期刊文献+
共找到7,903篇文章
< 1 2 250 >
每页显示 20 50 100
Spreading rate dependence of morphological characteristics in global oceanic transform faults 被引量:3
1
作者 Yiming Luo Jian Lin +1 位作者 Fan Zhang Meng Wei 《Acta Oceanologica Sinica》 SCIE CAS CSCD 2021年第4期39-64,共26页
We quantified the systematic variations in global transform fault morphology,revealing a first-order dependence on the spreading rate.(1)The average age offset of both the full transform and transform sub-segments dec... We quantified the systematic variations in global transform fault morphology,revealing a first-order dependence on the spreading rate.(1)The average age offset of both the full transform and transform sub-segments decrease with increasing spreading rate.(2)The average depth of both the transform valley and adjacent ridges are smaller in the fast compared to the slow systems,reflecting possibly density anomalies associated with warmer mantle at the fast systems and rifting at the slow ridges.However,the average depth difference between the transform valley and adjacent ridges is relatively constant from the fast to slow systems.(3)The nodal basin at a ridge-transform intersection is deeper and dominant at the ultraslow and slow systems,possibly reflecting a lower magma supply and stronger viscous resistance to mantle upwelling near a colder transform wall.In contrast,the nodal high,is most prominent in the fast,intermediate,and hotspot-influenced systems,where robust axial volcanic ridges extend toward the ridge-transform intersection.(4)Statistically,the average transform valley is wider at a transform system of larger age offset,reflecting thicker deforming plates flanking the transform fault.(5)The maximum magnitude of the transform earthquakes increases with age offset owing to an increase in the seismogenic area.Individual transform faults also exhibit significant anomalies owing to the complex local tectonic and magmatic processes. 展开更多
关键词 mid-ocean ridge transform fault MORPHOLOGY spreading rate transform earthquakes
在线阅读 下载PDF
Microearthquake reveals the lithospheric structure at midocean ridges and oceanic transform faults 被引量:1
2
作者 Zhiteng YU Jiabiao LI Weiwei DING 《Journal of Oceanology and Limnology》 SCIE CAS CSCD 2024年第3期697-700,共4页
Mid-ocean ridge and oceanic transforms are among the most prominent features on the seafloor surface and are crucial for understanding seafloor spreading and plate tectonic dynamics,but the deep structure of the ocean... Mid-ocean ridge and oceanic transforms are among the most prominent features on the seafloor surface and are crucial for understanding seafloor spreading and plate tectonic dynamics,but the deep structure of the oceanic lithosphere remains poorly understood.The large number of microearthquakes occurring along ridges and transforms provide valuable information for gaining an indepth view of the underlying detailed seismic structures,contributing