Pub Date : 2024-07-01DOI: 10.1784/insi.2024.66.7.438
Yue Chen, Xuehao Pan, Peiwen Shen
Welded joints frequently endure the composite stress of steel components. The presence of defects within these welded joints can significantly jeopardise the safety and performance of the welded structure. Magnetic memory testing technology has garnered substantial attention due to its ability to evaluate welding defects. However, the conventional zero-point pole theory, which serves as the foundation for defect assessment in practical detection, may lead to defect location and omission errors. In response to this challenge, scholars have conducted extensive research to accurately pinpoint the location and identify the types of defect within welds. This paper systematically reviews the mechanisms of magnetic memory welding defect detection, the factors that influence it, signal characteristic parameters, noise reduction in magnetic memory signals and the application of machine learning for quantitative assessment. By summarising these research advancements, this paper aims to address the current issues and provide guidance for the precise quantitative evaluation of welding defects in the future using metal magnetic memory technology.
{"title":"Overview of welding defect detection utilising metal magnetic memory technology","authors":"Yue Chen, Xuehao Pan, Peiwen Shen","doi":"10.1784/insi.2024.66.7.438","DOIUrl":"https://doi.org/10.1784/insi.2024.66.7.438","url":null,"abstract":"Welded joints frequently endure the composite stress of steel components. The presence of defects within these welded joints can significantly jeopardise the safety and performance of the welded structure. Magnetic memory testing technology has garnered substantial attention due to\u0000 its ability to evaluate welding defects. However, the conventional zero-point pole theory, which serves as the foundation for defect assessment in practical detection, may lead to defect location and omission errors. In response to this challenge, scholars have conducted extensive research\u0000 to accurately pinpoint the location and identify the types of defect within welds. This paper systematically reviews the mechanisms of magnetic memory welding defect detection, the factors that influence it, signal characteristic parameters, noise reduction in magnetic memory signals and the\u0000 application of machine learning for quantitative assessment. By summarising these research advancements, this paper aims to address the current issues and provide guidance for the precise quantitative evaluation of welding defects in the future using metal magnetic memory technology.","PeriodicalId":506650,"journal":{"name":"Insight - Non-Destructive Testing and Condition Monitoring","volume":"20 6","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141713810","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The detection and identification of noise from moving parts inside a sealed cavity is crucial for ensuring the reliability of sealed equipment. However, traditional noise recognition methods struggle to meet the stringent demands for high detection accuracy. Inspired by the idea of ensemble learning, this paper proposes a noise recognition method that combines recognition results with high-dimensional mapping to enhance the recognition of noise. Firstly, a built noise identification experimental system is used to collect signals. Then, features are filtered and extracted based on acoustic emission principles and signal properties. Ultimately, a new fusion method is devised incorporating recognition results as new features into the original dataset and designing multiple layers of single algorithms based on their individual strengths to enhance the feature extraction capabilities of the algorithm. In the first layer of the fusion algorithm, CatBoost learns from the original dataset and incorporates its recognition results into the dataset. XGBoost then trains on the new dataset as the training set. Finally, the sparse output matrix generated by XGBoost is input into a logistic regression (LR) algorithm for training and prediction. The proposed method is verified by experiments on datasets and the results show that the accuracy of this method is higher than that of a single recogniser. It also performs better than current mature stacking fusion methods and mapping-based fusion methods. This fusion approach is of great significance for improving noise recognition accuracy and for innovating fusion methods.
