Ndt & E International最新文献_第4页

Linear and nonlinear ultrasonic/acoustic characterization of thermally sprayed nickel coatings 热喷涂镍涂层的线性和非线性超声/声学特性

IF 4.5 2区材料科学 Q1 MATERIALS SCIENCE, CHARACTERIZATION & TESTING

Ndt & E International

Pub Date : 2025-11-21 DOI: 10.1016/j.ndteint.2025.103609

Jacob H. Brody , Prabhakaran Manogharan , Emma E. Peleg , Nathan W. Moore , Alper Erturk

Thermally sprayed coatings are extensively used in the energy and aerospace sectors to protect surfaces against wear, oxidation, corrosion, and for thermal insulation; therefore their characterization is vital to preserving and protecting key infrastructural capabilities. In this study, we conduct experiments to examine the feasibility of using nonlinear ultrasonic/acoustic techniques such as nonlinear resonance acoustic spectroscopy (NRAS) and second harmonic generation (SHG) to characterize thermally sprayed nickel coatings. Nonlinear elastic parameters have been proven to be highly sensitive to microstructural features in materials. This work characterizes the

α

and

β

nonlinearity parameters which are related to the degree of micro-damages (interlamellar gap/void structures) intrinsic to sprayed coatings. In this work, thermally sprayed nickel coatings are sprayed onto stainless steel substrates at two particle sizes (Ni 101 - fine powder and Ni 969 - coarse powder) and spray distances (101.6 and 203.2 mm stand-off distance). The sprayed coatings are separated from their substrates to determine if the nonlinearity parameters can effectively measure microstructural changes in these samples. The nonlinearity parameters measured independently from two different nonlinear ultrasonic/acoustic techniques show an increasing trend with increasing particle sizes and spray distances. Furthermore, NRAS measurements of coatings with and without the substrate indicate that the nonlinearity of thermally sprayed coatings can still be characterized on the thicker substrate. In addition to estimating the nonlinear parameters, porosity, hardness, wave speed and attenuation are also reported. Our results suggest that these nonlinear ultrasonic/acoustic techniques can be potentially used as an inspection and quality control tool for thermal spray coatings.

热喷涂涂层广泛用于能源和航空航天领域，以保护表面免受磨损、氧化、腐蚀和隔热；因此，它们的特征对于保存和保护关键的基础设施功能至关重要。在这项研究中，我们进行了实验，以检验使用非线性超声/声学技术，如非线性共振声光谱（NRAS）和二次谐波产生（SHG）来表征热喷涂镍涂层的可行性。非线性弹性参数对材料的微观结构特征非常敏感。本工作表征了与喷涂涂层固有微损伤（层间间隙/空洞结构）程度相关的α和β非线性参数。在这项工作中，热喷涂镍涂层以两种粒径（Ni 101 -细粉和Ni 969 -粗粉）和喷涂距离（101.6和203.2 mm的隔离距离）喷涂到不锈钢基体上。将喷涂后的涂层与基体分离，以确定非线性参数是否能有效测量样品的微观结构变化。两种不同的非线性超声/声学技术所测得的非线性参数随粒径的增大和喷雾距离的增大而增大。此外，有基材和没有基材的涂层的NRAS测量表明，在较厚的基材上仍然可以表征热喷涂涂层的非线性。除了估计非线性参数外，还报道了孔隙率，硬度，波速和衰减。我们的研究结果表明，这些非线性超声/声学技术可以作为热喷涂涂层的检测和质量控制工具。

{"title":"Linear and nonlinear ultrasonic/acoustic characterization of thermally sprayed nickel coatings","authors":"Jacob H. Brody , Prabhakaran Manogharan , Emma E. Peleg , Nathan W. Moore , Alper Erturk","doi":"10.1016/j.ndteint.2025.103609","DOIUrl":"10.1016/j.ndteint.2025.103609","url":null,"abstract":"<div><div>Thermally sprayed coatings are extensively used in the energy and aerospace sectors to protect surfaces against wear, oxidation, corrosion, and for thermal insulation; therefore their characterization is vital to preserving and protecting key infrastructural capabilities. In this study, we conduct experiments to examine the feasibility of using nonlinear ultrasonic/acoustic techniques such as nonlinear resonance acoustic spectroscopy (NRAS) and second harmonic generation (SHG) to characterize thermally sprayed nickel coatings. Nonlinear elastic parameters have been proven to be highly sensitive to microstructural features in materials. This work characterizes the <span><math><mi>α</mi></math></span> and <span><math><mi>β</mi></math></span> nonlinearity parameters which are related to the degree of micro-damages (interlamellar gap/void structures) intrinsic to sprayed coatings. In this work, thermally sprayed nickel coatings are sprayed onto stainless steel substrates at two particle sizes (Ni 101 - fine powder and Ni 969 - coarse powder) and spray distances (101.6 and 203.2 mm stand-off distance). The sprayed coatings are separated from their substrates to determine if the nonlinearity parameters can effectively measure microstructural changes in these samples. The nonlinearity parameters measured independently from two different nonlinear ultrasonic/acoustic techniques show an increasing trend with increasing particle sizes and spray distances. Furthermore, NRAS measurements of coatings with and without the substrate indicate that the nonlinearity of thermally sprayed coatings can still be characterized on the thicker substrate. In addition to estimating the nonlinear parameters, porosity, hardness, wave speed and attenuation are also reported. Our results suggest that these nonlinear ultrasonic/acoustic techniques can be potentially used as an inspection and quality control tool for thermal spray coatings.</div></div>","PeriodicalId":18868,"journal":{"name":"Ndt & E International","volume":"159 ","pages":"Article 103609"},"PeriodicalIF":4.5,"publicationDate":"2025-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145622284","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Multiporous-cascaded coil based high lift-off and dynamic electromagnetic thermography of rail defects inspection 基于多孔级联线圈的高升力和动态电磁热成像钢轨缺陷检测

