期刊文献+
共找到9,880篇文章
< 1 2 250 >
每页显示 20 50 100
融合注意力增强CNN与Transformer的电网关键节点识别
1
作者 黎海涛 乔禄 +2 位作者 杨艳红 谢冬雪 高文浩 《北京工业大学学报》 北大核心 2026年第2期117-129,共13页
为了精确识别电网关键节点以保障电力系统的可靠运行,提出一种基于融合拓扑特征与电气特征的双重自注意力卷积神经网络(convolutional neural network,CNN)的电网关键节点识别方法。首先,构建包含节点的局部拓扑特征、半局部拓扑特征、... 为了精确识别电网关键节点以保障电力系统的可靠运行,提出一种基于融合拓扑特征与电气特征的双重自注意力卷积神经网络(convolutional neural network,CNN)的电网关键节点识别方法。首先,构建包含节点的局部拓扑特征、半局部拓扑特征、电气距离及节点电压的多维特征集;然后,利用压缩-激励(squeeze-and-excitation,SE)自注意力机制改进CNN以增强对节点特征的提取能力,并引入多头自注意力的Transformer编码器以实现拓扑特征与电气特征的深度融合。结果表明:在IEEE 30节点和IEEE 118节点的标准测试系统上,该方法识别关键节点的准确性更高,并且在节点影响力评估和网络鲁棒性方面,得到的电网关键节点对网络的影响更大,鲁棒性更好,为电网的安全稳定运行提供了有效的决策支持。 展开更多
关键词 复杂网络 电网 关键节点识别 卷积神经网络(convolutional neural network cnn) 注意力 特征融合
在线阅读 下载PDF
Infrared spectroscopic analysis of O-H bond dynamics in one-dimensional confined water and bulk water
2
作者 ZHANG Lei WANG Tian-Qi FAN Yan-Ping 《红外与毫米波学报》 北大核心 2025年第1期78-85,共8页
In sub nanometer carbon nanotubes,water exhibits unique dynamic characteristics,and in the high-frequency region of the infrared spectrum,where the stretching vibrations of the internal oxygen-hydrogen(O-H)bonds are c... In sub nanometer carbon nanotubes,water exhibits unique dynamic characteristics,and in the high-frequency region of the infrared spectrum,where the stretching vibrations of the internal oxygen-hydrogen(O-H)bonds are closely related to the hydrogen bonds(H-bonds)network between water molecules.Therefore,it is crucial to analyze the relationship between these two aspects.In this paper,the infrared spectrum and motion characteristics of the stretching vibrations of the O-H bonds in one-dimensional confined water(1DCW)and bulk water(BW)in(6,6)single-walled carbon nanotubes(SWNT)are studied by molecular dynamics simulations.The results show that the stretching vibrations of the two O-H bonds in 1DCW exhibit different frequencies in the infrared spectrum,while the O-H bonds in BW display two identical main frequency peaks.Further analysis using the spring oscillator model reveals that the difference in the stretching amplitude of the O-H bonds is the main factor causing the change in vibration frequency,where an increase in stretching amplitude leads to a decrease in spring stiffness and,consequently,a lower vibration frequency.A more in-depth study found that the interaction of H-bonds between water molecules is the fundamental cause of the increased stretching amplitude and decreased vibration frequency of the O-H bonds.Finally,by analyzing the motion trajectory of the H atoms,the dynamic differences between 1DCW and BW are clearly revealed.These findings provide a new perspective for understanding the behavior of water molecules at the nanoscale and are of significant importance in advancing the development of infrared spectroscopy detection technology. 展开更多
关键词 one-dimensional confined water infrared spectroscopy hydrogen bonds
在线阅读 下载PDF
Density distribution of ground state of one-dimensional Bose gas with dipole interaction
3
作者 Shuchang Hao Yajiang Hao 《Chinese Physics B》 2025年第3期245-249,共5页
