The multi-principal element characteristic of high-entropy alloys has revolutionized the conventional alloy design concept of single-principal element,endowing them with excellent mechanical properties.However,owing t...The multi-principal element characteristic of high-entropy alloys has revolutionized the conventional alloy design concept of single-principal element,endowing them with excellent mechanical properties.However,owing to this multi-principal element nature,high-entropy alloys exhibit complex deformation behavior dominated by alternating and coupled deformation mechanisms.Therefore,elucidating these intricate deformation mechanisms remains a key challenge in current research.Neutron diffraction(ND)techniques offer distinct advantages over traditional microscopic methods for characterizing such complex deformation behavior.The strong penetration capability of neutrons enables in-situ,real-time,and non-destructive detection of structural evolution in most centimeter-level bulk samples under complex environments,and ND allows precise characterization of lattice site occupations for light elements,such as C and O,and neighboring elements.This review discussed the principles of ND,experiment procedures,and data analysis.Combining with recent advances in the research about face-centered cubic high-entropy alloy,typical examples of using ND to investigate the deformation behavior were summarized,ultimately revealing deformation mechanisms dominated by dislocations,stacking faults,twinning,and phase transformations.展开更多
为提高管道环焊缝超声衍射时差法(time of flight diffraction,TOFD)扫描图谱在背景信号干扰、样本量不均衡等情况下的缺陷识别效果,提出了一种改进的YOLOv5s网络模型.针对管道环焊缝TOFD图谱中缺陷形态不规则的特点,通过引入可变形卷积...为提高管道环焊缝超声衍射时差法(time of flight diffraction,TOFD)扫描图谱在背景信号干扰、样本量不均衡等情况下的缺陷识别效果,提出了一种改进的YOLOv5s网络模型.针对管道环焊缝TOFD图谱中缺陷形态不规则的特点,通过引入可变形卷积,使得网络自适应缺陷自身的形状特点,提高TOFD图谱中不规则缺陷的特征提取能力;针对TOFD扫描图谱中直通波和底面波等干扰波形对缺陷识别的影响,通过在网络不同深度分别添加自注意力机制,引导网络关注缺陷细微特征的同时抑制界面波对缺陷识别的影响;针对实际样本中各类缺陷不均衡的情况,采用SlideLoss损失函数代替原损失函数,提高网络对样本量较少的裂纹类缺陷的识别精度.对比试验结果表明,改进后的网络能够抑制TOFD图谱复杂背景干扰,提高样本不均衡条件下的识别率.相比原网络,整体平均识别率均值(mean Average Precision,mAP)和裂纹类缺陷的平均识别率(Average Precision,AP)分别提高了8.2%和7.3%.展开更多
基金National Key R&D Program of China(2023YFB3711904,2022YFA1603801)National Natural Science Foundation of China(12404230,52471181,52301213,52130108,52471005)+2 种基金National Nature Science Foundation of Zhejiang Province(LY23E010002)Open Fund of the China Spallation Neutron Source,Songshan Lake Science City(KFKT2023B11)Guangdong Basic and Applied Basic Research Foundation(2022A1515110805,2024A1515010878)。
文摘The multi-principal element characteristic of high-entropy alloys has revolutionized the conventional alloy design concept of single-principal element,endowing them with excellent mechanical properties.However,owing to this multi-principal element nature,high-entropy alloys exhibit complex deformation behavior dominated by alternating and coupled deformation mechanisms.Therefore,elucidating these intricate deformation mechanisms remains a key challenge in current research.Neutron diffraction(ND)techniques offer distinct advantages over traditional microscopic methods for characterizing such complex deformation behavior.The strong penetration capability of neutrons enables in-situ,real-time,and non-destructive detection of structural evolution in most centimeter-level bulk samples under complex environments,and ND allows precise characterization of lattice site occupations for light elements,such as C and O,and neighboring elements.This review discussed the principles of ND,experiment procedures,and data analysis.Combining with recent advances in the research about face-centered cubic high-entropy alloy,typical examples of using ND to investigate the deformation behavior were summarized,ultimately revealing deformation mechanisms dominated by dislocations,stacking faults,twinning,and phase transformations.
文摘为提高管道环焊缝超声衍射时差法(time of flight diffraction,TOFD)扫描图谱在背景信号干扰、样本量不均衡等情况下的缺陷识别效果,提出了一种改进的YOLOv5s网络模型.针对管道环焊缝TOFD图谱中缺陷形态不规则的特点,通过引入可变形卷积,使得网络自适应缺陷自身的形状特点,提高TOFD图谱中不规则缺陷的特征提取能力;针对TOFD扫描图谱中直通波和底面波等干扰波形对缺陷识别的影响,通过在网络不同深度分别添加自注意力机制,引导网络关注缺陷细微特征的同时抑制界面波对缺陷识别的影响;针对实际样本中各类缺陷不均衡的情况,采用SlideLoss损失函数代替原损失函数,提高网络对样本量较少的裂纹类缺陷的识别精度.对比试验结果表明,改进后的网络能够抑制TOFD图谱复杂背景干扰,提高样本不均衡条件下的识别率.相比原网络,整体平均识别率均值(mean Average Precision,mAP)和裂纹类缺陷的平均识别率(Average Precision,AP)分别提高了8.2%和7.3%.