IF 4.5 2区材料科学 Q1 MATERIALS SCIENCE, CHARACTERIZATION & TESTING

Ndt & E International

Pub Date : 2025-11-21 DOI: 10.1016/j.ndteint.2025.103599

Geng Yang , Haoran Li , Bin Gao , Xiaolong Lu , Junhong Qi , Dong Liu , Guiyun Tian , Xiaojie Xue , Xingcai Liu

Lift-off variability poses a significant challenge in electromagnetic thermography (ET) for detecting surface defects on rails, particularly under dynamic scanning conditions. To overcome this limitation, we propose a lift-off tolerant ET system based on a multiporous-cascaded excitation coil. This coil design enhances electromagnetic excitation and expands the effective heating footprint on the rail surface. Finite element simulations under various lift-off distances evaluate the coil's thermal response and excitation coverage. Compared to a conventional dual-turn elliptical coil, the proposed structure demonstrates superior heating intensity and spatial uniformity at large lift-off distances. Experimental validation at a scanning speed of 1 m/s with a 30 mm lift-off confirms the system's capability to detect multiple defect types with enhanced thermal contrast and signal robustness. These results validate the practical feasibility of the proposed system for dynamic rail inspection under large lift-off conditions.

在电磁热成像（ET）检测轨道表面缺陷时，特别是在动态扫描条件下，升空变异性对其提出了重大挑战。为了克服这一限制，我们提出了一种基于多孔级联励磁线圈的容升ET系统。这种线圈设计增强了电磁激励，扩大了轨道表面的有效加热足迹。在不同升力距离下的有限元模拟评估了线圈的热响应和激励覆盖。与传统的双匝椭圆线圈相比，所提出的结构在大起飞距离下具有优越的加热强度和空间均匀性。在扫描速度为1m /s、上升距离为30mm的情况下，实验验证了该系统能够通过增强热对比度和信号鲁棒性来检测多种缺陷类型。这些结果验证了该系统在大升离条件下动态钢轨检测的实际可行性。

{"title":"Multiporous-cascaded coil based high lift-off and dynamic electromagnetic thermography of rail defects inspection","authors":"Geng Yang , Haoran Li , Bin Gao , Xiaolong Lu , Junhong Qi , Dong Liu , Guiyun Tian , Xiaojie Xue , Xingcai Liu","doi":"10.1016/j.ndteint.2025.103599","DOIUrl":"10.1016/j.ndteint.2025.103599","url":null,"abstract":"<div><div>Lift-off variability poses a significant challenge in electromagnetic thermography (ET) for detecting surface defects on rails, particularly under dynamic scanning conditions. To overcome this limitation, we propose a lift-off tolerant ET system based on a multiporous-cascaded excitation coil. This coil design enhances electromagnetic excitation and expands the effective heating footprint on the rail surface. Finite element simulations under various lift-off distances evaluate the coil's thermal response and excitation coverage. Compared to a conventional dual-turn elliptical coil, the proposed structure demonstrates superior heating intensity and spatial uniformity at large lift-off distances. Experimental validation at a scanning speed of 1 m/s with a 30 mm lift-off confirms the system's capability to detect multiple defect types with enhanced thermal contrast and signal robustness. These results validate the practical feasibility of the proposed system for dynamic rail inspection under large lift-off conditions.</div></div>","PeriodicalId":18868,"journal":{"name":"Ndt & E International","volume":"159 ","pages":"Article 103599"},"PeriodicalIF":4.5,"publicationDate":"2025-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145622285","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A contribution by gradient explainability method for 1D-CNNs on ultrasonic data 梯度可解释性方法对超声数据一维cnn的贡献

IF 4.5 2区材料科学 Q1 MATERIALS SCIENCE, CHARACTERIZATION & TESTING

Ndt & E International

Pub Date : 2025-11-19 DOI: 10.1016/j.ndteint.2025.103592

Yuyang Liu, Sergio Cantero-Chinchilla, Anthony J. Croxford

This paper proposes the Contribution by Gradients (C-grad) method for interpreting neural network models applied to regression problems on ultrasonic data. As machine learning applications continue to expand in Non-Destructive Testing (NDT), the “black-box” nature of neural networks raises concerns about the consistency and interpretability of AI-generated solutions in industrial applications. The C-grad method addresses these challenges by quantifying input contributions through their rate of change across layers, providing detailed layer-by-layer insights into model decisions. Unlike traditional gradient algorithms focused solely on final outputs, C-grad backward propagates activation changes from intermediate layers to input arrays, enabling a more detailed breakdown of feature importance at each layer and enhancing interpretability. Applied to 1-dimensional convolutional neural networks (1D-CNNs) for ultrasonic A-scan time-traces data in corrosion profiling, the method shows superior stability and consistency over the traditional Gradient-based Class Activation Map method (Grad-CAM). By combining explanation infidelity testing with a peak-to-peak distance metric that correlates model explanations with physical echo features, C-grad offers a comprehensive framework for assessing model robustness to corrupted inputs. C-grad has proven successful in identifying crucial physical features in the A-scan time-traces for the 1D-CNNs trained on two corrosion profiling metrics: mean material thickness and roughness. These case studies under C-grad investigation provide detailed insights into model trustworthiness and guide architecture optimisation in ML-driven NDT.