Using the Bose-Fermi mapping method,we obtain the exact ground state wavefunction of one-dimensional(1D)Bose gas with the zero-range dipolar interaction in the strongly repulsive contact interaction limit.Its ground s... Using the Bose-Fermi mapping method,we obtain the exact ground state wavefunction of one-dimensional(1D)Bose gas with the zero-range dipolar interaction in the strongly repulsive contact interaction limit.Its ground state density distributions for both repulsive and attractive dipole interactions are exhibited.It is shown that in the case of the finite dipole interaction the density profiles do not change obviously with the increase of dipole interaction and display the typical shell structure of Tonks-Girardeau gases.As the repulsive dipole interaction is greatly strong,the density decreases at the center of the trap and displays a sunken valley.As the attractive dipole interaction increases,the density displays more oscillations and sharp peaks appear in the strong attraction limit,which mainly originate from the atoms occupying the low single particle levels. 展开更多
关键词 Bose gas one-dimensional dipole interaction
原文传递
The advanced development of one-dimensional transition metal dichalcogenide nanotubes:From preparation to application
4
作者 Fengshun Wang Huachao Ji +6 位作者 Zefei Wu Kang Chen Wenqi Gao Chen Wang Longlu Wang Jianmei Chen Dafeng Yan 《Chinese Chemical Letters》 2025年第5期187-197,共11页
Two-dimensional(2D)transition metal sulfides(TMDs)are emerging and highly well received 2D materials,which are considered as an ideal 2D platform for studying various electronic properties and potential applications d... Two-dimensional(2D)transition metal sulfides(TMDs)are emerging and highly well received 2D materials,which are considered as an ideal 2D platform for studying various electronic properties and potential applications due to their chemical diversity.Converting 2D TMDs into one-dimensional(1D)TMDs nanotubes can not only retain some advantages of 2D nanosheets but also providing a unique direction to explore the novel properties of TMDs materials in the 1D limit.However,the controllable preparation of high-quality nanotubes remains a major challenge.It is very necessary to review the advanced development of one-dimensional transition metal dichalcogenide nanotubes from preparation to application.Here,we first summarize a series of bottom-up synthesis methods of 1D TMDs,such as template growth and metal catalyzed method.Then,top-down synthesis methods are summarized,which included selfcuring and stacking of TMDs nanosheets.In addition,we discuss some key applications that utilize the properties of 1D-TMDs nanotubes in the areas of catalyst preparation,energy storage,and electronic devices.Last but not least,we prospect the preparation methods of high-quality 1D-TMDs nanotubes,which will lay a foundation for the synthesis of high-performance optoelectronic devices,catalysts,and energy storage components. 展开更多
关键词 one-dimensional transition metal sulfides NANOTUBES STRUCTURE Preparation method APPLICATIONS
原文传递
A tunable acoustic metasurface via one-dimensional mechanical adjustment for real-time focusing
5
作者 Jie Hu Mengqi Jiang +1 位作者 Rui Zang Yuhang Qian 《Chinese Physics B》 2025年第10期399-404,共6页