to understanding geodynamic processes within the oceanic lithosphere.Previous studies have indicated that the maximum depth of microseismicity is controlled by the 600-℃isotherm.However,this perspective is being challenged due to increasing observations of deep earthquakes that far exceed this suggested isotherm along mid-ocean ridges and oceanic transform faults.Several mechanisms have been proposed to explain these deep events,and we suggest that local geodynamic processes(e.g.,magma supply,mylonite shear zone,longlived faults,hydrothermal vents,etc.)likely play a more important role than previously thought. 展开更多
关键词 microearthquake mid-ocean ridge oceanic transform fault oceanic lithosphere thermal structure earthquake location
在线阅读 下载PDF
CENOZOIC ALTYN TRANSFORM FAULT OF THE NORTHERN PART OF THE TIBETAN PLATEAU
3
作者 Wang Genhou,Gao Jinhan,Wang Xiaoniu (China University of Geosciences, Beijing 100083,China) 《地学前缘》 EI CAS CSCD 2000年第S1期159-160,共2页
The transform fault is essentially a displacement fault whose terminal part is adjusted by other tectonic types, its displacement component is absorbed by other structures intersected with it by high angles or meet at... The transform fault is essentially a displacement fault whose terminal part is adjusted by other tectonic types, its displacement component is absorbed by other structures intersected with it by high angles or meet at right angles. The main elements of transform fault are the sleep\|dipping displacement faults and the adjusted structures intersected with it at high angles. According to the combination of tectonic features formed by its two ends of displacement fault and the structures intersected with it, the transform fault can be divided into three types, including the adjusted transform fault of extensional normal fault, the adjusted transform fault of compressive fold and thrust fault, and the compound transform fault. The transform fault is different from the displacement fault, its horizontal displacement may be increased or decreased or not be changed at all as the time of fault movement extended, but for parallel displacement the dislocation will be increased. Therefore, the study of transform fault is very important for the recognition of long time disputed displacement components of huge displacement fault. The traditional Altyn fault is the adjusting fault of the compression deformation of the Western Kunlun and Northern Qilian mountains of the northern margin of the Tibetan Plateau since Cenozoic. 展开更多
关键词 transform fault Altyn TIBETAN PLATEAU adjustment DISPLACEMENT fault
在线阅读 下载PDF
Transform faults and transfer faults:Plate boundary and intra-continental tectonic dynamics transition
4
作者 Yunpeng DONG Bo HUI +6 位作者 Bo WANG Shengsi SUN Huajian YAO Junlai LIU Rutao ZANG Yongcheng LI Qingxing LUO 《Science China Earth Sciences》 2025年第12期3867-3911,共45页
