针对变电站绝缘套管过热红外图像检测精度不高的问题,提出了基于改进YOLO第7版(you only look once version 7,YOLOv7)算法的检测技术。通过引入改良的跨阶段部分网络幽灵版本3(cross stage partial network ghost version 3,C3Ghost)...针对变电站绝缘套管过热红外图像检测精度不高的问题,提出了基于改进YOLO第7版(you only look once version 7,YOLOv7)算法的检测技术。通过引入改良的跨阶段部分网络幽灵版本3(cross stage partial network ghost version 3,C3Ghost)模块替换头部网络中的扩展高效层聚合网络(extended efficient layer aggregation network,E-ELAN)模块,优化了网络结构,增强了算法对小目标的识别能力。此外,整合了轻量级基于归一化的注意力模块(normalization-based attention module,NAM)到主干网络中以提高对红外图像特征的利用效率,并引入幽灵卷积(ghost convolution,GhostConv)模块替换了网络中的所有卷积,显著降低了模型的大小。结果表明,与YOLOv7初始算法相比,改进YOLOv7算法在F1评分和平均精确率均值上分别提高了19.51%和16.57%,算法的参数量减小了16.3 MB,且检测速度达到了41帧/s,充分证明了该算法在变电站实际应用中的有效性。该研究不仅显著提高了变电站绝缘套管过热红外图像检测的准确性,也能为后续相关技术的研究提供参考。展开更多
The rugosan fauna from the Guanyinqiao Bed （latest Ordovician, Hirnantian） of northern Guizhou, China is known to belong to the cold or cool-water type corals. The components of the fauna are solitary corals only, a...The rugosan fauna from the Guanyinqiao Bed （latest Ordovician, Hirnantian） of northern Guizhou, China is known to belong to the cold or cool-water type corals. The components of the fauna are solitary corals only, and corallite septa are generally strongly dilated, especially the streptelasmatid corals are dominant comprising 98% of the whole fauna. The Guanyinqiao Bed is rich in rugosans of 18 genera, which are streptelasnmtid Streptelasma （=Helicelasma）, Brachyelasma, Amplexobrachyelasma, Salvadorea, Grewingkia, Borelasma, CrassUasma, Leolasma, KenophyUum, UUernelasma, Paramplexoides, Siphonolasma, Pycnactoides, Dalmanophyllum, Bodophyllum, Axiphoria, Lambeophyllum and cystiphyllid Sinkiangolasma. Although this fauna was fairly abundant in a confined area （northern-northeastern Guizhou, southern Sichuan） during the Hirnantian age, the rugosan mass extinction （generic extinction rate 81%） happened at the end of the Hirnantian Stage. It is conduded that the mass extinction is related to the ending of maximum glaciation and ice cap melting in Gondwana in the southern hemisphere in the latest Hirnantian, resulting in rapid global sea-level rise in the earliest Silurian. In the Upper Yangtze Basin, the sea bottom environments were replaced by anoxic and warmer water during that time, so that the cool-water type rugosan became extinct. The present paper attempts to revise some already described rugose coral genera and species （He, 1978, 1985） and to supplement a few new forms from the Guanyinqiao Bed. Fourteen species of 12 genera are re-described and illustrated, of which one species- Grewingkia latifossulata is new. As a whole, the rugosan fauna of the Guanyinqiao Bed may be correlated with those contemporaneous of North Europe, Estonia and North America, indicating a dose biogeographic affinity to North Europe.展开更多
电力变压器是电网中的核心设备,其运行状态直接关系到系统的供电可靠性和安全。为了充分利用变压器多元状态监测信息,实现对变压器状态的精确感知,对在线监测参量进行了统计分析,确定了监测参数间存在一定相关性。提出基于多状态参量的...电力变压器是电网中的核心设备,其运行状态直接关系到系统的供电可靠性和安全。为了充分利用变压器多元状态监测信息,实现对变压器状态的精确感知,对在线监测参量进行了统计分析,确定了监测参数间存在一定相关性。提出基于多状态参量的回归特性进行设备状态分析的PLSPCA(partial least squares-principal component analysis)方法,即在设备正常运行时,利用偏最小二乘回归(PLS)方法挖掘状态参量之间的关联关系及回归方程,基于回归预测与实际之间的偏差构建了基于主成分分析(PCA)的变压器状态评判方法,通过控制图对比了变压器状态变化检出的能力。通过对变压器正常运行状态和异常运行状态监测数据进行分析,结果表明:多状态参数回归分析可以将变压器的多元监测数据直观地显示出来;该方法利用监测参数间的相关性,通过发现变压器运行过程中状态参数的霍特林统计量,可以提前检测变压器的异常运行状态。展开更多