Adjustable or programmable metamaterials offer versatile functions,while the complex multi-dimensional regulation increases workload,and hinders their applications in practical scenarios.To address these challenges,we... Adjustable or programmable metamaterials offer versatile functions,while the complex multi-dimensional regulation increases workload,and hinders their applications in practical scenarios.To address these challenges,we present a mechanically programmable acoustic metamaterial for real-time focal tuning via one-dimensional phase-gradient modulation in this paper.The device integrates a phase gradient structure with concave cavity channels and an x-shaped telescopic mechanical framework,enabling dynamic adjustment of inter-unit spacing(1 mm-3 mm)through a microcontroller-driven motor.By modulating the spacing between adjacent channels,the phase gradient is precisely controlled,allowing continuous focal shift from 50 mm to 300 mm along the x-axis at 7500 Hz.Broadband focusing is also discussed in the range6800 Hz-8100 Hz,with transmission coefficients exceeding 0.5,ensuring high efficiency and robust performance.Experimental results align closely with simulations,validating the design's effectiveness and adaptability.Unlike conventional programmable metamaterials requiring multi-dimensional parameter optimization,this approach simplifies real-time control through single-axis mechanical adjustment,significantly reducing operational complexity.Due to the advantages of broadband focusing,simple control mode,real-time monitoring,and so on,the device may have extensive applications in the fields of acoustic imaging,nondestructive testing,ultrasound medical treatment,etc. 展开更多
关键词 phase gradient structure one-dimensional mechanical adjustment adjustable focusing
原文传递
Rational Design of One-Dimensional Bifunctional PBA Nanocomposites as Efficient Electrocatalysts for Oxygen Evolution Reaction
6
作者 Songtao Zhang Yihao Chen +6 位作者 Wenhui Hu Xudong Chen Ziming Qiu Yichun Su Rongmei Zhu Mingbo Zheng Huan Pang 《Transactions of Tianjin University》 2025年第2期179-188,共10页
The oxygen evolution reaction(OER),a critical half-reaction in water electrolysis,has garnered significant attention.However,sluggish OER kinetics has emerged as a major impediment to efficient electrochemical energy c... The oxygen evolution reaction(OER),a critical half-reaction in water electrolysis,has garnered significant attention.However,sluggish OER kinetics has emerged as a major impediment to efficient electrochemical energy conversion.There is an urgent need to design novel electrocatalysts with optimized OER kinetics and enhanced intrinsic activity to improve overall OER performance.Herein,one-dimensional(1D)nanocomposites with high electrocatalytic activity were developed through the deposition of CoFePBA nanocubes onto the surface of MnO_(2) nanowires.The electronic structure of the nanocomposite surface was modified,and the synergistic effects between transition metals were leveraged to enhance catalytic activity through the deposition of Prussian blue analog(PBA)nanocubes on manganese dioxide nanowires.Specifically,CoFePBA featured an open crystal structure that offiered numerous electrochemical active sites and efficient charge transfer pathways.Additionally,the synergistic interactions between Co and Fe significantly reduced the OER overpotential.Additionally,the 1D rigid MnO_(2) acted as protective armor,ensuring the stability of active sites within CoFePBA during the OER.The synthesized MnO_(2)@CoFePBA achieved an overpotential of 1.614 V at 10 mA/cm^(2) and a small Tafel slope of 94 mV/dec and demonstrated stable performance for over 200 h.This work offers new insights into the rational design of various PBA-based nanocomposites with high activity and stability. 展开更多
