on the mechanical properties and microstructural characteristics of the repair joints so as to obtain an optimum repair welding process. The experimental results indicate that the groove defect can be removed by friction stir repair welding, and the offset repair welding process is superior to the symmetrical repair welding process. In the symmetrical repair welding process, a large number of fine cavity defects and an obvious aggregation of hard-brittle phase Al2Cu occur, accordingly the mechanical properties of the repair joint are weakened, and the fracture feature of repair joint is partially brittle and partially plastic. A good-quality repair joint can be obtained by the offset repair welding process, and the repair joint is fractured near the interface between the weld nugget zone and thermal-mechanically affected zone.

FEATURE ANALYSIS AND EXTRACTION OF INNER DEFECTS FOR SPOT WELD NUGGET OF ALUMINUM ALLOY

１ＩＮＴＲＯＤＵＣＴＩＯＮＳｐｏｔｗｅｌｄｉｎｇｈａｓｂｅｅｎｗｉｄｅｌｙｕｓｅｄｉｎａｌｕｍｉｎｕｍａｌｏｙｓｔｒｕｃｔｕｒｅｓｏｆｓｐａｃｅｃｒａｆｔａｎｄａｉｒｃｒａｆｔｄｕｅｔｏｉｔｓｃｈａｒａｃｔｅｒｉｓｔｉｃｓｏｆｔｅｃｈｎｏｌｏｇｙ．...

Study on intelligent image welding defects in recognition for the tube pressure vessels

Currently, the welding defects recognition of X-ray nondestructive inspection is principally carried out by manual work, which highly depends on the experience of the inspectors and costs plenty of workload. In this paper, an intelligent image processing and recognition method for the tube welding radiographic testing in large-scale pressure vessels is proposed. Firstly, the raw image is preprocessed by median filtering, pseudo point removing and non-lincar image enhancement. Secondly, the welded joints parts are separated from the whole image by edge detection and threshold segmentation algorithms. Then, the separated images are handled by FFT transformation. Finally, whether defects exist and the specific type of defects are judged by Support Vector Machine. Software developed basing on this method works stably on site, and experiments demonstrate that the recognition results are compliance with the JB/T 4730. 2 or ASME standards.

Classified recognition for metal magnetic memory signals of welding defects in API 5L X65 pipeline steel

Feature extraction and selection from signals is a key issue for metal magnetic memory testing technique. In order to realize the classification of metal magnetic memory signals of welding defects, four fractal analysis methods, such as box- counting, detrended fluctuation, minimal cover and rescaled-range analysis, were used to extract the feature signal after the original metal magnet memory signal was de-noising and differential processing, then the Karhunen-Lo^e transformation was adopted as classification tool to identify the defect signals. The result shows that this study can provide an efficient classification method for metal magnetic memory signal of welding defects.

Formative mechanism of the undercut defect in horizontal variable polarity plasma arc welding

Abstract Horizontal welding is important for heavy or huge welding structures. Keyhole mode variable polarity plasma arc welding of aluminum alloy plates with medium thickness was carried out in horizontal position. The characteristic of welding defects was introduced. Preliminary experiments indicated that the undercut defect could not be eliminated easily. The relationship between welding parameters and the undercut defect showed that this deject could be lessened by using higher heat input. The fluid flow of weld pool was observed by a high speed camera. The fluid flow in weld pool was not symmetric and much of molten metal gathered in the lower part. The fluid flow velocity in the lower part was bigger than that in the upper part. To this end, the formative mechanism of the undercut defect was proposed. The flowability of the molten metal was an influential factor for the undercut defect. A preheating method was designed to verify the formative mechanism.