本文提出了梯度贡献（C-grad）方法来解释应用于超声数据回归问题的神经网络模型。随着机器学习应用在无损检测（NDT）领域的不断扩展，神经网络的“黑箱”性质引发了人们对工业应用中人工智能生成解决方案的一致性和可解释性的担忧。C-grad方法通过通过各层的变化率量化输入贡献来解决这些挑战，为模型决策提供详细的逐层洞察。与传统的梯度算法只关注最终输出不同，C-grad将激活变化从中间层向后传播到输入数组，从而能够更详细地分解每一层的特征重要性，并增强可解释性。将该方法应用于一维卷积神经网络（1d - cnn）的腐蚀剖面超声a扫描时间轨迹数据中，与传统的基于梯度的类别激活图方法（Grad-CAM）相比，该方法具有更好的稳定性和一致性。通过将解释不忠测试与峰对峰距离度量（将模型解释与物理回波特征相关联）相结合，C-grad为评估模型对损坏输入的鲁棒性提供了一个全面的框架。C-grad已经被证明成功地识别了d - cnn在两个腐蚀剖面指标（平均材料厚度和粗糙度）上训练的a扫描时间轨迹中的关键物理特征。C-grad调查下的这些案例研究提供了对模型可信度的详细见解，并指导了机器学习驱动无损检测的架构优化。

{"title":"A contribution by gradient explainability method for 1D-CNNs on ultrasonic data","authors":"Yuyang Liu, Sergio Cantero-Chinchilla, Anthony J. Croxford","doi":"10.1016/j.ndteint.2025.103592","DOIUrl":"10.1016/j.ndteint.2025.103592","url":null,"abstract":"<div><div>This paper proposes the Contribution by Gradients (C-grad) method for interpreting neural network models applied to regression problems on ultrasonic data. As machine learning applications continue to expand in Non-Destructive Testing (NDT), the “black-box” nature of neural networks raises concerns about the consistency and interpretability of AI-generated solutions in industrial applications. The C-grad method addresses these challenges by quantifying input contributions through their rate of change across layers, providing detailed layer-by-layer insights into model decisions. Unlike traditional gradient algorithms focused solely on final outputs, C-grad backward propagates activation changes from intermediate layers to input arrays, enabling a more detailed breakdown of feature importance at each layer and enhancing interpretability. Applied to 1-dimensional convolutional neural networks (1D-CNNs) for ultrasonic A-scan time-traces data in corrosion profiling, the method shows superior stability and consistency over the traditional Gradient-based Class Activation Map method (Grad-CAM). By combining explanation infidelity testing with a peak-to-peak distance metric that correlates model explanations with physical echo features, C-grad offers a comprehensive framework for assessing model robustness to corrupted inputs. C-grad has proven successful in identifying crucial physical features in the A-scan time-traces for the 1D-CNNs trained on two corrosion profiling metrics: mean material thickness and roughness. These case studies under C-grad investigation provide detailed insights into model trustworthiness and guide architecture optimisation in ML-driven NDT.</div></div>","PeriodicalId":18868,"journal":{"name":"Ndt & E International","volume":"159 ","pages":"Article 103592"},"PeriodicalIF":4.5,"publicationDate":"2025-11-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145555258","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A comprehensive investigation of flexible and multi-dimensional simulation-based PoD analysis 基于柔性和多维仿真的PoD分析的综合研究

IF 4.5 2区材料科学 Q1 MATERIALS SCIENCE, CHARACTERIZATION & TESTING

Ndt & E International

Pub Date : 2025-11-17 DOI: 10.1016/j.ndteint.2025.103596

Nachman Malkiel , Anthony J. Croxford , Paul D. Wilcox

This study presents a comprehensive investigation into the simulation-based reliability assessment of NDE inspections, thereby enabling efficient, flexible, and multi-dimensional Probability of Detection (PoD) analysis. Building on existing literature and incorporating several new aspects, a complete framework is proposed and its various components are thoroughly discussed. The framework integrates validated, physics-based computer simulations with Kriging meta-modeling and Monte-Carlo integration, rigorously addressing challenging, multi-dimensional parameter spaces and uncertainty quantification. It constitutes an alternative to the traditional

\hat{a}

-versus-

a

method, which relies on oversimplifying statistical assumptions and is limited to one-dimensional PoD curves. While the Kriging-based approach provides a foundation for the more general stochastic case, the current research focuses on the simpler deterministic case as an initial step. An ultrasonic inspection example demonstrates this approach. First, a Finite Element model is validated against experimental results and its limitations are discussed. Then, the suggested framework is used to generate Probability of Detection (PoD) functions with uncertainty bounds highlighting its enhanced flexibility. Results are compared to ground-truth simulations, demonstrating the reliability of the analysis compared to traditional methods. Overall, the study takes a step forward towards the generation of a more efficient, robust, and adaptable NDE assessment, facilitating enhanced inspection design and confidence in reliability evaluations.

本研究对基于仿真的无损检测可靠性评估进行了全面的研究，从而实现了高效、灵活和多维的检测概率（PoD）分析。在现有文献的基础上，结合几个新的方面，提出了一个完整的框架，并对其各个组成部分进行了深入的讨论。该框架将经过验证的基于物理的计算机模拟与克里格元建模和蒙特卡罗集成相结合，严格解决具有挑战性的多维参数空间和不确定性量化问题。它构成了传统的a -对-a方法的替代方案，后者依赖于过度简化的统计假设，并且仅限于一维PoD曲线。虽然基于克里格的方法为更一般的随机情况提供了基础，但目前的研究主要集中在更简单的确定性情况下作为初始步骤。一个超声波检查的例子说明了这种方法。首先，根据实验结果对有限元模型进行了验证，并对其局限性进行了讨论。然后，使用该框架生成具有不确定性边界的检测概率（PoD）函数，以突出其增强的灵活性。结果与地面真值模拟进行了比较，证明了与传统方法相比分析的可靠性。总的来说，该研究朝着产生更有效、更稳健、适应性更强的无损检测评估迈出了一步，促进了检测设计的增强和可靠性评估的信心。