Large strike-slip faults play a crucial role in regulating the geometry and kinematics of the solid Earth's outer lithospheric plates and the structural deformation of internal continents.They not only control the... Large strike-slip faults play a crucial role in regulating the geometry and kinematics of the solid Earth's outer lithospheric plates and the structural deformation of internal continents.They not only control the geometric structures,motion properties,and direction of the lithospheric plates,but also regulate the complex tectonic stress field and strain state caused by differential motion among multiple blocks within the continent,maintaining the relative stability of the overall stress state of the lithosphere on the Earth's surface.According to the nature and significance of geotectonic structures,strike-slip faults can be divided into interplate types and intraplate tectonic types.Interplate strike-slip faults are transform faults,including oceanic transform faults and continental transform faults.Intraplate strike-slip faults can be divided into continental transfer faults and intraplate transcurrent faults.During the lateral movement of lithospheric plates along the Earth's surface,transform faults adjust the differences in the nature,direction,and rate of movement between different plates.Meanwhile,continental transfer faults and intraplate transcurrent faults adjust the location,nature,style,and differential stress of intraplate tectonic deformation.Strike-slip faults of varying types and scales interact in different ways to maintain the dynamic balance of matter and energy within Earth's lithospheric plates.Based on the concepts,tectonic significance,and recent research advances of strike-slip faults and classical transform faults,this paper summarizes the latest classification of strike-slip faults and their corresponding tectonic implications.It also updates the definitions,geometric characteristics,and kinematic features of oceanic transform faults,continental transform faults,continental transfer faults,and intraplate transcurrent faults.Through typical global examples,this paper comprehensively analyzes the deep structure,structural geometry and kinematic characteristics,evolution process,geological significance,and seismic hazards of different types of strike-slip faults.Furthermore,the frontier science issue and research strategies for the study of oceanic transform faults,continental transfer faults,and intraplate transcurrent faults are summarized as well. 展开更多
关键词 Oceanic transform fault Continental transform fault Continental transfer fault Strike-slip fault Plate tectonics and continental dynamics
原文传递
Lithospheric structure and geodynamics of oceanic transform fault
5
作者 Fan ZHANG Jian LIN +5 位作者 Sanzhong LI Yanhui SUO Pengcheng ZHOU Yiming LUO Caicai ZHA Xubo ZHANG 《Science China Earth Sciences》 2025年第12期3829-3845,共17页
