Global energy and environmental issues are becoming increasingly problematic,and the vibration and noise problem of 110 kV transformers,which are the most widely distributed,have attracted widespread attention from bo...Global energy and environmental issues are becoming increasingly problematic,and the vibration and noise problem of 110 kV transformers,which are the most widely distributed,have attracted widespread attention from both inside and outside the industry.DC bias is one of the main contributing factors to vibration noise during the normal operation of transformers.To clarify the vibration and noise mechanism of a 110 kV transformer under a DC bias,a multi-field coupling model of a 110 kV transformer was established using the finite element method.The electromagnetic,vibration,and noise characteristics during the DC bias process were compared and quantified through field circuit coupling in parallel with the power frequency of AC,harmonic,and DC power sources.It was found that a DC bias can cause significant distortions in the magnetic flux density,force,and displacement distributions of the core and winding.The contributions of the DC bias effect to the core and winding are different at Kdc=0.85.At this point,the core approached saturation,and the increase in the core force and displacement slowed.However,the saturation of the core increased the leakage flux,and the stress and displacement of the winding increased faster.The sound field distribution characteristics of the 110 kV transformer under a DC bias are related to the force characteristics.When the DC bias coefficient was 1.25,the noise sound pressure level reached 73.6 dB.展开更多
Nowadays,ensuring thequality of networkserviceshas become increasingly vital.Experts are turning toknowledge graph technology,with a significant emphasis on entity extraction in the identification of device configurat...Nowadays,ensuring thequality of networkserviceshas become increasingly vital.Experts are turning toknowledge graph technology,with a significant emphasis on entity extraction in the identification of device configurations.This research paper presents a novel entity extraction method that leverages a combination of active learning and attention mechanisms.Initially,an improved active learning approach is employed to select the most valuable unlabeled samples,which are subsequently submitted for expert labeling.This approach successfully addresses the problems of isolated points and sample redundancy within the network configuration sample set.Then the labeled samples are utilized to train the model for network configuration entity extraction.Furthermore,the multi-head self-attention of the transformer model is enhanced by introducing the Adaptive Weighting method based on the Laplace mixture distribution.This enhancement enables the transformer model to dynamically adapt its focus to words in various positions,displaying exceptional adaptability to abnormal data and further elevating the accuracy of the proposed model.Through comparisons with Random Sampling(RANDOM),Maximum Normalized Log-Probability(MNLP),Least Confidence(LC),Token Entrop(TE),and Entropy Query by Bagging(EQB),the proposed method,Entropy Query by Bagging and Maximum Influence Active Learning(EQBMIAL),achieves comparable performance with only 40% of the samples on both datasets,while other algorithms require 50% of the samples.Furthermore,the entity extraction algorithm with the Adaptive Weighted Multi-head Attention mechanism(AW-MHA)is compared with BILSTM-CRF,Mutil_Attention-Bilstm-Crf,Deep_Neural_Model_NER and BERT_Transformer,achieving precision rates of 75.98% and 98.32% on the two datasets,respectively.Statistical tests demonstrate the statistical significance and effectiveness of the proposed algorithms in this paper.展开更多
Ischemic stroke is an important disease leading to death and disability for all human beings, and the key to its treatment lies in the early opening of obstructed vessels and restoration of perfusion to the local infa...Ischemic stroke is an important disease leading to death and disability for all human beings, and the key to its treatment lies in the early opening of obstructed vessels and restoration of perfusion to the local infarcted area. Intravenous thrombolysis with tissue plasminogen activator (tPA) is one of the effective therapies to achieve revascularization, but it faces strict indications with a narrow therapeutic time window, and significantly increases the incidence of hemorrhagic transformation, HT, after reperfusion of the infarcted foci, which