{"title":"A comprehensive investigation of flexible and multi-dimensional simulation-based PoD analysis","authors":"Nachman Malkiel , Anthony J. Croxford , Paul D. Wilcox","doi":"10.1016/j.ndteint.2025.103596","DOIUrl":"10.1016/j.ndteint.2025.103596","url":null,"abstract":"<div><div>This study presents a comprehensive investigation into the simulation-based reliability assessment of NDE inspections, thereby enabling efficient, flexible, and multi-dimensional Probability of Detection (PoD) analysis. Building on existing literature and incorporating several new aspects, a complete framework is proposed and its various components are thoroughly discussed. The framework integrates validated, physics-based computer simulations with Kriging meta-modeling and Monte-Carlo integration, rigorously addressing challenging, multi-dimensional parameter spaces and uncertainty quantification. It constitutes an alternative to the traditional <span><math><mrow><mover><mi>a</mi><mo>ˆ</mo></mover></mrow></math></span>-versus-<span><math><mrow><mi>a</mi></mrow></math></span> method, which relies on oversimplifying statistical assumptions and is limited to one-dimensional PoD curves. While the Kriging-based approach provides a foundation for the more general stochastic case, the current research focuses on the simpler deterministic case as an initial step. An ultrasonic inspection example demonstrates this approach. First, a Finite Element model is validated against experimental results and its limitations are discussed. Then, the suggested framework is used to generate Probability of Detection (PoD) functions with uncertainty bounds highlighting its enhanced flexibility. Results are compared to ground-truth simulations, demonstrating the reliability of the analysis compared to traditional methods. Overall, the study takes a step forward towards the generation of a more efficient, robust, and adaptable NDE assessment, facilitating enhanced inspection design and confidence in reliability evaluations.</div></div>","PeriodicalId":18868,"journal":{"name":"Ndt & E International","volume":"159 ","pages":"Article 103596"},"PeriodicalIF":4.5,"publicationDate":"2025-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145578739","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A steady-state wavefield polarity-aware U-Net for ultrasonic guided waves depth-resolved imaging 用于超声导波深度分辨成像的稳态波场极性感知U-Net

IF 4.5 2区材料科学 Q1 MATERIALS SCIENCE, CHARACTERIZATION & TESTING

Ndt & E International

Pub Date : 2025-11-15 DOI: 10.1016/j.ndteint.2025.103597

Xuan Li , Lishuai Liu , Jiang Lu , Yuncheng Zhang , Dazhi Cong , Yanxun Xiang

The laser scanning full wavefield capture (LSFWC) technique enables rapid acquisition of two‐dimensional maps of local velocity variations of ultrasonic guided waves (UGWs), thereby facilitating quantitative spatial visualization of damage in thin‐walled structures. However, this technique faces a trade-off among spatial resolution, the accuracy of damage estimation, and detection efficiency: the spatial sampling rate influences the limit of velocity estimation, whereas increasing the sampling rate renders the full wavefield data acquisition extremely time-consuming. To tackle these difficulties, we propose a U-Net based imaging approach to reconstruct high-quality depth-resolved images from steady-state wavefield polarity (SSWP) of spatial sub-Nyquist sampled UGWs. The developed SSWP-aware U-Net can automatically segment the wavefield, extract local phase information, and perform encoding and decoding to establish approximate nonlinear mapping between wavefield and spatial distribution of UGWs velocity variations. The selection of SSWP as input features greatly accelerates the acquisition of the training dataset, and further enhances the robustness and generalization ability of U-Net trained on simulated dataset. We experimentally compared the proposed imaging method to wavenumber estimation methods. The results demonstrated that the proposed SSWP-aware U-Net is significantly superior to traditional methods in terms of image reconstruction quality and defect depth resolution especially with local spatial sub-Nyquist sampling.

激光扫描全波场捕获（LSFWC）技术能够快速获取超声导波（ugw）局部速度变化的二维地图，从而促进薄壁结构损伤的定量空间可视化。然而，该技术面临着空间分辨率、损伤估计精度和检测效率之间的权衡：空间采样率影响速度估计的极限，而增加采样率使得全波场数据采集非常耗时。为了解决这些困难，我们提出了一种基于U-Net的成像方法，从空间亚奈奎斯特采样的ugw的稳态波场极性（SSWP）重建高质量的深度分辨率图像。开发的感知sswp的U-Net可以自动分割波场，提取局部相位信息，进行编解码，建立波场与ugw速度变化空间分布的近似非线性映射关系。选择SSWP作为输入特征大大加快了训练数据集的获取速度，进一步增强了U-Net在模拟数据集上训练的鲁棒性和泛化能力。我们实验比较了所提出的成像方法和波数估计方法。结果表明，基于sswp感知的U-Net在图像重建质量和缺陷深度分辨率方面明显优于传统方法，特别是在局部空间亚奈奎斯特采样方面。

{"title":"A steady-state wavefield polarity-aware U-Net for ultrasonic guided waves depth-resolved imaging","authors":"Xuan Li , Lishuai Liu , Jiang Lu , Yuncheng Zhang , Dazhi Cong , Yanxun Xiang","doi":"10.1016/j.ndteint.2025.103597","DOIUrl":"10.1016/j.ndteint.2025.103597","url":null,"abstract":"<div><div>The laser scanning full wavefield capture (LSFWC) technique enables rapid acquisition of two‐dimensional maps of local velocity variations of ultrasonic guided waves (UGWs), thereby facilitating quantitative spatial visualization of damage in thin‐walled structures. However, this technique faces a trade-off among spatial resolution, the accuracy of damage estimation, and detection efficiency: the spatial sampling rate influences the limit of velocity estimation, whereas increasing the sampling rate renders the full wavefield data acquisition extremely time-consuming. To tackle these difficulties, we propose a U-Net based imaging approach to reconstruct high-quality depth-resolved images from steady-state wavefield polarity (SSWP) of spatial sub-Nyquist sampled UGWs. The developed SSWP-aware U-Net can automatically segment the wavefield, extract local phase information, and perform encoding and decoding to establish approximate nonlinear mapping between wavefield and spatial distribution of UGWs velocity variations. The selection of SSWP as input features greatly accelerates the acquisition of the training dataset, and further enhances the robustness and generalization ability of U-Net trained on simulated dataset. We experimentally compared the proposed imaging method to wavenumber estimation methods. The results demonstrated that the proposed SSWP-aware U-Net is significantly superior to traditional methods in terms of image reconstruction quality and defect depth resolution especially with local spatial sub-Nyquist sampling.</div></div>","PeriodicalId":18868,"journal":{"name":"Ndt & E International","volume":"158 ","pages":"Article 103597"},"PeriodicalIF":4.5,"publicationDate":"2025-11-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145576322","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Magnetic resistance perturbation testing for high-sensitivity detection of subsurface and surface cracks in carbon steel 高灵敏度检测碳钢表面和地下裂纹的磁阻摄动试验