The oceanic transform fault(OTF)is one of the three types of plate boundaries.It provides an important channel for the exchange of material and energy within the Earth's interior,which is significant for understan... The oceanic transform fault(OTF)is one of the three types of plate boundaries.It provides an important channel for the exchange of material and energy within the Earth's interior,which is significant for understanding global plate tectonics.In recent years,substantial progress has been made in the study of OTFs,including their morphology,crustal accretion mode,stress and strain state,brittle-ductile deformation structures,and segmentation and symmetry of seismicity.The combined effects of magmatism and tectonism shape the morphology and structure of OTF:the spreading rate and the age offsets reflect the thermal structure of the OTF,and thus affect the intensity of magmatism;meanwhile,the stress within the lithosphere and plate motion control the tectonic features and formation of faults.Three-dimensional dynamic models have demonstrated that increased magmatism at mid-ocean ridges destabilizes the fault,thereby facilitating the dynamic evolution of the OTF.Moreover,the maximum depth of earthquakes on OTF is controlled by the thermal structure of the lithosphere,which is crucial for characterizing the frictional properties of faults and understanding their seismic behaviors.Based on recent comprehensive research findings,this paper reviews the tectonic features,three-dimensional morphological structure,and lithospheric thermal structure of OTF,and discusses important scientific issues,including the magmatic-tectonic co-evolution and geodynamic mechanisms of OTF.Future research will combine high-resolution observations and theoretical simulations to further elucidate the processes and mechanisms of OTF,providing important advances to global plate tectonic theory. 展开更多
关键词 Mid-ocean ridge Oceanic transform fault GEODYNAMICS Magma supply Thermal evolution Numerical modeling
原文传递
Structure and deformation of continental transform faults
6
作者 Jikun FENG Huajian YAO +1 位作者 Yunpeng DONG Yang LI 《Science China Earth Sciences》 2025年第12期3846-3866,共21页
Transform faults represent one of the three primary types of plate boundaries in plate tectonics theory and constitute an essential component of this framework.In general,they are classified into oceanic and continent... Transform faults represent one of the three primary types of plate boundaries in plate tectonics theory and constitute an essential component of this framework.In general,they are classified into oceanic and continental transform faults based on the nature of their separated plates.Owing to significant differences in properties between continental and oceanic lithospheres,continental transform faults exhibit more complex structures than their oceanic counterparts.Continental transform faults are strike-slip boundaries where stress and strain are highly