文摘针对变电站绝缘套管过热红外图像检测精度不高的问题,提出了基于改进YOLO第7版(you only look once version 7,YOLOv7)算法的检测技术。通过引入改良的跨阶段部分网络幽灵版本3(cross stage partial network ghost version 3,C3Ghost)模块替换头部网络中的扩展高效层聚合网络(extended efficient layer aggregation network,E-ELAN)模块,优化了网络结构,增强了算法对小目标的识别能力。此外,整合了轻量级基于归一化的注意力模块(normalization-based attention module,NAM)到主干网络中以提高对红外图像特征的利用效率,并引入幽灵卷积(ghost convolution,GhostConv)模块替换了网络中的所有卷积,显著降低了模型的大小。结果表明,与YOLOv7初始算法相比,改进YOLOv7算法在F1评分和平均精确率均值上分别提高了19.51%和16.57%,算法的参数量减小了16.3 MB,且检测速度达到了41帧/s,充分证明了该算法在变电站实际应用中的有效性。该研究不仅显著提高了变电站绝缘套管过热红外图像检测的准确性,也能为后续相关技术的研究提供参考。
文摘The rugosan fauna from the Guanyinqiao Bed （latest Ordovician, Hirnantian） of northern Guizhou, China is known to belong to the cold or cool-water type corals. The components of the fauna are solitary corals only, and corallite septa are generally strongly dilated, especially the streptelasmatid corals are dominant comprising 98% of the whole fauna. The Guanyinqiao Bed is rich in rugosans of 18 genera, which are streptelasnmtid Streptelasma （=Helicelasma）, Brachyelasma, Amplexobrachyelasma, Salvadorea, Grewingkia, Borelasma, CrassUasma, Leolasma, KenophyUum, UUernelasma, Paramplexoides, Siphonolasma, Pycnactoides, Dalmanophyllum, Bodophyllum, Axiphoria, Lambeophyllum and cystiphyllid Sinkiangolasma. Although this fauna was fairly abundant in a confined area （northern-northeastern Guizhou, southern Sichuan） during the Hirnantian age, the rugosan mass extinction （generic extinction rate 81%） happened at the end of the Hirnantian Stage. It is conduded that the mass extinction is related to the ending of maximum glaciation and ice cap melting in Gondwana in the southern hemisphere in the latest Hirnantian, resulting in rapid global sea-level rise in the earliest Silurian. In the Upper Yangtze Basin, the sea bottom environments were replaced by anoxic and warmer water during that time, so that the cool-water type rugosan became extinct. The present paper attempts to revise some already described rugose coral genera and species （He, 1978, 1985） and to supplement a few new forms from the Guanyinqiao Bed. Fourteen species of 12 genera are re-described and illustrated, of which one species- Grewingkia latifossulata is new. As a whole, the rugosan fauna of the Guanyinqiao Bed may be correlated with those contemporaneous of North Europe, Estonia and North America, indicating a dose biogeographic affinity to North Europe.
文摘电力变压器是电网中的核心设备,其运行状态直接关系到系统的供电可靠性和安全。为了充分利用变压器多元状态监测信息,实现对变压器状态的精确感知,对在线监测参量进行了统计分析,确定了监测参数间存在一定相关性。提出基于多状态参量的回归特性进行设备状态分析的PLSPCA(partial least squares-principal component analysis)方法,即在设备正常运行时,利用偏最小二乘回归(PLS)方法挖掘状态参量之间的关联关系及回归方程,基于回归预测与实际之间的偏差构建了基于主成分分析(PCA)的变压器状态评判方法,通过控制图对比了变压器状态变化检出的能力。通过对变压器正常运行状态和异常运行状态监测数据进行分析,结果表明:多状态参数回归分析可以将变压器的多元监测数据直观地显示出来;该方法利用监测参数间的相关性,通过发现变压器运行过程中状态参数的霍特林统计量,可以提前检测变压器的异常运行状态。