IF 4.5 2区材料科学 Q1 MATERIALS SCIENCE, CHARACTERIZATION & TESTING

Ndt & E International

Pub Date : 2025-11-14 DOI: 10.1016/j.ndteint.2025.103595

Jianxi Ding , Xin'an Yuan , Wei Li , Xiaokang Yin , Xiao Li , Dong Hu , Zichen Nie , Dehui Wang , Jianming Zhao

The alternating current field measurement (ACFM) is an excellent method for detecting cracks in underwater carbon steel structures. ACFM obtains crack information based on secondary electromagnetic field perturbation. The weak secondary electromagnetic field is insufficient for detecting subsurface and surface cracks. This study proposes a magnetic resistance perturbation testing (MRPT) method based on measured primary magnetic field in ACFM. Weak DC magnetization is employed to generate localized permeability perturbations surrounding surface and subsurface cracks. The permeability modifies the magnetic reluctance of carbon steel in the ACFM excitation field zone, leading to detectable amplitude variations in the primary magnetic field. The magnetic permeability disturbance transmits subsurface crack signatures to the surface detection zone and enhances the detectability of surface cracks. Being the spatial integration of permeability, the magnetic reluctance in the ACFM excitation zone effectively amplifies weak permeability disturbances of crack. This mechanism endows the MRPT method with enhanced sensitivity to both subsurface and surface cracks in any direction. Simulation and experiments on different cracks have been conducted to validate the efficiency of the MRPT method. The experimental results show that horizontal and vertical subsurface cracks with a depth of 11.5 mm can be detected with high sensitivity. The detection accuracy of surface cracks is improved.

交流电场测量（ACFM）是一种检测水下碳钢结构裂纹的有效方法。ACFM基于二次电磁场扰动获得裂纹信息。微弱的二次电磁场不足以探测地表和地下裂缝。本文提出了一种基于ACFM一次磁场测量的磁阻摄动测试方法。采用弱直流磁化产生表面和地下裂纹周围的局部磁导率微扰。磁导率改变了碳钢在ACFM励磁区的磁阻，导致一次磁场的可检测振幅变化。磁导率扰动将地下裂纹特征传递到地表探测区，提高了地表裂纹的可探测性。作为磁导率的空间积分，ACFM激励区的磁阻有效地放大了裂纹的弱磁导率扰动。这种机制使MRPT方法在任何方向上对地下和地表裂缝都具有更高的灵敏度。通过对不同裂纹的仿真和实验，验证了该方法的有效性。实验结果表明，该方法可以探测到深度为11.5 mm的水平和垂直地下裂缝，灵敏度较高。提高了表面裂纹的检测精度。

{"title":"Magnetic resistance perturbation testing for high-sensitivity detection of subsurface and surface cracks in carbon steel","authors":"Jianxi Ding , Xin'an Yuan , Wei Li , Xiaokang Yin , Xiao Li , Dong Hu , Zichen Nie , Dehui Wang , Jianming Zhao","doi":"10.1016/j.ndteint.2025.103595","DOIUrl":"10.1016/j.ndteint.2025.103595","url":null,"abstract":"<div><div>The alternating current field measurement (ACFM) is an excellent method for detecting cracks in underwater carbon steel structures. ACFM obtains crack information based on secondary electromagnetic field perturbation. The weak secondary electromagnetic field is insufficient for detecting subsurface and surface cracks. This study proposes a magnetic resistance perturbation testing (MRPT) method based on measured primary magnetic field in ACFM. Weak DC magnetization is employed to generate localized permeability perturbations surrounding surface and subsurface cracks. The permeability modifies the magnetic reluctance of carbon steel in the ACFM excitation field zone, leading to detectable amplitude variations in the primary magnetic field. The magnetic permeability disturbance transmits subsurface crack signatures to the surface detection zone and enhances the detectability of surface cracks. Being the spatial integration of permeability, the magnetic reluctance in the ACFM excitation zone effectively amplifies weak permeability disturbances of crack. This mechanism endows the MRPT method with enhanced sensitivity to both subsurface and surface cracks in any direction. Simulation and experiments on different cracks have been conducted to validate the efficiency of the MRPT method. The experimental results show that horizontal and vertical subsurface cracks with a depth of 11.5 mm can be detected with high sensitivity. The detection accuracy of surface cracks is improved.</div></div>","PeriodicalId":18868,"journal":{"name":"Ndt & E International","volume":"159 ","pages":"Article 103595"},"PeriodicalIF":4.5,"publicationDate":"2025-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145578738","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Accuracy of polarized laser scattering detection of surface/subsurface damage in ground quartz glass 磨砂石英玻璃表面/亚表面损伤的偏振激光散射检测精度

IF 4.5 2区材料科学 Q1 MATERIALS SCIENCE, CHARACTERIZATION & TESTING

Ndt & E International

Pub Date : 2025-11-12 DOI: 10.1016/j.ndteint.2025.103593

Shengwei Song , Pengfei Xu , Jingfei Yin , Han Haitjema

Quartz glass is widely used in high precision optical systems and instruments due to its excellent physical, chemical, and mechanical properties. Developing technology has increasing requirements on the surface integrity of a quartz glass part. However, there is a lack of high-efficiency and precision detection to the surface and subsurface damage in transparent quartz glass, which hinders the improvement of the surface integrity during a machining process. This study proposes a nondestructive dual sensor polarized laser scattering (PLS) system to detect the surface and subsurface damage in quartz glass. One sensor coupled with an integrating sphere is adopted to detect the surface damage. The other sensor is used to detect the subsurface damage. Both the surface and subsurface damage was measured by conventional methods to provide comparisons. Results show that the proposed dual-sensor PLS system represents high accuracies in detecting surface and subsurface damage in quartz glass. The detection errors of surface damage are less than 2 %. The detection errors of subsurface damage are about 6 %. The PLS signal is sensitive to the damage depth in quartz glass. It represents a constant sensitivity to the damage ranging from sub-micrometer to a few micrometers, which is beneficial to quantitively characterizing and evaluating the damage. Therefore, the dual sensor PLS system could achieve a detection with resolution to submicrometer scale. This study paves the way to the nondestructive detection of the damage in such transparent materials.