关键词 one-dimensional materials NANOCOMPOSITES Prussian blue analog(PBA) Synergistic effect ELECTROCATALYSTS
在线阅读 下载PDF
Thermodynamics of classical one-dimensional generalized nonlinear Klein-Gordon lattice model
7
作者 Hu-Wei Jia Ning-Hua Tong 《Chinese Physics B》 2025年第8期381-396,共16页
We study the thermodynamic properties of the classical one-dimensional generalized nonlinear Klein-Gordon lattice model(n≥2)by using the cluster variation method with linear response theory.The results of this method... We study the thermodynamic properties of the classical one-dimensional generalized nonlinear Klein-Gordon lattice model(n≥2)by using the cluster variation method with linear response theory.The results of this method are exact in the thermodynamic limit.We present the single-site reduced densityρ^((1))(z),averages such as(z^(2)),<|z^(n)|>,and<(z_(1)-z_(2))^(2)>,the specific heat C_(v),and the static correlation functions.We analyze the scaling behavior of these quantities and obtain the exact scaling powers at the low and high temperatures.Using these results,we gauge the accuracy of the projective truncation approximation for theφ^(4)lattice model. 展开更多
关键词 cluster variation method linear response theory one-dimensional generalized nonlinear Klein-Gordon lattice model
原文传递
面向可重构结构的CNN模型混合压缩方法
8
作者 刘朋飞 蒋林 +1 位作者 李远成 吴海 《现代电子技术》 北大核心 2026年第1期167-173,共7页
随着卷积神经网络规模的不断扩大,其参数量和计算量显著增加,导致硬件面临严重的访存瓶颈,限制了计算效率。为解决这一问题,文中提出一种面向可重构结构的CNN混合压缩新方法,该方法采用先剪枝后量化的策略,通过基于一阶泰勒展开的滤波... 随着卷积神经网络规模的不断扩大,其参数量和计算量显著增加,导致硬件面临严重的访存瓶颈,限制了计算效率。为解决这一问题,文中提出一种面向可重构结构的CNN混合压缩新方法,该方法采用先剪枝后量化的策略,通过基于一阶泰勒展开的滤波器剪枝、基于阈值的全连接层权值剪枝和混合精度自适应量化策略,来减少模型参数量和计算复杂度,并部署在自研的可重构处理器上。实验结果表明,所提方法在VGG16和ResNet18模型上分别实现了31.4倍和7.9倍的压缩比,精度仅下降1.20%和0.74%。在基于VirtexUltraScale VU440 FPGA开发板搭建的可重构阵列处理器上,压缩后的VGG16模型执行周期最大降低了62.7%。证明所提方法适合资源有限的边缘计算设备。 展开更多
关键词 卷积神经网络 模型压缩 结构化剪枝 自适应量化 并行计算 可重构结构
在线阅读 下载PDF
Ultraprecision and highly uniform nanoscale conformality enabled by introducing oxide buffer layer for one-dimensional grating standard
9
作者 Ya-Xin Zhang Song Wang +9 位作者 Chen-Ying Wang Yi-Fan Zhao Feng Han Di Liu Peng-Cheng Zhang Nan Zhu Kun Zheng Wei Ren Wei-Xuan Jing Zhuang-De Jiang 《Rare Metals》 2025年第7期4779-4788,共10页
One-dimensional nano-grating standard(ODNGS)is widely recognized as a crucial nanometric standard for metrological technology.However,achieving the ultratiny size of ODNGS with high consistent uniformity and low rough... One-dimensional nano-grating standard(ODNGS)is widely recognized as a crucial nanometric standard for metrological technology.However,achieving the ultratiny size of ODNGS with high consistent uniformity and low roughness by conventional processes such as the inductively coupled plasma(ICP)etching methodpresents a significant challenge in obtaining accurate calibration values.In this work,a 50-nm ODNGS with a conformal buffer layer(Al_(2)O_(3))is