concentrated.They typically extend for hundreds to thousands of kilometers and have experienced tens to hundreds of kilometers of strike-slip displacement.These faults may appear as a single master fault or as complex fault systems with multiple branches.Their deep structures and deformation patterns at varying depths offer critical insights into the structure and rheological behavior of the continental lithosphere.Imaging fine-scale structures of continental transform faults via geophysical methods is crucial for understanding their nature and evolution.Seismic anisotropy results provide key constraints on their deep deformation characteristics.This paper reviews geophysical studies from typical continental transform fault regions and investigates their deep structure and deformation mechanisms by integrating geological and geodetic observations.Although these fault systems are structurally diverse,several common features emerge.(1)Nearly all continental transform faults cut through the entire crust and extend into the upper mantle,with significant seismic anisotropy observed within the fault zones.(2)Regardless of whether the fault is a single narrow structure or a branching system,uppercrustal segments typically form narrow zones of strain concentration where brittle friction accommodates slip and seismicity is concentrated.The shear zone broadens with depth,reaching tens of kilometers in width within the lithospheric mantle.(3)The width of a continental transform fault correlates with the nature of the lithosphere it cross-cuts.Narrow shear zones form in rigid and ancient lithosphere,otherwise,broader distributed deformation occurs.(4)Non-strike-slip components(compression or tension)significantly influence fault zone complexity.Recent ocean drilling programs have advanced understanding of oceanic transform faults,yet knowledge of continental transform fault structure and evolution remains limited.Advances in seismic imaging and observational techniques will enable higher-resolution characterization of these faults,providing new constraints on their seismic behavior and earthquake migration patterns. 展开更多
关键词 Continental transform fault Strike-slip fault fault zone structure Seismic anisotropy Structural deformation
原文传递
基于改进Transformer网络的新型电力系统故障诊断技术研究
7
作者 张超 《自动化应用》 2026年第5期150-152,共3页
针对新型电力系统故障诊断的复杂性和实时性需求,提出了一种基于层次化注意力机制的Transformer故障诊断方法。该方法采用局部-全局双分支协同学习架构,对电压、电流、阻抗、功率、频率偏差、谐波含量、负荷率、设备温度这8项关键参数... 针对新型电力系统故障诊断的复杂性和实时性需求,提出了一种基于层次化注意力机制的Transformer故障诊断方法。该方法采用局部-全局双分支协同学习架构,对电压、电流、阻抗、功率、频率偏差、谐波含量、负荷率、设备温度这8项关键参数进行深度挖掘,实现对短路、接地、过载、绝缘、操作故障这5类典型故障的准确识别。仿真结果表明,在满载运行工况下,所提方法对这5类故障的诊断准确率分别达到96.27%,94.58%,92.56%,94.28%,94.67%。系统响应时间均处于毫秒级别,能满足配电网实时故障处理对时延的严格要求,为新型电力系统智能运维提供了可靠的技术支撑。 展开更多
关键词 电力系统 多故障诊断 改进transformer 故障定位
在线阅读 下载PDF
自适应Transformer-LSTM的滚动轴承故障预测方法研究
8
作者 董辛旻 职帅轩 崔富源 《重庆理工大学学报(自然科学)》 北大核心 2026年第2期96-104,共9页