石英玻璃由于其优异的物理、化学和机械性能，被广泛应用于高精度光学系统和仪器中。技术的发展对石英玻璃零件的表面完整性提出了越来越高的要求。然而，目前缺乏对透明石英玻璃表面和亚表面损伤的高效、精确检测，阻碍了加工过程中表面完整性的提高。本研究提出了一种无损双传感器偏振激光散射（PLS）系统，用于检测石英玻璃表面和亚表面损伤。采用一个传感器与积分球耦合来检测表面损伤。另一个传感器用于检测亚表面损伤。表面和亚表面损伤均采用常规方法进行测量，以便进行比较。结果表明，所提出的双传感器PLS系统在检测石英玻璃表面和亚表面损伤方面具有较高的精度。表面损伤的检测误差小于2%。亚表面损伤的检测误差约为6%。PLS信号对石英玻璃的损伤深度敏感。它对亚微米到几微米范围内的损伤具有恒定的灵敏度，有利于对损伤进行定量表征和评价。因此，双传感器PLS系统可以实现亚微米级分辨率的检测。本研究为此类透明材料损伤的无损检测铺平了道路。

{"title":"Accuracy of polarized laser scattering detection of surface/subsurface damage in ground quartz glass","authors":"Shengwei Song , Pengfei Xu , Jingfei Yin , Han Haitjema","doi":"10.1016/j.ndteint.2025.103593","DOIUrl":"10.1016/j.ndteint.2025.103593","url":null,"abstract":"<div><div>Quartz glass is widely used in high precision optical systems and instruments due to its excellent physical, chemical, and mechanical properties. Developing technology has increasing requirements on the surface integrity of a quartz glass part. However, there is a lack of high-efficiency and precision detection to the surface and subsurface damage in transparent quartz glass, which hinders the improvement of the surface integrity during a machining process. This study proposes a nondestructive dual sensor polarized laser scattering (PLS) system to detect the surface and subsurface damage in quartz glass. One sensor coupled with an integrating sphere is adopted to detect the surface damage. The other sensor is used to detect the subsurface damage. Both the surface and subsurface damage was measured by conventional methods to provide comparisons. Results show that the proposed dual-sensor PLS system represents high accuracies in detecting surface and subsurface damage in quartz glass. The detection errors of surface damage are less than 2 %. The detection errors of subsurface damage are about 6 %. The PLS signal is sensitive to the damage depth in quartz glass. It represents a constant sensitivity to the damage ranging from sub-micrometer to a few micrometers, which is beneficial to quantitively characterizing and evaluating the damage. Therefore, the dual sensor PLS system could achieve a detection with resolution to submicrometer scale. This study paves the way to the nondestructive detection of the damage in such transparent materials.</div></div>","PeriodicalId":18868,"journal":{"name":"Ndt & E International","volume":"158 ","pages":"Article 103593"},"PeriodicalIF":4.5,"publicationDate":"2025-11-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145525782","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Machine learning-based phased array ultrasonic testing for graded materials: A comprehensive defect detection and quantification framework 基于机器学习的梯度材料相控阵超声检测：一个全面的缺陷检测和量化框架

IF 4.5 2区材料科学 Q1 MATERIALS SCIENCE, CHARACTERIZATION & TESTING

Ndt & E International

Pub Date : 2025-11-11 DOI: 10.1016/j.ndteint.2025.103594

Yongqiang Zhang , Hai Zhou , Pubo Li , Yongxin Lu , Gang Shen

Functionally graded materials (FGMs) are widely used in high-end fields like aerospace and energy for their customizable gradient properties, yet accurate detection of subtle defects in their inhomogeneous structures remains a key challenge for conventional non-destructive testing (NDT) techniques. To address this, this study proposes a multi-task learning-based phased array ultrasonic testing (PAUT) system for FGM defect inspection, featuring a multi-task neural network integrating CNN, RNN, and ensemble learning, plus gradient-corrected acoustic modeling, multi-scale feature extraction, and 3D reconstruction. A physic dataset was built based on Ti6Al4V-ZrO₂ FGM acoustic properties, incorporating gradient-induced wave distortion and Gaussian/speckle synthetic noise. The system's CNN extracts B-scan spatial features and LSTM captures A-scan temporal dependencies, enabling synergistic defect localization and quantification via a combined loss function optimized by Pareto multi-objective strategy. Experimental results show high detection accuracy for different size defects. Transfer learning adapts it to Al₂O₃-Ni FGMs and trained/validated on 5 defect types with Bayesian uncertainty quantification ensuring reliability. This work provides a physics-informed solution for FGM inspection, overcoming single-modal NDT and homogeneous-material model limitations, and supports intelligent testing system generalization in FGM-based high-end manufacturing.