successfully obtained,indicating outstanding stability and abrasion resistance.Remarkably,the introduction of hydrogen silsesquioxane(HSQ)and amorphous Al_(2)O_(3)simultaneously guarantees an incredibly small expanded uncertainty(0.5 nm)and repeatability of the standard uniformity(less than 0.3 nm)in the grating dimensions.TheⅠ-Ⅴcurves of ODNGS with an Al_(2)O_(3)buffer layer at room temperature(RT)and200℃are depicted respectively to showcase the sustained favorable insulation properties.Notably,the nanostructure fluctuation,line edge roughness(LER)and line width roughness(LWR)of the standard can be decreased obviously by 64.1%,63%and 70%,respectively.Our results suggest that the ODNGS with Al_(2)O_(3)exhibits exceptional precision and robust calibration reliability for calibrating nanoscale measuring instruments.It holds tremendous potential for manufacturing high-precision nanostructures and grating arrays with precisely controllable dimensions,which will play a pivotal role in the fabrication of microfluidics chips,metasurface and photodetectors in the future. 展开更多
关键词 Nanoscale conformality Calibration one-dimensional grating standard Al_(2)O_(3)conformality
原文传递
Modulating electronic properties of carbon nanotube via constructing one-dimensional vdW heterostructures
10
作者 Wenqi Lv Weili Li +1 位作者 Wei Ji Yanning Zhang 《Chinese Physics B》 2025年第6期512-516,共5页
Controlling charge polarity in the semiconducting single-walled carbon nanotubes(CNTs) by substitutional doping is a difficult work due to their extremely strong C–C bonding. In this work, an inner doping strategy is... Controlling charge polarity in the semiconducting single-walled carbon nanotubes(CNTs) by substitutional doping is a difficult work due to their extremely strong C–C bonding. In this work, an inner doping strategy is explored by filling CNTs with one-dimensional(1D)-TM_(6)Te_(6) nanowires to form TM_(6)Te_(6)@CNT-(16,0) 1D van der Waals heterostructures(1D-vd WHs). The systematic first-principles studies on the electronic properties of 1D-vd WHs show that N-type doping CNTs can be formed by charge transfer from TM_(6)Te_(6) nanowires to CNTs, without introducing additional carrier scattering.Particularly, contribution from both T M(e.g., Sc and Y) and Te atoms strengthens the charge transfer. The outside CNTs further confine the dispersion of Te-p orbitals in nanowires that deforms the C-π states at the bottom of the conduction band to quasi sp^(3) hybridization. Our study provides an inner doping strategy that can effectively confine the charge polarity of CNTs and further broaden its applications in some novel nano-devices. 展开更多
关键词 electronic modification of CNTs one-dimensional(1D)vdW heterostructures inner doping density functional theory
原文传递
On the interfacial behavior of a one-dimensional hexagonal piezoelectric quasicrystal film based on the beam theory
11
作者 Wenkai ZHANG C.S.LU +2 位作者 Minghao ZHAO Cuiying FAN Huayang DANG 《Applied Mathematics and Mechanics(English Edition)》 2025年第2期289-304,共16页