针对滚动轴承故障预测中,模型预测精度低、参数难以确定等问题,提出一种自适应Transformer编码器与长短期记忆神经网络(long short-term memory, LSTM)的滚动轴承故障预测方法。提取轴承全周期振动信号24维时频特征,并基于信息熵、鲁棒... 针对滚动轴承故障预测中,模型预测精度低、参数难以确定等问题,提出一种自适应Transformer编码器与长短期记忆神经网络(long short-term memory, LSTM)的滚动轴承故障预测方法。提取轴承全周期振动信号24维时频特征,并基于信息熵、鲁棒性和相关性构建综合指标对特征进行筛选。采用核主元分析将优选特征进行融合,构建能够反应轴承退化状态的退化指标。综合时间序列全局特性与长短期特性,构建Transformer编码器与LSTM结合的故障预测模型,并引入极光优化算法(polar lights optimization, PLO)进行参数自适应寻优,应用XJTU-SY与IMS滚动轴承数据集进行方法分析验证。结果表明:所提模型在轴承故障预测中表现出良好效果,相较于经典的时域卷积网络(temporal convolutional network, TCN)、双向门控循环神经网络(gate recurrent unit, GRU)等,预测精度有较大提升。 展开更多
关键词 滚动轴承 故障预测 数据驱动 transformER 优化算法
在线阅读 下载PDF
基于CNN-Transformer-Cross Attention的滚动轴承故障诊断
9
作者 郑文超 张梅 《煤矿机械》 2026年第4期188-192,共5页
滚动轴承是煤机核心部件,若发生故障,易导致停机与安全风险。提出了一种融合快速傅里叶变换(FFT)、卷积神经网络(CNN)、Transformer及Cross Attention的故障诊断方法。该方法首先通过FFT提取频率特征,随后结合CNN的局部特征提取能力、Tr... 滚动轴承是煤机核心部件,若发生故障,易导致停机与安全风险。提出了一种融合快速傅里叶变换(FFT)、卷积神经网络(CNN)、Transformer及Cross Attention的故障诊断方法。该方法首先通过FFT提取频率特征,随后结合CNN的局部特征提取能力、Transformer的全局建模能力及Cross Attention的信息融合能力,全面提升模型的识别能力,实现滚动轴承故障的精确识别。实验结果表明,该方法的故障诊断准确率可达98%,具有高精度、强鲁棒性的特点,适用于煤矿设备的智能运维。 展开更多
关键词 轴承 故障诊断 FFT CNN transformER Cross Attention
原文传递
有效诊断Vision Transformer网络的滚动轴承故障诊断方法
10
作者 罗志勇 李明周 董鑫 《重庆邮电大学学报(自然科学版)》 北大核心 2026年第1期146-155,共10页
针对滚动轴承故障诊断中特征提取不完整和诊断效率低的问题,提出了有效诊断Vision Transformer(EDViT)网络。采用基于峰度的加权融合策略,合并传感器信息;利用短时傅里叶变换,将融合后的信号转换为时频图像;依次应用EDViT的双重注意卷... 针对滚动轴承故障诊断中特征提取不完整和诊断效率低的问题,提出了有效诊断Vision Transformer(EDViT)网络。采用基于峰度的加权融合策略,合并传感器信息;利用短时傅里叶变换,将融合后的信号转换为时频图像;依次应用EDViT的双重注意卷积模块和双分支补丁视觉变换模块来提取局部和全局特征,使用分类器进行故障分类。实验验证在凯斯西储大学轴承数据集上进行。结果表明,EDViT模型具有出色的特征提取能力、快速的收敛速度和较高的诊断准确性。与其他方法的对比表明,EDViT模型具有很强的泛化能力和鲁棒性。 展开更多
关键词 有效诊断Vision transformer网络 滚动轴承 故障诊断
在线阅读 下载PDF
基于Transform-BiGRU和Swin-Transform-CBAM模型的双通道滚动轴承故障诊断
11
作者 徐纪龙 陈蕊 +1 位作者 孟召杰 袁长慧 《电子设计工程》 2026年第6期109-114,共6页
针对滚动轴承故障诊断中单通道特征提取不足的问题,提出一种基于连续小波变换的双通道轴承故障诊断模型。该模型构建了堆叠重构的信号通道与连续小波变换的二维图像通道;采用Transform-BiGRU网络提取时序特征,通过Swin-Transformer提取... 针对滚动轴承故障诊断中单通道特征提取不足的问题,提出一种基于连续小波变换的双通道轴承故障诊断模型。该模型构建了堆叠重构的信号通道与连续小波变换的二维图像通道;采用Transform-BiGRU网络提取时序特征,通过Swin-Transformer提取空间特征,并引入CBAM注意力机制强化关键信息;通过对两个通道提取的特征拼接融合,进行故障诊断。基于CWRU轴承测试集的实验表明,该模型的诊断准确率达到99.83%;通过对比实验表明,相较于传统的故障诊断模型,该文模型诊断准确率提升了7.99%~21.29%,验证了该模型在滚动轴承故障诊断中的有效性。 展开更多
关键词 滚动轴承 连续小波变换 双通道 深度学习 故障诊断
在线阅读 下载PDF
基于Transformer物理机理-数据协同驱动的航空发动机异常预测方法
12
作者 万海洋 彭博文 +3 位作者 刘芬 曹擎星 孙文兵 李巍华 《应用基础与工程科学学报》 北大核心 2026年第1期24-32,共9页
针对航空发动机早期异常预测的需求,提出了一种基于Transformer架构的物理机理与数据协同驱动框架.该框架将物理约束融入轴承故障仿真数据的生成过程中,为Leffler-Kernel增强的Transformer模型提供具有物理一致性的仿真训练样本,从而显... 针对航空发动机早期异常预测的需求,提出了一种基于Transformer架构的物理机理与数据协同驱动框架.该框架将物理约束融入轴承故障仿真数据的生成过程中,为Leffler-Kernel增强的Transformer模型提供具有物理一致性的仿真训练样本,从而显著提升了模型的物理可信度和泛化能力.在真实航空发动机实验平台上开展的人工故障轴承替换试验中,所采集的转子与机匣振动信号验证了该方法在早期故障识别中的优越预测性能.该方法具备良好的工程应用潜力,可为航空发动机系统的预测性维护与故障预防提供有力支持. 展开更多
关键词 物理机理-数据协同驱动 故障诊断 Leffler-Kernel transformER 航空发动机 异常预测
原文传递
基于Transformer的多任务协同监控架构研究
13
作者 牛云鹏 索智文 +3 位作者 王惠伟 屈波 周超逸 张丽芳 《煤炭工程》 北大核心 2026年第1期35-42,共8页