功能梯度材料（fgm）因其可定制的梯度特性而广泛应用于航空航天和能源等高端领域，但在其非均匀结构中精确检测细微缺陷仍然是传统无损检测（NDT）技术的一个关键挑战。为了解决这一问题，本研究提出了一种基于多任务学习的相控阵超声检测（PAUT）系统，用于FGM缺陷检测，该系统采用集成CNN、RNN和集成学习的多任务神经网络，加上梯度校正声学建模、多尺度特征提取和三维重建。基于Ti6Al4V-ZrO2 FGM声学特性，结合梯度波畸变和高斯/散斑合成噪声，建立物理数据集。系统的CNN提取b扫描空间特征，LSTM捕获a扫描时间依赖关系，通过Pareto多目标策略优化的组合损失函数实现协同缺陷定位和量化。实验结果表明，对不同尺寸的缺陷均具有较高的检测精度。迁移学习使其适应于Al2O3-Ni fgm，并通过贝叶斯不确定性量化对5种缺陷类型进行训练/验证，以确保可靠性。这项工作为FGM检测提供了物理信息解决方案，克服了单模态无损检测和均质材料模型的限制，并支持基于FGM的高端制造中的智能测试系统推广。

{"title":"Machine learning-based phased array ultrasonic testing for graded materials: A comprehensive defect detection and quantification framework","authors":"Yongqiang Zhang , Hai Zhou , Pubo Li , Yongxin Lu , Gang Shen","doi":"10.1016/j.ndteint.2025.103594","DOIUrl":"10.1016/j.ndteint.2025.103594","url":null,"abstract":"<div><div>Functionally graded materials (FGMs) are widely used in high-end fields like aerospace and energy for their customizable gradient properties, yet accurate detection of subtle defects in their inhomogeneous structures remains a key challenge for conventional non-destructive testing (NDT) techniques. To address this, this study proposes a multi-task learning-based phased array ultrasonic testing (PAUT) system for FGM defect inspection, featuring a multi-task neural network integrating CNN, RNN, and ensemble learning, plus gradient-corrected acoustic modeling, multi-scale feature extraction, and 3D reconstruction. A physic dataset was built based on Ti6Al4V-ZrO<sub>2</sub> FGM acoustic properties, incorporating gradient-induced wave distortion and Gaussian/speckle synthetic noise. The system's CNN extracts B-scan spatial features and LSTM captures A-scan temporal dependencies, enabling synergistic defect localization and quantification via a combined loss function optimized by Pareto multi-objective strategy. Experimental results show high detection accuracy for different size defects. Transfer learning adapts it to Al<sub>2</sub>O<sub>3</sub>-Ni FGMs and trained/validated on 5 defect types with Bayesian uncertainty quantification ensuring reliability. This work provides a physics-informed solution for FGM inspection, overcoming single-modal NDT and homogeneous-material model limitations, and supports intelligent testing system generalization in FGM-based high-end manufacturing.</div></div>","PeriodicalId":18868,"journal":{"name":"Ndt & E International","volume":"158 ","pages":"Article 103594"},"PeriodicalIF":4.5,"publicationDate":"2025-11-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145576321","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Quantitative measurement method of wall thickness thinning for ferromagnetic components based on ramp-excited magnetization 基于斜坡激发磁化的铁磁元件壁厚减薄定量测量方法

IF 4.5 2区材料科学 Q1 MATERIALS SCIENCE, CHARACTERIZATION & TESTING

Ndt & E International

Pub Date : 2025-11-05 DOI: 10.1016/j.ndteint.2025.103590

Zhiyang Deng , Suyu Zhang , Xiaochun Song , Pan Qi , Yini Song , Yihua Kang

The evaluation of internal wall thickness thinning of ferromagnetic components is of vital importance in engineering applications. Existing non-destructive testing methods are difficult to ensure the detection accuracy and adaptability. This paper proposes a method for detecting wall thinning of ferromagnetic components based on ramp-excited magnetization (RE-MAG). Theoretical modeling and simulation modeling analysis show that under the ramp-excited magnetization, the permeability and voltage change curves will have two maximum values and one minimum value. There is a good linear relationship between the time difference

Δ t (Δ t^{'})

between these three extreme points and the wall thickness thinning. The experimental results show that the method is effective in detecting the thinning of the inner wall thickness of ferromagnetic components, and the relative error of the measurement results is up to 7.5 %. In addition, the measurement results are also related to the thinning width, magnetizing current frequency and magnetizing current strength, and different ferromagnetic materials have different calibration curves. The characteristic parameters of this method are easy to extract with good linearity, and the electromagnetic parameters of the specimen need not be known in advance.

铁磁元件的壁厚减薄评价在工程应用中具有重要意义。现有的无损检测方法难以保证检测精度和适应性。提出了一种基于斜坡激发磁化（RE-MAG）的铁磁元件壁薄检测方法。理论建模和仿真建模分析表明，在斜坡激励磁化作用下，磁导率和电压变化曲线存在两个最大值和一个最小值。这三个极值点之间的时间差Δt（Δt’）与壁厚减薄之间存在良好的线性关系。实验结果表明，该方法可以有效地检测铁磁元件内壁厚度的变薄，测量结果的相对误差可达7.5%。此外，测量结果还与减薄宽度、磁化电流频率和磁化电流强度有关，不同的铁磁材料具有不同的校准曲线。该方法的特征参数提取方便，线性度好，且不需要预先知道试样的电磁参数。

