提出了一种数学形态学与GG(Gath-Geva)模糊聚类相结合的旋转机械故障诊断方法,通过对滚动轴承信号的多尺度形态运算得到信号的形态谱,定量反映了信号在不同尺度下的形态变化特征。为进一步对滚动轴承信号进行故障识别,提取出基于形态学...提出了一种数学形态学与GG(Gath-Geva)模糊聚类相结合的旋转机械故障诊断方法,通过对滚动轴承信号的多尺度形态运算得到信号的形态谱,定量反映了信号在不同尺度下的形态变化特征。为进一步对滚动轴承信号进行故障识别,提取出基于形态学操作的分形维数和描述不同信号形态特征的指标即形态谱熵,并把这2个参数作为GG聚类的故障特征向量,进行聚类分析,同时对GG聚类与FCM(fuzzy center means)聚类和GK(Gustafaon-Kessel)聚类进行了比较。实验证明了基于数学形态学与GG聚类相结合的机械故障诊断方法的有效性,且证明了GG聚类更适合对不同形状、大小和密度的空间故障数据模糊聚类,聚类效果更好。展开更多
Late Mesozoic granitoids are widespread in the Great Xing’an Range(GXR), which is part of a large igneous province in eastern China. The geodynamic setting of the Late Mesozoic granitoids is still debated, and there ...Late Mesozoic granitoids are widespread in the Great Xing’an Range(GXR), which is part of a large igneous province in eastern China. The geodynamic setting of the Late Mesozoic granitoids is still debated, and there have been two dominant models proposed, subduction and thermal erosion. This study discusses the geodynamic mechanisms from a new perspective on ages of the granitoids and fractal dimensions of their shape. Our results show that granitoids become gradually older from South GXR to North GXR to Erguna Block(EB) in the Jurassic, and opposite in the Cretaceous. The fractal dimensions of the Perimeter-area model(DAP) exhibit the same features. The values of DAP are smaller from South GXR(0.673 1) to North GXR(0.628 0) to EB(0.607 9) in the Jurassic, and larger from South GXR(0.609 6) to North GXR(0.630 2) to EB(0.639 9) in the Cretaceous. This implies that the geometrical irregularities of the granitoids are shaped by subduction rather than thermal erosion. These spatial variations could be best explained by the subduction of the Pacific Plate and consequent granitoid magmatism in the Late Mesozoic, thus providing a new fractal evidence for Pacific Plate subduction mechanism and opening a new possibility method for studing plate movement.展开更多
In the last decade,micro-architected structures have gained significant attention in academia and industry for their lightweight,strong,and thermally efficient properties.Inspired by biomimicry design,this paper prese...In the last decade,micro-architected structures have gained significant attention in academia and industry for their lightweight,strong,and thermally efficient properties.Inspired by biomimicry design,this paper presents a novel ribbed family of additively manufactured Micro-Architected Domes(MAD).The design incorporates tetrapod pyramid unit cells,golden ratio-based fractal patterns,Schoen’s Minimal Gyroid,and spherical geometry.The study focuses on dome radius,height,and azimuth/elevation partitioning as input variables,with the main output being ribbed micro-cell diameter.The relationships between unit-cells’diameter and input variables were established through problem-solving and numerical computations:linear dependency with the dome radius and hyperbolic dependency with the azimuth and elevation partitioning.The proposed design successfully adhered to the Surface-to-Volume ratio of Schoen’s Minimal Gyroid,achieving an average volume relative density of 2.5%,confirming its lightweight nature.The feasibility of the design was further supported by fabricating three specimens using Filament Fused Fabrication.This research showcases the potential of biomimicry-inspired micro-architected structures,paving the way for innovative applications in various fields.展开更多
输电线路故障信号是一种典型的非线性信号,分形几何理论为描述非线性故障信号的特性提供了一个有力的分析工具。针对传统分形维数的局限性,本文提出了一种基于局域均值分解(local mean decomposition,LMD)-形态学的分形维数-Elman神经...输电线路故障信号是一种典型的非线性信号,分形几何理论为描述非线性故障信号的特性提供了一个有力的分析工具。针对传统分形维数的局限性,本文提出了一种基于局域均值分解(local mean decomposition,LMD)-形态学的分形维数-Elman神经网络的输电线路故障选相新方法。该方法通过对故障电流进行相模转换后,对单一线模分量进行LMD分解得到若干乘积函数(product function,PF)分量,然后选取前4个PF分量进行数学形态学的分形维数估计,最后形成特征向量作为Elman神经网络的输入参数。仿真试验表明:提出的故障分类识别方法能快速、准确地识别各类故障,并且不易受故障初始角、故障位置和过渡电阻的影响,与传统的BP神经网络相比,Elman神经网络具有更好的效果,为准确判断输电线路故障选相提供了一种快速有效的新方法。展开更多
文摘提出了一种数学形态学与GG(Gath-Geva)模糊聚类相结合的旋转机械故障诊断方法,通过对滚动轴承信号的多尺度形态运算得到信号的形态谱,定量反映了信号在不同尺度下的形态变化特征。为进一步对滚动轴承信号进行故障识别,提取出基于形态学操作的分形维数和描述不同信号形态特征的指标即形态谱熵,并把这2个参数作为GG聚类的故障特征向量,进行聚类分析,同时对GG聚类与FCM(fuzzy center means)聚类和GK(Gustafaon-Kessel)聚类进行了比较。实验证明了基于数学形态学与GG聚类相结合的机械故障诊断方法的有效性,且证明了GG聚类更适合对不同形状、大小和密度的空间故障数据模糊聚类,聚类效果更好。
文摘针对滚动轴承状态监测实时性差、故障诊断准确率低的问题,提出一种基于改进局部均值分解(ILMD)和数学形态学分形理论的特征提取算法,并结合概率神经网络(PNN)完成对轴承状态的智能化识别分类.该算法首先通过ILMD分解轴承原始振动信号,选取相关性系数最大的两阶分量,求取其分形维数作为特征向量;其次,结合盒维数理论,将“形态学覆盖面积”作为第三维特征向量,同时构建起三维特征矩阵;最后,将特征矩阵输入PNN以完成状态的识别分类.使用西储大学实测轴承数据验证算法,结果表明,该算法不仅能够精确识别不同状态的轴承,还能有效分类同种故障下不同损伤程度的轴承状态,平均识别率超过99.6%,平均识别时间0.21 s.