煤矿智能管控面临动态响应滞后和多源数据割裂的挑战,针对传统模型难以捕捉井下瞬态异常和协同分析多模态数据的问题。文章提出基于Transformer的多任务自适应架构(MTA-Transformer),通过跨模态特征融合与共享编码器,统一建模设备振动... 煤矿智能管控面临动态响应滞后和多源数据割裂的挑战,针对传统模型难以捕捉井下瞬态异常和协同分析多模态数据的问题。文章提出基于Transformer的多任务自适应架构(MTA-Transformer),通过跨模态特征融合与共享编码器,统一建模设备振动、瓦斯浓度等数据,实现开采环境的多尺度动态感知,解决对开采环境的动态监控与风险超前预警问题。实验表明,在轴承故障检测任务中,该模型准确率达93.5%,误报率(FAR)为2.0%,响应时间在5 ms内,较传统模型有较大提升;瓦斯浓度预测NRMSE为7.83%,预测区间覆盖概率(PICP)达91.7%,超前预警时效可达6 h。MTA-Transformer为矿山智能化建设提供了可落地的模式。 展开更多
关键词 煤矿智能管控 transformER 多任务协同 故障诊断 瓦斯浓度预测
在线阅读 下载PDF
基于Transformer的无人机故障检测研究
14
作者 张自旺 沈剑 +3 位作者 王晓光 刘繁 曹卓 贺斌娜 《机械设计与制造工程》 2026年第1期82-86,共5页
无人机故障检测作为保障飞行安全的核心技术,当前研究多依赖于仿真实验数据,并且传统方法难以有效捕捉飞行数据中的长程时空依赖关系。针对这些挑战,提出了一种基于Transformer架构的无人机故障检测方法,通过可学习位置编码和多头自注... 无人机故障检测作为保障飞行安全的核心技术,当前研究多依赖于仿真实验数据,并且传统方法难以有效捕捉飞行数据中的长程时空依赖关系。针对这些挑战,提出了一种基于Transformer架构的无人机故障检测方法,通过可学习位置编码和多头自注意力机制,构建传感器数据的时空依赖关系;同时结合焦点损失函数缓解类别不平衡问题。实验结果表明,该方法在真实飞行数据集上准确率达95%、F1分数达94%,相比基于LSTM和随机森林的故障检测方法展现更优的综合性能,并且在实时检测模拟中具有良好的可靠性,充分验证了其在真实飞行场景中的工程适用性。 展开更多
关键词 无人机 故障检测 transformER 焦点损失函数 实时检测模拟
在线阅读 下载PDF
基于主成分分析和Transformer模型的离心泵耦合故障智能诊断方法
15
作者 姜大连 许文 +3 位作者 吴尉 陈涛 周许杰 董亮 《机电工程》 北大核心 2026年第3期627-636,共10页
针对离心泵单一故障、耦合故障及不同程度故障在诊断过程中存在模式识别难度大、特征提取复杂度高及数据非线性可分等问题,提出了一种基于主成分分析(PCA)与Transformer模型的智能故障诊断方法。首先,对多源异构传感器采集的原始振动信... 针对离心泵单一故障、耦合故障及不同程度故障在诊断过程中存在模式识别难度大、特征提取复杂度高及数据非线性可分等问题,提出了一种基于主成分分析(PCA)与Transformer模型的智能故障诊断方法。首先,对多源异构传感器采集的原始振动信号进行了归一化预处理,以消除量纲差异并提升数据可靠性;然后,采用PCA对数据进行了降维处理,有效剔除了冗余信息,保留了最具判别性的关键特征;接着,完成预处理后,根据数据规模调整了Transformer模型中注意力机制头数等参数,并在优化后的网络结构上对PCA-Transformer模型进行了深度学习训练;最后,采用实验验证了PCA-Transformer模型在故障识别中的有效性。研究结果表明:预处理后的模型诊断准确率提升至99%,较原始数据提高了3%;训练准确率提升了2.5%,训练损失降低了75%。该方法在离心泵的单一故障、耦合故障及不同发展阶段的识别中均表现出良好性能,可为设备的智能监测与安全运行提供有力支撑。 展开更多
关键词 离心泵 故障诊断模型 主成分分析 transformer理论 注意力机制 数据降维处理 智能监测
在线阅读 下载PDF
一种基于多层融合-CNN-Transformer的防火切断阀故障诊断模型
16
作者 何阳 熊晓燕 +2 位作者 王伟杰 李翔宇 兰媛 《机电工程》 北大核心 2026年第2期269-279,共11页
防火切断阀作为飞机液压系统中的关键组件,一旦发生堵塞或阀芯偏移等故障,会使主回路供油不足、压力降低,导致液压负载元件故障,造成严重后果。因此,提出了一种基于多层融合-卷积神经网络(CNN)-Transformer的模型,用于防火切断阀的故障... 防火切断阀作为飞机液压系统中的关键组件,一旦发生堵塞或阀芯偏移等故障,会使主回路供油不足、压力降低,导致液压负载元件故障,造成严重后果。因此,提出了一种基于多层融合-卷积神经网络(CNN)-Transformer的模型,用于防火切断阀的故障诊断。首先,由于切断阀出口处缺乏压力测点,无法利用压差信号进行故障诊断,需采集三轴加速度信号并对其进行特征层预处理;然后,将处理好的数据输入至CNN-Transformer网络进行了训练与分类,CNN的小卷积层能够有效提取局部特征,Transformer则能够对全局特征进行捕捉;最后,针对发动机泵和增压泵等其他元件的振动干扰,利用Dempster-Shafer(DS)证据理论对位于切断阀入口处和出口处的两个加速度传感器的训练结果进行了决策层融合,以提高最终诊断的准确性和可靠性;在搭建的飞机液压系统试验台上对基于多层融合-CNN-Transformer的防火切断阀的故障诊断方法进行了实验验证,并与现有主流方法进行了对比。研究结果表明:基于多层融合-CNN-Transformer的防火切断阀的故障诊断方法在防火切断阀故障诊断中表现出最高的准确率,实验数据在阀芯开口为70%、80%、90%和100%的工况下的平均识别准确率达到99.5%。该方法可为飞行器液压系统中关键元件的智能诊断提供一种高可靠性的技术路线。 展开更多
关键词 防火切断阀 故障诊断 卷积神经网络 transformER Dempster-Shafer(DS)证据理论 多层融合模型
在线阅读 下载PDF
基于CNN-BiLSTM-Transformer的舰船中压直流全电推进系统故障诊断设计 被引量:2
17
作者 张建良 韩涛 季瑞松 《实验技术与管理》 北大核心 2025年第1期11-18,共8页