{"title":"Quantitative measurement method of wall thickness thinning for ferromagnetic components based on ramp-excited magnetization","authors":"Zhiyang Deng , Suyu Zhang , Xiaochun Song , Pan Qi , Yini Song , Yihua Kang","doi":"10.1016/j.ndteint.2025.103590","DOIUrl":"10.1016/j.ndteint.2025.103590","url":null,"abstract":"<div><div>The evaluation of internal wall thickness thinning of ferromagnetic components is of vital importance in engineering applications. Existing non-destructive testing methods are difficult to ensure the detection accuracy and adaptability. This paper proposes a method for detecting wall thinning of ferromagnetic components based on ramp-excited magnetization (RE-MAG). Theoretical modeling and simulation modeling analysis show that under the ramp-excited magnetization, the permeability and voltage change curves will have two maximum values and one minimum value. There is a good linear relationship between the time difference <span><math><mrow><mo>Δ</mo><mi>t</mi><mrow><mo>(</mo><mrow><mo>Δ</mo><msup><mi>t</mi><mo>′</mo></msup></mrow><mo>)</mo></mrow></mrow></math></span> between these three extreme points and the wall thickness thinning. The experimental results show that the method is effective in detecting the thinning of the inner wall thickness of ferromagnetic components, and the relative error of the measurement results is up to 7.5 %. In addition, the measurement results are also related to the thinning width, magnetizing current frequency and magnetizing current strength, and different ferromagnetic materials have different calibration curves. The characteristic parameters of this method are easy to extract with good linearity, and the electromagnetic parameters of the specimen need not be known in advance.</div></div>","PeriodicalId":18868,"journal":{"name":"Ndt & E International","volume":"158 ","pages":"Article 103590"},"PeriodicalIF":4.5,"publicationDate":"2025-11-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145466173","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Precise measurement of oxide film thickness on fuel cladding utilizing a stacked array probe and a SVR model 基于堆叠阵列探针和SVR模型的燃料包壳氧化膜厚度精确测量

IF 4.5 2区材料科学 Q1 MATERIALS SCIENCE, CHARACTERIZATION & TESTING

Ndt & E International

Pub Date : 2025-11-05 DOI: 10.1016/j.ndteint.2025.103591

Liuxin Ge , Qi Xiao , Ming Li , Songling Huang , Na Zhang , Chaofeng Ye

Oxide films accumulated on cladding tubes in nuclear reactors impede heat transfer and exacerbate corrosion. Measurement of the thickness of oxide film is imperative to maintain the safety of nuclear plants. The evaluation of oxide film thickness in nuclear power plant applications frequently involves the use of eddy current testing (ECT), which requires a calibration phase during the measurement process. Inaccuracies occur when the object’s material differs from the calibration samples. To address this issue, this paper presents an innovative approach that combines stacked array coils with a support vector regression (SVR) model to increase the measurement precision. The stacked array sensor comprises multiple coils separated by insulation layers with thicknesses of a few tens of microns, which function as internal distance standards to calibrate the oxide film thickness. The SVR model is developed to calculate the thickness of the oxide film from the measurement by solving the regression problem. This method mitigates the influence of the material property. A three-dimensional numerical model is developed to calculate the signals for various material conductivities. The simulation data are employed to train the SVR model and assess the accuracy of the approach. The findings demonstrate that the root mean square error and the coefficient of determination for the oxide film thickness calculation are 0.30

μ

m and 1.00, respectively. Furthermore, an experimental system is established and data are collected to validate the feasibility of the method. The experimental results indicate that the maximum inaccuracy remains below

2 μ m

even with different material conductivities.

核反应堆包层管上堆积的氧化膜阻碍了传热，加剧了腐蚀。氧化膜厚度的测量对维护核电站的安全至关重要。在核电站应用中，氧化膜厚度的评估经常涉及到涡流测试（ECT）的使用，在测量过程中需要一个校准阶段。当物体的材料与校准样品不同时，就会出现不准确性。为了解决这一问题，本文提出了一种将堆叠阵列线圈与支持向量回归（SVR）模型相结合的创新方法，以提高测量精度。堆叠阵列传感器由多个线圈组成，由厚度为几十微米的绝缘层隔开，作为内部距离标准，用于校准氧化膜厚度。通过求解回归问题，建立SVR模型，从测量结果计算氧化膜厚度。这种方法减轻了材料性能的影响。建立了一个三维数值模型来计算各种材料电导率的信号。利用仿真数据对SVR模型进行训练，并对方法的精度进行评估。结果表明，计算氧化膜厚度的均方根误差为0.30 μm，决定系数为1.00。建立了实验系统，并进行了数据采集，验证了该方法的可行性。实验结果表明，在不同的材料电导率下，最大误差仍保持在2μm以下。

{"title":"Precise measurement of oxide film thickness on fuel cladding utilizing a stacked array probe and a SVR model","authors":"Liuxin Ge , Qi Xiao , Ming Li , Songling Huang , Na Zhang , Chaofeng Ye","doi":"10.1016/j.ndteint.2025.103591","DOIUrl":"10.1016/j.ndteint.2025.103591","url":null,"abstract":"<div><div>Oxide films accumulated on cladding tubes in nuclear reactors impede heat transfer and exacerbate corrosion. Measurement of the thickness of oxide film is imperative to maintain the safety of nuclear plants. The evaluation of oxide film thickness in nuclear power plant applications frequently involves the use of eddy current testing (ECT), which requires a calibration phase during the measurement process. Inaccuracies occur when the object’s material differs from the calibration samples. To address this issue, this paper presents an innovative approach that combines stacked array coils with a support vector regression (SVR) model to increase the measurement precision. The stacked array sensor comprises multiple coils separated by insulation layers with thicknesses of a few tens of microns, which function as internal distance standards to calibrate the oxide film thickness. The SVR model is developed to calculate the thickness of the oxide film from the measurement by solving the regression problem. This method mitigates the influence of the material property. A three-dimensional numerical model is developed to calculate the signals for various material conductivities. The simulation data are employed to train the SVR model and assess the accuracy of the approach. The findings demonstrate that the root mean square error and the coefficient of determination for the oxide film thickness calculation are 0.30 <span><math><mi>μ</mi></math></span>m and 1.00, respectively. Furthermore, an experimental system is established and data are collected to validate the feasibility of the method. The experimental results indicate that the maximum inaccuracy remains below <span><math><mrow><mn>2</mn><mspace></mspace><mi>μ</mi><mi>m</mi></mrow></math></span> even with different material conductivities.</div></div>","PeriodicalId":18868,"journal":{"name":"Ndt & E International","volume":"158 ","pages":"Article 103591"},"PeriodicalIF":4.5,"publicationDate":"2025-11-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145466174","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0