{"title":"Noise recognition of moving parts in a sealed cavity based on the fusion of recognition results and high-dimensional mapping","authors":"Yajie Gao, Yuhang Zhang, Yuansong Liu, Chao Li, Zhigang Sun, Guotao Wang","doi":"10.1784/insi.2024.66.7.424","DOIUrl":"https://doi.org/10.1784/insi.2024.66.7.424","url":null,"abstract":"The detection and identification of noise from moving parts inside a sealed cavity is crucial for ensuring the reliability of sealed equipment. However, traditional noise recognition methods struggle to meet the stringent demands for high detection accuracy. Inspired by the idea of\u0000 ensemble learning, this paper proposes a noise recognition method that combines recognition results with high-dimensional mapping to enhance the recognition of noise. Firstly, a built noise identification experimental system is used to collect signals. Then, features are filtered and extracted\u0000 based on acoustic emission principles and signal properties. Ultimately, a new fusion method is devised incorporating recognition results as new features into the original dataset and designing multiple layers of single algorithms based on their individual strengths to enhance the feature\u0000 extraction capabilities of the algorithm. In the first layer of the fusion algorithm, CatBoost learns from the original dataset and incorporates its recognition results into the dataset. XGBoost then trains on the new dataset as the training set. Finally, the sparse output matrix generated\u0000 by XGBoost is input into a logistic regression (LR) algorithm for training and prediction. The proposed method is verified by experiments on datasets and the results show that the accuracy of this method is higher than that of a single recogniser. It also performs better than current mature\u0000 stacking fusion methods and mapping-based fusion methods. This fusion approach is of great significance for improving noise recognition accuracy and for innovating fusion methods.","PeriodicalId":506650,"journal":{"name":"Insight - Non-Destructive Testing and Condition Monitoring","volume":"42 9","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141709641","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The miniaturisation, ultra-thinness and high-density multi-layer structure of advanced microelectronic packages complicate the propagation mechanism of ultrasonic waves. In this paper, a finite element model is used to simulate ultrasonic wave propagation in flip chip packages, investigating the laws of transmission and reflection at the lamination boundaries. The acoustic field of ultrasonic transducers is simulated using MATLAB and Abaqus software. The angular spectrum method (ASM) based on the Fourier transform is adopted to more precisely reveal the distribution characteristics and attenuation relationship of near‐field ultrasonic waves. The influence of the frequency and size of the ultrasonic transducer on the propagation characteristics of ultrasonic waves is analysed. Based on an acoustic field map generated by the detection model, the waveform conversions of acoustic waves in a multi-layer structure are analysed. The results show that ultrasonic waves are mainly presented in the form of reflected and transmitted waves at the layered interface and the model with a perfectly matched layer (PML) has higher accuracy. Therefore, this method is applied to ultrasonic testing in a flip chip package, which cannot only effectively exclude interference from boundary reflection but also greatly improve the reliability of waveform conversions analysis.
{"title":"Modelling and simulation for the investigation on the ultrasonic propagation mechanism in advanced microelectronic packages","authors":"Yuan Chen, Dengxue Liu, Yuhui Fan, Zhongyang Wang, Xiang Wan, Ming Dong","doi":"10.1784/insi.2024.66.7.415","DOIUrl":"https://doi.org/10.1784/insi.2024.66.7.415","url":null,"abstract":"The miniaturisation, ultra-thinness and high-density multi-layer structure of advanced microelectronic packages complicate the propagation mechanism of ultrasonic waves. In this paper, a finite element model is used to simulate ultrasonic wave propagation in flip chip packages, investigating\u0000 the laws of transmission and reflection at the lamination boundaries. The acoustic field of ultrasonic transducers is simulated using MATLAB and Abaqus software. The angular spectrum method (ASM) based on the Fourier transform is adopted to more precisely reveal the distribution characteristics\u0000 and attenuation relationship of near‐field ultrasonic waves. The influence of the frequency and size of the ultrasonic transducer on the propagation characteristics of ultrasonic waves is analysed. Based on an acoustic field map generated by the detection model, the waveform conversions\u0000 of acoustic waves in a multi-layer structure are analysed. The results show that ultrasonic waves are mainly presented in the form of reflected and transmitted waves at the layered interface and the model with a perfectly matched layer (PML) has higher accuracy. Therefore, this method is applied\u0000 to ultrasonic testing in a flip chip package, which cannot only effectively exclude interference from boundary reflection but also greatly improve the reliability of waveform conversions analysis.","PeriodicalId":506650,"journal":{"name":"Insight - Non-Destructive Testing and Condition Monitoring","volume":"53 3","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141696760","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-07-01DOI: 10.1784/insi.2024.66.7.409