针对舰船中压直流全电推进系统结构复杂度高、单元耦合性强、运行环境多变等特点造成的故障诊断准确性低和实时性差等问题,开展了基于CNN-BiLSTM-Transformer的故障诊断设计。首先,基于卷积神经网络CNN构建单点特征级联网络,开展单一时... 针对舰船中压直流全电推进系统结构复杂度高、单元耦合性强、运行环境多变等特点造成的故障诊断准确性低和实时性差等问题,开展了基于CNN-BiLSTM-Transformer的故障诊断设计。首先,基于卷积神经网络CNN构建单点特征级联网络,开展单一时刻下故障信号空间特征的深入提取,以提升故障特征提取的有效性;其次,以双向长短期记忆网络BiLSTM为核心设计多点特征依赖网络,利用门控机制和双向时序学习机制,实现故障信号在多个时刻之间特征依赖关系的有效学习,以提升故障诊断的准确性;然后,以Transformer为核心建立序列特征并行处理网络,通过自注意力机制实现对故障特征上下文关系的精确刻画,进而利用多头注意力机制实现特征序列的并行处理,以提升故障诊断的实时性;最后,设计舰船中压直流全电推进系统故障诊断实验方案,并开展不同故障模式下的诊断性能评估。该文方法在多种故障模式下诊断准确率和实时性均优于现有的主流故障诊断方法,有助于为舰船中压直流全电推进系统的安全运行提供更有力的技术保障。 展开更多
关键词 舰船 中压直流 全电推进系统 故障诊断 transformER
在线阅读 下载PDF
基于CNN-Transformer模型的堆垛机振动故障诊断研究
18
作者 孙晓霞 王博 孟文俊 《起重运输机械》 2025年第20期78-85,共8页
堆垛机在实际运行过程中存在缺乏故障数据的挑战,且利用阈值设定判断堆垛机的运行情况误差较大。为了解决此问题,文中提出一种多源数据样本融合到结构映射策略,利用卷积神经网络的局部特征提取能力与Transformer的时序建模优势构建了CNN... 堆垛机在实际运行过程中存在缺乏故障数据的挑战,且利用阈值设定判断堆垛机的运行情况误差较大。为了解决此问题,文中提出一种多源数据样本融合到结构映射策略,利用卷积神经网络的局部特征提取能力与Transformer的时序建模优势构建了CNN-Transformer模型进行堆垛机的振动故障识别。实验结果表明:该模型展现出优良的故障识别精度与泛化性能。 展开更多
关键词 堆垛机 多源数据融合 CNN transformer 振动故障诊断
在线阅读 下载PDF
Improved BP Neural Network for Transformer Fault Diagnosis 被引量:42
19
作者 SUN Yan-jing ZHANG Shen MIAO Chang-xin LI Jing-meng 《Journal of China University of Mining and Technology》 EI 2007年第1期138-142,共5页
The back propagation (BP)-based artificial neural nets (ANN) can identify complicated relationships among dissolved gas contents in transformer oil and corresponding fault types, using the highly nonlinear mapping nat... The back propagation (BP)-based artificial neural nets (ANN) can identify complicated relationships among dissolved gas contents in transformer oil and corresponding fault types, using the highly nonlinear mapping nature of the neural nets. An efficient BP-ALM (BP with Adaptive Learning Rate and Momentum coefficient) algorithm is proposed to reduce the training time and avoid being trapped into local minima, where the learning rate and the momentum coefficient are altered at iterations. We developed a system of transformer fault diagnosis based on Dissolved Gases Analysis (DGA) with a BP-ALM algorithm. Training patterns were selected from the results of a Refined Three-Ratio method (RTR). Test results show that the system has a better ability of quick learning and global convergence than other methods and a superior performance in fault diagnosis compared to convectional BP-based neural networks and RTR. 展开更多
关键词 transformer fault diagnosis BACK-PROPAGATION artificial neural network momentum coefficient
在线阅读 下载PDF
多变量时序标记Transformer及其在电潜泵故障诊断中的应用 被引量:9
20
作者 李康 李爽 +2 位作者 高小永 李强 张来斌 《控制与决策》 北大核心 2025年第4期1145-1153,共9页
电潜泵故障诊断对于确保安全可靠采油至关重要,但是,电潜泵数据呈现出的多变量、非线性和动态变化等复杂特性为该任务带来了严峻挑战.近年来,深度学习在复杂数据特征提取方面表现出的强大能力催生了一系列基于神经网络的电潜泵故障诊断... 电潜泵故障诊断对于确保安全可靠采油至关重要,但是,电潜泵数据呈现出的多变量、非线性和动态变化等复杂特性为该任务带来了严峻挑战.近年来,深度学习在复杂数据特征提取方面表现出的强大能力催生了一系列基于神经网络的电潜泵故障诊断方法.然而,多数方法忽略了电潜泵数据的动态特性以及长时依赖特征提取困难的问题.针对上述问题,提出一种多变量时序标记Transformer神经网络来实现电潜泵故障诊断.该模型设计新的多变量时间序列标记策略,继承引入多头注意力机制和残差连接的传统Transformer神经网络编码器在长时依赖特征提取方面的优势,用前向神经网络替代传统Transformer神经网络解码器来简化模型复杂度.通过对油田现场故障数据分析,验证所提出方法的有效性.实验结果表明,所提出方法实现了10类电潜泵故障的精确诊断,相比于流行的深度学习方法诊断性能更优. 展开更多
关键词 电潜泵 transformer神经网络 深度学习 特征提取 故障诊断 多变量时序标记
原文传递
上一页 1 2 250 下一页 到第
使用帮助 返回顶部