铝合金铸件已经广泛应用于汽车、飞机等重要工业,其质量直接影响到机械零部件的安全使用。针对铝合金铸件的X射线图像表面和内部缺陷多样化和细微问题,提出了一种融合局部和全局特征的X射线图像铸造缺陷检测方法。首先,将高效通道注意...铝合金铸件已经广泛应用于汽车、飞机等重要工业,其质量直接影响到机械零部件的安全使用。针对铝合金铸件的X射线图像表面和内部缺陷多样化和细微问题,提出了一种融合局部和全局特征的X射线图像铸造缺陷检测方法。首先,将高效通道注意力模块和经典网络ResNet-50进行融合构成新的基础卷积神经网络,以新的网络为骨干构建双分支网络模型。然后,提出了细节信息定位提取(detailed information location and extraction,DILE)模块,该模块定位到包含丰富判别性信息的局部区域。最后,将由DILE得到的局部图像结合原始图像作为网络的输入,构建了一个融合局部和全局特征的双分支网络模型。对全局区域的学习有助于在复杂背景下提取有意义的细微信息,对局部区域的学习可以进一步提高分类效果。该方法在真实汽车铸件的X射线图像数据集上进行测试训练,测试集准确率达98.3%。实验结果表明,该方法相较于其他常规方法更有效。展开更多
Subsurface defects were fluorescently tagged with nanoscale quantum dots and scanned layer by layer using confocal fluorescence microscopy to obtain images at various depths. Subsurface damage depths of fused silica o...Subsurface defects were fluorescently tagged with nanoscale quantum dots and scanned layer by layer using confocal fluorescence microscopy to obtain images at various depths. Subsurface damage depths of fused silica optics were characterized quantitatively by changes in the fluorescence intensity of feature points. The fluorescence intensity vs scan depth revealed that the maximum fluorescence intensity decreases sharply when the scan depth exceeds a critical value. The subsurface damage depth could be determined by the actual embedded depth of the quantum dots. Taper polishing and magnetorheological finishing were performed under the same conditions to verify the effectiveness of the nondestructive fluorescence method. The results indicated that the quantum dots effectively tagged subsurface defects of fused-silica optics, and that the nondestructive detection method could effectively evaluate subsurface damage depths.展开更多
Local defected resonance(LDR)is a recently-developed non-destructive testing method,which identifies damage by detecting the vibrational response of the structural surface under the wideband ultrasonic excitation. The...Local defected resonance(LDR)is a recently-developed non-destructive testing method,which identifies damage by detecting the vibrational response of the structural surface under the wideband ultrasonic excitation. The concept of LDR is studied and applied for damage imaging of delamination in composite laminates. Aiming at the problem of poor anti-noise ability and inaccurate damage identification in traditional detection process,an LDR-based multi-frequency method is proposed. Experimental results show that the proposed method can realize the localization and imaging of delamination damage in composite materials.展开更多
The oscillating voltage test is a nondestructive detection method for partial discharge of XLPE （cross linked polyethylene） cable and has been applied recently. This paper made three kinds of varying severity artifi...The oscillating voltage test is a nondestructive detection method for partial discharge of XLPE （cross linked polyethylene） cable and has been applied recently. This paper made three kinds of varying severity artificial defect models of cable joints in 10 kV XLPE cable. Oscillating voltage is applied to the model, by use of pulse current method to detect partial discharge signals. In order to study the statistical characteristics of partial discharge of cable joint under the oscillating voltage, three-dimensional statistical map has been made. The results show that for the same kind of defects, with the increases of the defect severity, the discharge interval extended, the magnitude and the number of partial discharge increase, for different kinds of defects, obvious differences exist among the maps, this may established a foundation for the further study of the partial discharge pattern recognition of XLPE cable under oscillating voltage.展开更多
文摘铝合金铸件已经广泛应用于汽车、飞机等重要工业,其质量直接影响到机械零部件的安全使用。针对铝合金铸件的X射线图像表面和内部缺陷多样化和细微问题,提出了一种融合局部和全局特征的X射线图像铸造缺陷检测方法。首先,将高效通道注意力模块和经典网络ResNet-50进行融合构成新的基础卷积神经网络,以新的网络为骨干构建双分支网络模型。然后,提出了细节信息定位提取(detailed information location and extraction,DILE)模块,该模块定位到包含丰富判别性信息的局部区域。最后,将由DILE得到的局部图像结合原始图像作为网络的输入,构建了一个融合局部和全局特征的双分支网络模型。对全局区域的学习有助于在复杂背景下提取有意义的细微信息,对局部区域的学习可以进一步提高分类效果。该方法在真实汽车铸件的X射线图像数据集上进行测试训练,测试集准确率达98.3%。实验结果表明,该方法相较于其他常规方法更有效。
基金Project(JCKY2016212A506-0503) supported by the Science Challenge Project of ChinaProject(51475106) supported by the National Natural Science Foundation of China
文摘Subsurface defects were fluorescently tagged with nanoscale quantum dots and scanned layer by layer using confocal fluorescence microscopy to obtain images at various depths. Subsurface damage depths of fused silica optics were characterized quantitatively by changes in the fluorescence intensity of feature points. The fluorescence intensity vs scan depth revealed that the maximum fluorescence intensity decreases sharply when the scan depth exceeds a critical value. The subsurface damage depth could be determined by the actual embedded depth of the quantum dots. Taper polishing and magnetorheological finishing were performed under the same conditions to verify the effectiveness of the nondestructive fluorescence method. The results indicated that the quantum dots effectively tagged subsurface defects of fused-silica optics, and that the nondestructive detection method could effectively evaluate subsurface damage depths.
基金supported by the National Natural Science Foundation of China(Nos.51875227,51805261,51775267)the Natural Science Foundation of Jiangsu Province(Nos.BK20181286,BK20180430)the Aviation Science Fund(No.20161552014)
文摘Local defected resonance(LDR)is a recently-developed non-destructive testing method,which identifies damage by detecting the vibrational response of the structural surface under the wideband ultrasonic excitation. The concept of LDR is studied and applied for damage imaging of delamination in composite laminates. Aiming at the problem of poor anti-noise ability and inaccurate damage identification in traditional detection process,an LDR-based multi-frequency method is proposed. Experimental results show that the proposed method can realize the localization and imaging of delamination damage in composite materials.
文摘The oscillating voltage test is a nondestructive detection method for partial discharge of XLPE （cross linked polyethylene） cable and has been applied recently. This paper made three kinds of varying severity artificial defect models of cable joints in 10 kV XLPE cable. Oscillating voltage is applied to the model, by use of pulse current method to detect partial discharge signals. In order to study the statistical characteristics of partial discharge of cable joint under the oscillating voltage, three-dimensional statistical map has been made. The results show that for the same kind of defects, with the increases of the defect severity, the discharge interval extended, the magnitude and the number of partial discharge increase, for different kinds of defects, obvious differences exist among the maps, this may established a foundation for the further study of the partial discharge pattern recognition of XLPE cable under oscillating voltage.