V. Arora, R. Mulaveesala, S. K. Bhambhu, S. Sharma, I. Singh, P Das, A. Sharma, G. Dua
Thermal non‐destructive testing and evaluation (NDT&E) is crucial in ensuring the quality and safety of industrial materials, components and structures. It serves as a key tool for assessing their operational reliability, thus enhancing safety in a wide range of industries. There is a growing demand for dependable, swift, remote and secure inspection and assessment techniques to detect hidden flaws, especially for sustainable solutions, which prompts adjustments in design and manufacturing standards. Hidden defects often emerge during the service life of these materials and structures due to various stress factors, potentially resulting in catastrophic failures. This study delves into an optimal and dependable experimental method for conducting fast, remote and secure inspections and assessments of carbon fibre‐reinforced polymer (CFRP) materials using infrared imaging (IRI) as part of thermal non‐destructive testing and evaluation (TNDT&E). Additionally, it examines the post-processing approach associated with this technique. This perspective also sheds light on the current state-of-the-art of infrared imaging methods employed in TNDT&E, emphasising the strengths and weaknesses in their ability to detect subsurface defects present within the material. Most of the methods discussed in previous research primarily focus on the thermal differences in specific areas of a sample using processed thermal images, even though these images come from analysing a series of images captured over time. This study highlights the latest research in thermal/infrared non‐destructive testing and evaluation, along with related post-processing techniques. It not only aims to show hidden subsurface defects through thermal differences but also provides information about how these defects change over time.
{"title":"Thermal non‐destructive testing and evaluation for inspection of carbon fibre‐reinforced polymers","authors":"V. Arora, R. Mulaveesala, S. K. Bhambhu, S. Sharma, I. Singh, P Das, A. Sharma, G. Dua","doi":"10.1784/insi.2024.66.7.409","DOIUrl":"https://doi.org/10.1784/insi.2024.66.7.409","url":null,"abstract":"Thermal non‐destructive testing and evaluation (NDT&E) is crucial in ensuring the quality and safety of industrial materials, components and structures. It serves as a key tool for assessing their operational reliability, thus enhancing safety in a wide range of industries.\u0000 There is a growing demand for dependable, swift, remote and secure inspection and assessment techniques to detect hidden flaws, especially for sustainable solutions, which prompts adjustments in design and manufacturing standards. Hidden defects often emerge during the service life of these\u0000 materials and structures due to various stress factors, potentially resulting in catastrophic failures. This study delves into an optimal and dependable experimental method for conducting fast, remote and secure inspections and assessments of carbon fibre‐reinforced polymer (CFRP) materials\u0000 using infrared imaging (IRI) as part of thermal non‐destructive testing and evaluation (TNDT&E). Additionally, it examines the post-processing approach associated with this technique. This perspective also sheds light on the current state-of-the-art of infrared imaging methods employed\u0000 in TNDT&E, emphasising the strengths and weaknesses in their ability to detect subsurface defects present within the material. Most of the methods discussed in previous research primarily focus on the thermal differences in specific areas of a sample using processed thermal images, even\u0000 though these images come from analysing a series of images captured over time. This study highlights the latest research in thermal/infrared non‐destructive testing and evaluation, along with related post-processing techniques. It not only aims to show hidden subsurface defects\u0000 through thermal differences but also provides information about how these defects change over time.","PeriodicalId":506650,"journal":{"name":"Insight - Non-Destructive Testing and Condition Monitoring","volume":"13 6","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141711176","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-07-01DOI: 10.1784/insi.2024.66.7.400
S. Kenderian, T. Case, P. M. Adams, A. Gregorian
To simulate cracks, a series of slits with depths ranging from 17.2% to 40% of 301 stainless steel (SS) plate thickness are fabricated for this study. The examination includes a microfocus X-ray tube and a digital detector array (DDA) with 75 ??m pixel pitch. The basis of this study is to determine the effect of detector scintillator blur, noise and pixelation on the contrast of the resultant radiographic image and, ultimately, on the detectability of a crack. Many crack detectability studies are performed on notched plates rather than true cracks. As the notch width becomes smaller than a pixel, the grey value of the image it generates will be averaged over the entire footprint of the pixel. Therefore, the contrast with neighbouring (no notch) pixels is reduced. This study develops a relationship between the material property, detector parameters, notch geometry and expected contrast that would result from the radiograph image. This relationship can be used as a transfer function towards evaluating the resultant contrast from a notch and extrapolating the results towards a hairline crack.