基金supported by the National Key R & D Program of China (No. 2016YFC0600501)the Open Research Project of The Hubei Key Laboratory of Intelligent Geo-Information Processing 295 (No. KLIGIP-2017A03)the National and Nature Science Foundation of China (Nos. 41430320, 41572315)
文摘Late Mesozoic granitoids are widespread in the Great Xing’an Range(GXR), which is part of a large igneous province in eastern China. The geodynamic setting of the Late Mesozoic granitoids is still debated, and there have been two dominant models proposed, subduction and thermal erosion. This study discusses the geodynamic mechanisms from a new perspective on ages of the granitoids and fractal dimensions of their shape. Our results show that granitoids become gradually older from South GXR to North GXR to Erguna Block(EB) in the Jurassic, and opposite in the Cretaceous. The fractal dimensions of the Perimeter-area model(DAP) exhibit the same features. The values of DAP are smaller from South GXR(0.673 1) to North GXR(0.628 0) to EB(0.607 9) in the Jurassic, and larger from South GXR(0.609 6) to North GXR(0.630 2) to EB(0.639 9) in the Cretaceous. This implies that the geometrical irregularities of the granitoids are shaped by subduction rather than thermal erosion. These spatial variations could be best explained by the subduction of the Pacific Plate and consequent granitoid magmatism in the Late Mesozoic, thus providing a new fractal evidence for Pacific Plate subduction mechanism and opening a new possibility method for studing plate movement.
基金The authors would like to thank Ecole Nationale Superieure d’Arts&Metiers de Meknes,Moulay Ismail University,Morocco for providing Ansys SpaceClaim R21.Many thanks also to Euromed Center of Research,Euromed University of Fes,Morocco for the availability of Matlab(2022)that allowed performing all the numerical computations,as well as the access to the VOLUMIC Stream 30 Ultra 3D printer for MAD prototyping.
文摘In the last decade,micro-architected structures have gained significant attention in academia and industry for their lightweight,strong,and thermally efficient properties.Inspired by biomimicry design,this paper presents a novel ribbed family of additively manufactured Micro-Architected Domes(MAD).The design incorporates tetrapod pyramid unit cells,golden ratio-based fractal patterns,Schoen’s Minimal Gyroid,and spherical geometry.The study focuses on dome radius,height,and azimuth/elevation partitioning as input variables,with the main output being ribbed micro-cell diameter.The relationships between unit-cells’diameter and input variables were established through problem-solving and numerical computations:linear dependency with the dome radius and hyperbolic dependency with the azimuth and elevation partitioning.The proposed design successfully adhered to the Surface-to-Volume ratio of Schoen’s Minimal Gyroid,achieving an average volume relative density of 2.5%,confirming its lightweight nature.The feasibility of the design was further supported by fabricating three specimens using Filament Fused Fabrication.This research showcases the potential of biomimicry-inspired micro-architected structures,paving the way for innovative applications in various fields.
文摘输电线路故障信号是一种典型的非线性信号,分形几何理论为描述非线性故障信号的特性提供了一个有力的分析工具。针对传统分形维数的局限性,本文提出了一种基于局域均值分解(local mean decomposition,LMD)-形态学的分形维数-Elman神经网络的输电线路故障选相新方法。该方法通过对故障电流进行相模转换后,对单一线模分量进行LMD分解得到若干乘积函数(product function,PF)分量,然后选取前4个PF分量进行数学形态学的分形维数估计,最后形成特征向量作为Elman神经网络的输入参数。仿真试验表明:提出的故障分类识别方法能快速、准确地识别各类故障,并且不易受故障初始角、故障位置和过渡电阻的影响,与传统的BP神经网络相比,Elman神经网络具有更好的效果,为准确判断输电线路故障选相提供了一种快速有效的新方法。