In this paper,the mechanical response of a one-dimensional(1D)hexagonal piezoelectric quasicrystal(PQC)thin film is analyzed under electric and temperature loads.Based on the Euler-Bernoulli beam theory,a theoretical ... In this paper,the mechanical response of a one-dimensional(1D)hexagonal piezoelectric quasicrystal(PQC)thin film is analyzed under electric and temperature loads.Based on the Euler-Bernoulli beam theory,a theoretical model is proposed,resulting in coupled governing integral equations that account for interfacial normal and shear stresses.To numerically solve these integral equations,an expansion method using orthogonal Chebyshev polynomials is employed.The results provide insights into the interfacial stresses,axial force,as well as axial and vertical deformations of the PQC film.Additionally,fracture criteria,including stress intensity factors,mode angles,and the J-integral,are evaluated.The solution is compared with the membrane theory,neglecting the normal stress and bending deformation.Finally,the effects of stiffness and aspect ratio on the PQC film are thoroughly discussed.This study serves as a valuable guide for controlling the mechanical response and conducting safety assessments of PQC film systems. 展开更多
关键词 one-dimensional(1D)hexagonal piezoelectric quasicrystal(PQC)film beam theory electric and temperature loads Chebyshev polynomial interfacial behavior
在线阅读 下载PDF
基于改进CNN的煤矿掘进工作面超前探测异常体识别方法
12
作者 原野 《煤矿现代化》 2026年第1期162-165,共4页
煤矿掘进工作面超前探测中,异常体识别的探测技术存在局限性,导致精度不足。传统的探测方法,如某些物理探测手段,受巷道空间的限制,其探测范围和精度相对有限。在处理复杂地质条件时,无法准确识别异常体的位置和形态。为此,开展基于改... 煤矿掘进工作面超前探测中,异常体识别的探测技术存在局限性,导致精度不足。传统的探测方法,如某些物理探测手段,受巷道空间的限制,其探测范围和精度相对有限。在处理复杂地质条件时,无法准确识别异常体的位置和形态。为此,开展基于改进卷积神经网络(CNN)的识别方法研究。通过预处理声波远距离超前物探数据,包括去噪、增强和归一化等步骤,提升数据质量。利用基于改进CNN的模型对探测图像进行异常体特征提取,该模型通过优化卷积层、引入注意力机制和调整超参数,有效提高了特征提取的准确性和鲁棒性。最后,基于提取的特征向量,采用SVM分类器实现异常体的识别分类。通过对比实验证明,该方法相较于现有方法在异常体识别准确率和效率有显著提升。 展开更多
关键词 改进cnn 掘进 超前探测 异常体识别
在线阅读 下载PDF
基于CNN与格栅优化的输变电工程造价趋势预测研究
13
作者 孙永彦 丁艳 +1 位作者 叶慧男 王天佑 《粘接》 2026年第2期552-555,共4页
设计要求的严格与多变常导致工程造价的增加,进而使评价指标不稳定。为此,提出采用卷积神经网络(Convolutional Neural Network,CNN)与格栅优化的输变电工程造价趋势预测方法。首先,利用数学方法估算输变电工程造价成本,并获取相应的造... 设计要求的严格与多变常导致工程造价的增加,进而使评价指标不稳定。为此,提出采用卷积神经网络(Convolutional Neural Network,CNN)与格栅优化的输变电工程造价趋势预测方法。首先,利用数学方法估算输变电工程造价成本,并获取相应的造价数据,对原始造价数据进行标准化处理,选取施工造价的预测评价指标,利用熵权法筛选特征,随后计算施工造价指标的特征相似度,构建神经网络预测模型,并引入均方误差(MSE)作为损失函数进行训练,从而建立了输变电工程造价预测模型。基于CNN特征提取与格栅优化算法计算关联度,以预测输变电工程造价趋势,最后,采用均值强化算法对造价数据进行平均化处理,所得平均值即为最终的工程造价预测值。 展开更多
关键词 输变电工程 造价预测 cnn 格栅优化 特征提取
在线阅读 下载PDF
基于CNN-BiLSTM-SSA的锅炉再热器壁温预测模型
14
作者 徐世明 何至谦 +6 位作者 彭献永 商忠宝 范景玮 王俊略 曲舒杨 刘洋 周怀春 《动力工程学报》 北大核心 2026年第1期121-130,共10页
针对锅炉高温再热器壁温动态特点,提出了一种基于稀疏自注意力(SSA)、卷积神经网络(CNN)及双向长短期记忆神经网络(BiLSTM)相融合的再热器壁温软测量模型。首先,采用核主成分分析(KPCA)算法对原始候选变量进行筛选降维,选择前26个主成... 针对锅炉高温再热器壁温动态特点,提出了一种基于稀疏自注意力(SSA)、卷积神经网络(CNN)及双向长短期记忆神经网络(BiLSTM)相融合的再热器壁温软测量模型。首先,采用核主成分分析(KPCA)算法对原始候选变量进行筛选降维,选择前26个主成分变量作为模型的最终输入。其次,考虑利用CNN捕捉局部相关性,BiLSTM学习数据的长期序列依赖性的优势,使用卷积神经网络-双向长短期记忆神经网络(CNN-BiLSTM)捕捉时序数据中的短期和长期依赖关系,引入稀疏自注意力SSA机制,通过为不同特征部分分配自适应权重,从而增强CNN-BiLSTM模型的特征提取与建模能力,最后利用在役1000 MW超超临界锅炉的历史数据进行仿真实验。结果表明:CNN-BiLSTM-SSA模型在高温再热器壁温预测中的均方根误差(RMSE)、平均绝对误差(MAE)及平均绝对百分比误差(MAPE)分别为4.92℃、3.81℃和0.6241%,相应的指标均优于CNN、LSTM、BiLSTM、CNN-LSTM和CNN-BiLSTM模型。 展开更多
关键词 再热器壁温软测量 深度学习 卷积神经网络 长短期记忆网络 注意力机制 核主成分分析 cnn-BiLSTM
在线阅读 下载PDF
Reconstruction of pile-up events using a one-dimensional convolutional autoencoder for the NEDA detector array
15