{"title":"Contrast-based notch-to-crack transfer function for digital radiography","authors":"S. Kenderian, T. Case, P. M. Adams, A. Gregorian","doi":"10.1784/insi.2024.66.7.400","DOIUrl":"https://doi.org/10.1784/insi.2024.66.7.400","url":null,"abstract":"To simulate cracks, a series of slits with depths ranging from 17.2% to 40% of 301 stainless steel (SS) plate thickness are fabricated for this study. The examination includes a microfocus X-ray tube and a digital detector array (DDA) with 75 ??m pixel pitch. The basis\u0000 of this study is to determine the effect of detector scintillator blur, noise and pixelation on the contrast of the resultant radiographic image and, ultimately, on the detectability of a crack. Many crack detectability studies are performed on notched plates rather than true cracks. As the\u0000 notch width becomes smaller than a pixel, the grey value of the image it generates will be averaged over the entire footprint of the pixel. Therefore, the contrast with neighbouring (no notch) pixels is reduced. This study develops a relationship between the material property, detector parameters,\u0000 notch geometry and expected contrast that would result from the radiograph image. This relationship can be used as a transfer function towards evaluating the resultant contrast from a notch and extrapolating the results towards a hairline crack.","PeriodicalId":506650,"journal":{"name":"Insight - Non-Destructive Testing and Condition Monitoring","volume":"3 11","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141696404","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The sparse total focusing method (TFM) has been shown to enhance the computational efficacy of ultrasound imaging but the image quality of ultrasound regrettably deteriorates with an increase in the sparsity rate of array elements. Deep learning has made remarkable advancements in image processing and cycle-consistent generative adversarial networks (CycleGANs) have been extensively employed to reconstruct diverse image categories. However, due to the incomplete extraction of image feature information by the generator and discriminator in a CycleGAN, high-quality sparse TFM images cannot be directly reconstructed using CycleGANs. There is also a risk of losing crucial feature information related to minor defects. As a result, this paper modifies the generator and discriminator in the CycleGAN to construct a new relativistic discriminator and coordinate attention CycleGAN (RCACycleGAN) model, which enables high-precision reconstruction of sparse TFM images. The addition of the coordinate attention module to the CycleGAN enhances the defective feature representation by fully considering the channel and spatial correlation between regions and using the fusion of spatially perceived feature maps in different directions. It solves the problem of easy loss of defective key feature information. The relativistic discriminator replaces the PatchGAN discriminator in the CycleGAN and evaluates the quality of both real and sparse TFM reconstructed images to ensure a relative image quality evaluation. This process solves the problem of unstable image quality of the sparse TFM reconstructed image. Experimental results demonstrate that RCACycleGAN can stably reconstruct sparse TFM images even in small sample dataset scenarios. The proposed network model reconstructs images with better accuracy, including in terms of structural similarity, defect roundness and area, and has a shorter training time than several existing network models.