基于声发射的钢桥面板焊接气孔缺陷在线识别

为实现正交异性钢桥面板机器人智能化焊接过程中缺陷的在线监测,提出了一种基于快速傅里叶变换和支持向量机的气孔缺陷声发射识别方法.通过开展机器人焊接实验,揭示了钢桥面板焊接及缺陷产生过程的声发射特征.无损伤与气孔缺陷2种工况信号的幅值、振铃计数、峰值频率和中心频率等参数重叠交叉严重、相关性不显著,而气孔缺陷信号的傅里叶频谱存在更多高频能量分布,因此以频谱为输入建立2种工况的径向基核支持向量机模型.实验结果表明,与朴素贝叶斯、随机森林和线性核支持向量机模型相比,径向基核支持向量机模型拥有更高的正确率(95.4%)和召回率(94.3%),能够用于焊接过程气孔缺陷的在线识别,具有较强的鲁棒性和实用性.

基于改进YOLOv8的地铁列车焊缝缺陷轻量化检测方法

针对现有的地铁列车车体焊接质量检测技术存在检测模型较大、检测精度和效率较低的问题,提出一种基于改进YOLOv8的焊缝缺陷轻量化检测方法。首先,利用相控阵超声波检测仪采集对接焊缝内部缺陷图像,通过图像预处理制作成焊缝缺陷数据集。然后,在YOLOv8模型基础上,利用Inner-SIoU优化原有损失函数、采用C2f-PConv替换C2f模块、引入大型可分离核注意力(LSKA)模块和挤压激励(SE)注意力机制,建立了基于改进YOLOv8的地铁列车车体焊缝缺陷质量检测模型,以提高焊缝缺陷特征提取和多尺度特征融合的能力。最后,利用改进的YOLOv8模型对焊缝缺陷数据集进行训练和测试。结果表明,改进的YOLOv8模型大小为7.91 M,对于焊缝缺陷的检测精度达到98.30%,检测速度达到138.9帧/s,与YOLOv8原始模型相比,模型更小,检测精度更高。

Defects of friction stir welded 1060 aluminum alloy and their effects on joint tensile property

9. 6 mm thick 1060-H24 aluminum alloy plates were friction stir welded and the influencing factors on groove and tunnel defects were examined. Results show that the welding speed range for achieving a groove-free joint is enlarged with increasing the rotating speed. The tunnel size decreases with decreasing the welding speed under the same rotating speed. Excessive or insufficient shoulder plunge depth will cause defective joints. At a relatively low rotating speed and high welding speed, the tool having a larger shoulder diameter has a larger range of processing parameters to obtain a groove-free joint. Moreover, the tensile fracture behaviors of the defective and defect-free samples are different.

焊缝缺陷三维成像及亚像素定量方法研究

为实现工业设备中焊缝结构内部缺陷准确地定量分析,针对构造缺陷三维重构模型时缺陷图谱的过分割和欠分割导致定量分析不够精确的问题,提出了一种基于图谱颜色分割的新算法。利用该算法处理缺陷切片并进行三维重构获取其长度,应用Canny算子与三次样条插值法结合获取缺陷亚像素边缘,基于缺陷区域连通标记算法和累加法可获取缺陷面积和体积的方法。研究结果表明:与实际数据对比,未焊透、未熔合、夹渣缺陷重构的长度误差分别为8.18%、5%、9.62%,基于此对分割缺陷的亚像素边缘内连通域进行标记以准确地获取其面积及体积。该方法能够在不破坏焊缝结构的情况下即时获取焊缝内部缺陷的形貌并对其进行精准且全面的定量分析。

航空铝合金薄板搅拌摩擦焊内部缺陷的超声相控阵检测

针对3 mm厚度下航空铝合金薄板搅拌摩擦焊焊缝的检测需求,基于典型缺陷特征以及缺陷对焊接接头性能的影响规律分析,明确了搅拌摩擦焊无损检测应重点关注的缺陷类型;设计了不同深度和尺寸的隧道以及未焊透缺陷模拟样件,利用模拟缺陷样件验证了超声相控阵检测技术对铝合金搅拌摩擦焊中1 mm埋深0.2 mm孔径的隧道缺陷以及1.6 mm埋深2 mm长、0.2 mm深未焊透缺陷的检测能力,对3 mm厚度下搅拌摩擦焊实际样件的检测结果与X射线检测结果及断口观察结果吻合,超声相控阵检测技术可以用于3 mm以下航空铝合金薄板搅拌摩擦焊焊缝的检测。