作者 J.M.Deltoro G.Jaworski +15 位作者 A.Goasduff V.González A.Gadea M.Palacz J.J.Valiente-Dobón J.Nyberg S.Casans A.E.Navarro-Antón E.Sanchis G.de Angelis A.Boujrad S.Coudert T.Dupasquier S.Ertürk O.Stezowski R.Wadsworth 《Nuclear Science and Techniques》 2025年第2期62-70,共9页
Pulse pile-up is a problem in nuclear spectroscopy and nuclear reaction studies that occurs when two pulses overlap and distort each other,degrading the quality of energy and timing information.Different methods have ... Pulse pile-up is a problem in nuclear spectroscopy and nuclear reaction studies that occurs when two pulses overlap and distort each other,degrading the quality of energy and timing information.Different methods have been used for pile-up rejection,both digital and analogue,but some pile-up events may contain pulses of interest and need to be reconstructed.The paper proposes a new method for reconstructing pile-up events acquired with a neutron detector array(NEDA)using an one-dimensional convolutional autoencoder(1D-CAE).The datasets for training and testing the 1D-CAE are created from data acquired from the NEDA.The new pile-up signal reconstruction method is evaluated from the point of view of how similar the reconstructed signals are to the original ones.Furthermore,it is analysed considering the result of the neutron-gamma discrimination based on charge comparison,comparing the result obtained from original and reconstructed signals. 展开更多
关键词 1D-CAE Autoencoder CAE Convolutional neural network(cnn) Neutron detector Neutron-gamma discrimination(NGD) Machine learning Pulse shape discrimination Pile-up pulse
在线阅读 下载PDF
Interfacial analysis of a penny-shaped one-dimensional hexagonal functionally graded piezoelectric quasicrystal film on a temperature-dependent substrate
16
作者 Kai LUO Cuiying FAN +2 位作者 Minghao ZHAO CSLU Huayang DANG 《Applied Mathematics and Mechanics(English Edition)》 2025年第12期2297-2316,共20页
In this paper,we investigate the interfacial behavior of a thin,penny-shaped,one-dimensional(1D)hexagonal functionally graded(FG)piezoelectric quasicrystal(PQC)film bonded on a temperature-dependent elastic substrate ... In this paper,we investigate the interfacial behavior of a thin,penny-shaped,one-dimensional(1D)hexagonal functionally graded(FG)piezoelectric quasicrystal(PQC)film bonded on a temperature-dependent elastic substrate under thermal and electrical loads.The problem is modeled as axisymmetric based on the membrane theory,with the peeling stress and bending moment being disregarded.A potential theory method,combined with the Hankel transform technique,is utilized to derive the displacement field on the substrate surface.With perfect interfacial bonding assumption,an integral equation governing the phonon interfacial shear stress is formulated and numerically solved by the Chebyshev polynomials.Explicit expressions are derived for the interfacial shear stress,the internal stresses within the PQC film and the substrate,the axial strain,and the stress intensity factors(SIFs).Numerical simulations are conducted to investigate the effects of the film's aspect ratio,material inhomogeneity,material mismatch,and temperature-dependent material properties on its mechanical response.The results provide insights for the functional design and reliability assessment of FG PQC film/substrate systems. 展开更多