{"title":"A novel RCACycleGAN model is proposed for the high-precision reconstruction of sparse TFM images","authors":"Zhouteng Liu, Liming Li, Wenfa Zhu, Yanxun Xiang, Guopeng Fan, Hui Zhang","doi":"10.1784/insi.2024.66.5.272","DOIUrl":"https://doi.org/10.1784/insi.2024.66.5.272","url":null,"abstract":"The sparse total focusing method (TFM) has been shown to enhance the computational efficacy of ultrasound imaging but the image quality of ultrasound regrettably deteriorates with an increase in the sparsity rate of array elements. Deep learning has made remarkable advancements in image\u0000 processing and cycle-consistent generative adversarial networks (CycleGANs) have been extensively employed to reconstruct diverse image categories. However, due to the incomplete extraction of image feature information by the generator and discriminator in a CycleGAN, high-quality sparse TFM\u0000 images cannot be directly reconstructed using CycleGANs. There is also a risk of losing crucial feature information related to minor defects. As a result, this paper modifies the generator and discriminator in the CycleGAN to construct a new relativistic discriminator and coordinate attention\u0000 CycleGAN (RCACycleGAN) model, which enables high-precision reconstruction of sparse TFM images. The addition of the coordinate attention module to the CycleGAN enhances the defective feature representation by fully considering the channel and spatial correlation between regions and using the\u0000 fusion of spatially perceived feature maps in different directions. It solves the problem of easy loss of defective key feature information. The relativistic discriminator replaces the PatchGAN discriminator in the CycleGAN and evaluates the quality of both real and sparse TFM reconstructed\u0000 images to ensure a relative image quality evaluation. This process solves the problem of unstable image quality of the sparse TFM reconstructed image. Experimental results demonstrate that RCACycleGAN can stably reconstruct sparse TFM images even in small sample dataset scenarios. The proposed\u0000 network model reconstructs images with better accuracy, including in terms of structural similarity, defect roundness and area, and has a shorter training time than several existing network models.","PeriodicalId":506650,"journal":{"name":"Insight - Non-Destructive Testing and Condition Monitoring","volume":"2020 41","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141026743","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-05-01DOI: 10.1784/insi.2024.66.5.281
Tang Yuqi, Guo Keke, Chonglin Zhao, Wenlian Wang
Ensuring sufficient preload of bolted connections is crucial for maintaining the safety of railway operations. Accidents resulting from loose bolts caused by inadequate preload are a recurring concern. Traditional bolt preload detection methods suffer from reliability issues and bulky equipment requirements. To address these challenges, this study proposes a novel solution utilising passive wireless measurement systems. However, the communication effectiveness of these systems can be significantly impacted by the surrounding metal environment. This research introduces an intelligent clover washer design to mitigate the adverse effects of eddy currents induced by the metal environment. The design incorporates a new clover antenna and washer structure, effectively reducing the influence of the metal environment and improving the communication quality. The proposed design has undergone comprehensive prototyping, simulation and experimental verification. The results demonstrate a significant improvement over the traditional circular antenna and washer combination under similar conditions. Specifically, the sensing distance of the new clover washer is enhanced by 60% and the stable communication distance is improved by 75%. The experimental results highlight the ability of the new clover smart washer to generate a stronger spatial magnetic field and exhibit reduced susceptibility to the metal washer, thereby enhancing communication effectiveness.
{"title":"Bolt loosening monitoring with passive wireless-based smart washers","authors":"Tang Yuqi, Guo Keke, Chonglin Zhao, Wenlian Wang","doi":"10.1784/insi.2024.66.5.281","DOIUrl":"https://doi.org/10.1784/insi.2024.66.5.281","url":null,"abstract":"Ensuring sufficient preload of bolted connections is crucial for maintaining the safety of railway operations. Accidents resulting from loose bolts caused by inadequate preload are a recurring concern. Traditional bolt preload detection methods suffer from reliability issues and bulky\u0000 equipment requirements. To address these challenges, this study proposes a novel solution utilising passive wireless measurement systems. However, the communication effectiveness of these systems can be significantly impacted by the surrounding metal environment. This research introduces an\u0000 intelligent clover washer design to mitigate the adverse effects of eddy currents induced by the metal environment. The design incorporates a new clover antenna and washer structure, effectively reducing the influence of the metal environment and improving the communication quality. The proposed\u0000 design has undergone comprehensive prototyping, simulation and experimental verification. The results demonstrate a significant improvement over the traditional circular antenna and washer combination under similar conditions. Specifically, the sensing distance of the new clover washer is\u0000 enhanced by 60% and the stable communication distance is improved by 75%. The experimental results highlight the ability of the new clover smart washer to generate a stronger spatial magnetic field and exhibit reduced susceptibility to the metal washer, thereby enhancing communication effectiveness.","PeriodicalId":506650,"journal":{"name":"Insight - Non-Destructive Testing and Condition Monitoring","volume":"12 7","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141051442","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-05-01DOI: 10.1784/insi.2024.66.5.294
Zhenheng Xu, Zhong Liu, Bing Tian, Q. Lv, Hu Liu
Existing deep learning (DL) algorithms are based on a large amount of training data and they face challenges in effectively extracting fault features when dealing with few-shot fault diagnoses. Model-agnostic meta-learning (MAML) also faces some challenges, including the limited capability of the basic convolutional neural network (CNN) with a single convolutional kernel to extract fault features comprehensively, as well as the instability of model training due to the inner and outer double-layer loops. To address these issues, this paper presents a multi-step loss meta-learning method based on multi-scale feature extraction (MFEML). Firstly, an improved multi-scale feature extraction module (IMFEM) is designed to solve the problem of the insufficient feature extraction capability of the CNN. Secondly, the multi-step loss is used to reconstruct the meta-loss to address the issue of MAML training instability. Finally, the experimental results of two datasets demonstrate the effectiveness of the MFEML.
{"title":"A multi-step loss meta-learning method based on multi-scale feature extraction for few-shot fault diagnosis","authors":"Zhenheng Xu, Zhong Liu, Bing Tian, Q. Lv, Hu Liu","doi":"10.1784/insi.2024.66.5.294","DOIUrl":"https://doi.org/10.1784/insi.2024.66.5.294","url":null,"abstract":"Existing deep learning (DL) algorithms are based on a large amount of training data and they face challenges in effectively extracting fault features when dealing with few-shot fault diagnoses. Model-agnostic meta-learning (MAML) also faces some challenges, including the limited capability\u0000 of the basic convolutional neural network (CNN) with a single convolutional kernel to extract fault features comprehensively, as well as the instability of model training due to the inner and outer double-layer loops. To address these issues, this paper presents a multi-step loss meta-learning\u0000 method based on multi-scale feature extraction (MFEML). Firstly, an improved multi-scale feature extraction module (IMFEM) is designed to solve the problem of the insufficient feature extraction capability of the CNN. Secondly, the multi-step loss is used to reconstruct the meta-loss to address\u0000 the issue of MAML training instability. Finally, the experimental results of two datasets demonstrate the effectiveness of the MFEML.","PeriodicalId":506650,"journal":{"name":"Insight - Non-Destructive Testing and Condition Monitoring","volume":"23 16","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141048882","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Arc magnetic field analysis is a valuable approach for assessing the stability of the arc welding process, yet existing methods lack the ability to effectively quantify the disorder within the process. Through an investigation into the characteristics of the arc magnetic field signal, it was observed that the occurrence of low-frequency random fluctuations in arc magnetic field power, induced by unstable factors such as bubbles or short circuits, contributed to increased complexity and randomness in the arc magnetic field signals. To visualise the arc magnetic field signals in a time-frequency domain, a spectrogram was employed, revealing a strong correlation between the distribution of maximum power spectral density (PSD) in the spectrogram and the stability of the arc welding process. Furthermore, a novel method based on sample entropy was introduced to provide a quantitative measure of this relationship. A comprehensive quantitative assessment indicator called arc magnetic field sample entropy (AMFSE) was proposed. This indicator effectively mitigates the influence of varying parameters on the quantitative results, enabling a more accurate and consistent representation of the stability of the arc welding process. The proposed method was validated through testing, yielding an accuracy rate exceeding 90%.