关键词 one-dimensional(1D)hexagonal piezoelectric quasicrystal(PQC)film functionally graded(FG)material three-dimensional(3D)interfacial behavior stress intensity factor(SIF) temperature-dependent material property
在线阅读 下载PDF
基于多维故障特征提取的CNN-BiGRU-ATT多分支配电网故障定位
17
作者 张玉敏 王德龙 +4 位作者 张晓 吉兴全 张祥星 黄心月 王学林 《中国电力》 北大核心 2026年第1期163-174,共12页
针对多分支配电网故障定位在微弱故障条件下故障特征提取困难的问题,提出了基于多维故障特征提取的卷积神经网络(convolution neural network,CNN)-双向门控循环单元(bidirectional gated recurrent unit,BiGRU)-注意力机制(attention m... 针对多分支配电网故障定位在微弱故障条件下故障特征提取困难的问题,提出了基于多维故障特征提取的卷积神经网络(convolution neural network,CNN)-双向门控循环单元(bidirectional gated recurrent unit,BiGRU)-注意力机制(attention mechanism,ATT)多分支配电网故障定位方法。首先,分析不同故障位置和故障分支的行波特性,采用基于直线检测(line segment detector,LSD)的波头标定方法提取故障波头的坐标、幅值和斜率等信息,利用主成分分析法(principal component analysis,PCA)构造与故障位置成映射关系的多维故障特征空间;其次,构建CNN-BiGRU-ATT故障定位模型,深入挖掘时序特征和幅值特征与故障位置之间的关联;最后,结合分类与回归任务,分别实现故障区段定位与精准定位。在有限样本的情况下,区段定位准确率达99.6429%,精准定位误差55.77 m,跨工况误差最低2.95 m。结果表明,该模型能有效关联多维故障特征与故障信息,较对比模型具有更优的故障定位精度稳定性与场景泛化能力。 展开更多
关键词 故障定位 多分支配电网 LSD 多维故障特征 cnn-BiGRU-ATT
在线阅读 下载PDF
基于条件高斯PAC-Bayes的机载CNN分类器安全性评估 被引量:1
18
作者 马赞 白杰 +2 位作者 陈勇 刘瑞华 张艳婷 《航空学报》 北大核心 2025年第4期217-230,共14页
针对机器学习技术的固有不确定输出特性给航空器适航安全性定量评估造成的挑战,在SAE ARP4761标准框架下,基于条件高斯PAC-Bayes泛化理论提出一种面向卷积神经网络(CNN)分类功能的系统安全性评估方法。首先,基于PAC-Bayes理论,通过条件... 针对机器学习技术的固有不确定输出特性给航空器适航安全性定量评估造成的挑战,在SAE ARP4761标准框架下,基于条件高斯PAC-Bayes泛化理论提出一种面向卷积神经网络(CNN)分类功能的系统安全性评估方法。首先,基于PAC-Bayes理论,通过条件高斯分布改进训练方法,优化泛化界,获取CNN模型不确定性量化表示。其次,提出一种基于泛化界置信度的软件不确定性与硬件可靠性融合方法,获取CNN部件的综合失效基础数据,支持整机/系统的定量安全性评估。最后,以基于CNN的全球导航卫星系统干扰信号识别模块装机为案例,表明该方法对适航安全性评估的有效支撑作用,为CNN技术的装机应用提供了必要的适航符合性保证。同时也实验验证基于条件高斯的方法比标准PAC-Bayes及Vapnik-Chervonenkis维都具有更紧的计算边界。 展开更多
关键词 机载cnn分类器 PAC-Bayes SAE ARP4761 条件高斯 适航安全性
原文传递
基于CNN-LSTM-Attention 组合模型的黄金周旅游客流预测——以大理州为例 被引量:2
19
作者 戢晓峰 郭雅诗 +2 位作者 陈方 黄志文 李武 《干旱区资源与环境》 北大核心 2025年第3期200-208,共9页
黄金周旅游客流预测一直是区域旅游管理的重大现实需求,能够为黄金周旅游组织提供更为精准的数据支持。文中基于百度迁徙数据和百度搜索指数数据,以卷积神经网络(CNN)、长短期记忆网络(LSTM)以及注意力机制(Attention)为基准,构建了CNN-... 黄金周旅游客流预测一直是区域旅游管理的重大现实需求,能够为黄金周旅游组织提供更为精准的数据支持。文中基于百度迁徙数据和百度搜索指数数据,以卷积神经网络(CNN)、长短期记忆网络(LSTM)以及注意力机制(Attention)为基准,构建了CNN-LSTM-Attention组合模型,对大理州黄金周日度旅游客流人数进行了预测,并基于SHAP算法进行了影响因素分析。结果显示:1)CNN-LSTM-Attention组合模型的预测精度优于RF模型、SVM模型、CNN模型、LSTM模型和CNN-LSTM模型。2)引入百度搜索指数特征后,模型的均方根误差(RMSE)、平均绝对百分比误差(MAPE)、决定系数(R^(2))表现最优,表明百度搜索指数的加入在一定程度上提升了模型的预测精度。文中所构模型为黄金周旅游客流预测提供了新思路。 展开更多
关键词 客流预测 黄金周 卷积神经网络(cnn) 长短期记忆网络(LSTM) 注意力机制
原文传递
基于CNN模型的地震数据噪声压制性能对比研究 被引量:1
20
作者 张光德 张怀榜 +3 位作者 赵金泉 尤加春 魏俊廷 杨德宽 《石油物探》 北大核心 2025年第2期232-246,共15页
地震噪声的压制是地震勘探中地震数据处理的重要研究内容之一。准确地压制地震噪声和提取地震信号中的有效信息是地震勘探和地震监测的一项关键步骤。传统的地震噪声压制方法存在一些不足之处,如灵活性不足、难以处理复杂噪声、有效信... 地震噪声的压制是地震勘探中地震数据处理的重要研究内容之一。准确地压制地震噪声和提取地震信号中的有效信息是地震勘探和地震监测的一项关键步骤。传统的地震噪声压制方法存在一些不足之处,如灵活性不足、难以处理复杂噪声、有效信息损失以及依赖人工提取特征等局限性。为克服传统方法的不足,采用时频域变换并结合深度学习方法进行地震噪声压制,并验证其应用效果。通过构建5个神经网络模型(FCN、Unet、CBDNet、SwinUnet以及TransUnet)对经过时频变换的地震信号进行噪声压制。为了定量评估实验方法的去噪性能,引入了峰值信噪比(PSNR)、结构相似性指数(SSIM)和均方根误差(RMSE)3个指标,比较不同方法的噪声压制性能。数值实验结果表明,基于时频变换的卷积神经网络(CNN)方法对常见的地震噪声类型(包括随机噪声、海洋涌浪噪声、陆地面波噪声)具有较好的噪声压制效果,能够提高地震数据的信噪比。而Transformer模块的引入可进一步提高对上述3种常见地震数据噪声类型的压制效果,进一步提升CNN模型的去噪性能。尽管该方法在数值实验中取得了较好的应用效果,但仍有进一步优化的空间可供探索,比如改进网络结构以适应更复杂的地震信号,并探索与其他先进技术结合,以提升地震噪声压制性能。 展开更多
关键词 地震噪声压制 深度学习 卷积神经网络(cnn) 时频变换 TRANSFORMER
在线阅读 下载PDF
上一页 1 2 250 下一页 到第
使用帮助 返回顶部