{"title":"Sample entropy-based quantitative assessment of the arc magnetic field spectrum for improved arc welding quality","authors":"Senming Zhong, Ping Yao, Yunyi Huang, Xiaojun Wang, Jianbin Luo, Shunjian Liang","doi":"10.1784/insi.2024.66.5.287","DOIUrl":"https://doi.org/10.1784/insi.2024.66.5.287","url":null,"abstract":"Arc magnetic field analysis is a valuable approach for assessing the stability of the arc welding process, yet existing methods lack the ability to effectively quantify the disorder within the process. Through an investigation into the characteristics of the arc magnetic field signal,\u0000 it was observed that the occurrence of low-frequency random fluctuations in arc magnetic field power, induced by unstable factors such as bubbles or short circuits, contributed to increased complexity and randomness in the arc magnetic field signals. To visualise the arc magnetic field signals\u0000 in a time-frequency domain, a spectrogram was employed, revealing a strong correlation between the distribution of maximum power spectral density (PSD) in the spectrogram and the stability of the arc welding process. Furthermore, a novel method based on sample entropy was introduced to provide\u0000 a quantitative measure of this relationship. A comprehensive quantitative assessment indicator called arc magnetic field sample entropy (AMFSE) was proposed. This indicator effectively mitigates the influence of varying parameters on the quantitative results, enabling a more accurate and consistent\u0000 representation of the stability of the arc welding process. The proposed method was validated through testing, yielding an accuracy rate exceeding 90%.","PeriodicalId":506650,"journal":{"name":"Insight - Non-Destructive Testing and Condition Monitoring","volume":"2018 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141027013","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Alkaline battery defect detection is crucial for ensuring product quality and providing diagnostic feedback. Recently, high-performance deep learning algorithms have been introduced to recognise defects in alkaline batteries. However, the majority of deep learning-based methods overlook the significant data imbalance issues in alkaline battery electrode images, potentially resulting in performance degradation. Therefore, a voting-based recognition algorithm containing three parts is proposed in this study. Firstly, a resampling and training method is developed to provide richer information and stronger constraints. Secondly, a weak classification framework based on an improved convolutional neural network (CNN) is designed to provide fine-grained category representations. Finally, a voting-based prediction approach is proposed to improve accuracy and obtain the final results. Visualisation results demonstrate that the proposed algorithm has a stronger clustering ability and uses voting-based prediction to improve performance. Comparison research shows that the proposed method significantly improves the recall of minority classes and the precision of majority classes and reaches a state-of-the-art F1 score of 0.982 for alkaline battery defect recognition, which is 0.041 higher than the basic CNN model.
碱性电池缺陷检测对于确保产品质量和提供诊断反馈至关重要。最近，人们引入了高性能深度学习算法来识别碱性电池中的缺陷。然而，大多数基于深度学习的方法都忽略了碱性电池电极图像中严重的数据不平衡问题，可能导致性能下降。因此，本研究提出了一种基于投票的识别算法，包含三个部分。首先，开发了一种重采样和训练方法，以提供更丰富的信息和更强的约束。其次，设计了一个基于改进的卷积神经网络（CNN）的弱分类框架，以提供细粒度的类别表示。最后，提出了一种基于投票的预测方法，以提高准确性并获得最终结果。可视化结果表明，所提出的算法具有更强的聚类能力，并利用基于投票的预测来提高性能。对比研究表明，所提出的方法显著提高了少数类别的召回率和多数类别的精确度，在碱性电池缺陷识别方面达到了最先进的 F1 分数 0.982，比基本 CNN 模型高出 0.041。
{"title":"Resolving data imbalance in alkaline battery defect detection: a voting-based deep learning approach","authors":"Zhenying Xu, Bangguo Han, Liling Han, Yucheng Tao, Yun Wang, Ying-Jun Lei","doi":"10.1784/insi.2024.66.5.305","DOIUrl":"https://doi.org/10.1784/insi.2024.66.5.305","url":null,"abstract":"Alkaline battery defect detection is crucial for ensuring product quality and providing diagnostic feedback. Recently, high-performance deep learning algorithms have been introduced to recognise defects in alkaline batteries. However, the majority of deep learning-based methods overlook\u0000 the significant data imbalance issues in alkaline battery electrode images, potentially resulting in performance degradation. Therefore, a voting-based recognition algorithm containing three parts is proposed in this study. Firstly, a resampling and training method is developed to provide\u0000 richer information and stronger constraints. Secondly, a weak classification framework based on an improved convolutional neural network (CNN) is designed to provide fine-grained category representations. Finally, a voting-based prediction approach is proposed to improve accuracy and obtain\u0000 the final results. Visualisation results demonstrate that the proposed algorithm has a stronger clustering ability and uses voting-based prediction to improve performance. Comparison research shows that the proposed method significantly improves the recall of minority classes and the precision\u0000 of majority classes and reaches a state-of-the-art F1 score of 0.982 for alkaline battery defect recognition, which is 0.041 higher than the basic CNN model.","PeriodicalId":506650,"journal":{"name":"Insight - Non-Destructive Testing and Condition Monitoring","volume":"48 2","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141053092","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：