greatly reduces the incidence of patients with ischemic stroke. which significantly increases the incidence of hemorrhagic transformation (HT) after reperfusion of the infarcted focus, greatly reducing patient utilization and clinical benefit. Since the mechanism of HT has not been fully elucidated, and the related molecular mechanisms are complex and interactive, there is no specific and effective therapy to avoid the occurrence of HT. In this article, we focus on the research progress on the mechanism of HT after tPA intravenous thrombolysis in ischemic stroke patients from the aspects of vascular integrity disruption, oxidative stress, and neuroinflammatory response and the corresponding therapeutic strategies, in order to improve the safety and prognosis of tPA intravenous thrombolysis in the clinic.展开更多
针对多变量时序(Multivariate Time Series,MTS)分类中长序列数据难以捕捉时序特征的问题,提出一种基于双向稀疏Transformer的时序分类模型BST(Bidirectional Sparse Transformer),提高了MTS分类任务的准确度.BST模型使用Transformer框...针对多变量时序(Multivariate Time Series,MTS)分类中长序列数据难以捕捉时序特征的问题,提出一种基于双向稀疏Transformer的时序分类模型BST(Bidirectional Sparse Transformer),提高了MTS分类任务的准确度.BST模型使用Transformer框架,构建了一种基于活跃度得分的双向稀疏注意力机制.基于KL散度构建活跃度评价函数,并将评价函数的非对称问题转变为对称权重问题.据此,对原有查询矩阵、键值矩阵进行双向稀疏化,从而降低原Transformer模型中自注意力机制运算的时间复杂度.实验结果显示,BST模型在9个长序列数据集上取得最高平均排名,在临界差异图中领先第2名35.7%,对于具有强时序性的乙醇浓度数据集(Ethanol Concentration,EC),分类准确率提高30.9%.展开更多
基金supported by the Key R&D Program of Shandong Province(2021CXGC010210).
文摘Global energy and environmental issues are becoming increasingly problematic,and the vibration and noise problem of 110 kV transformers,which are the most widely distributed,have attracted widespread attention from both inside and outside the industry.DC bias is one of the main contributing factors to vibration noise during the normal operation of transformers.To clarify the vibration and noise mechanism of a 110 kV transformer under a DC bias,a multi-field coupling model of a 110 kV transformer was established using the finite element method.The electromagnetic,vibration,and noise characteristics during the DC bias process were compared and quantified through field circuit coupling in parallel with the power frequency of AC,harmonic,and DC power sources.It was found that a DC bias can cause significant distortions in the magnetic flux density,force,and displacement distributions of the core and winding.The contributions of the DC bias effect to the core and winding are different at Kdc=0.85.At this point,the core approached saturation,and the increase in the core force and displacement slowed.However,the saturation of the core increased the leakage flux,and the stress and displacement of the winding increased faster.The sound field distribution characteristics of the 110 kV transformer under a DC bias are related to the force characteristics.When the DC bias coefficient was 1.25,the noise sound pressure level reached 73.6 dB.
基金supported by the National Key R&D Program of China(2019YFB2103202).
文摘Nowadays,ensuring thequality of networkserviceshas become increasingly vital.Experts are turning toknowledge graph technology,with a significant emphasis on entity extraction in the identification of device configurations.This research paper presents a novel entity extraction method that leverages a combination of active learning and attention mechanisms.Initially,an improved active learning approach is employed to select the most valuable unlabeled samples,which are subsequently submitted for expert labeling.This approach successfully addresses the problems of isolated points and sample redundancy within the network configuration sample set.Then the labeled samples are utilized to train the model for network configuration entity extraction.Furthermore,the multi-head self-attention of the transformer model is enhanced by introducing the Adaptive Weighting method based on the Laplace mixture distribution.This enhancement enables the transformer model to dynamically adapt its focus to words in various positions,displaying exceptional adaptability to abnormal data and further elevating the accuracy of the proposed model.Through comparisons with Random Sampling(RANDOM),Maximum Normalized Log-Probability(MNLP),Least Confidence(LC),Token Entrop(TE),and Entropy Query by Bagging(EQB),the proposed method,Entropy Query by Bagging and Maximum Influence Active Learning(EQBMIAL),achieves comparable performance with only 40% of the samples on both datasets,while other algorithms require 50% of the samples.Furthermore,the entity extraction algorithm with the Adaptive Weighted Multi-head Attention mechanism(AW-MHA)is compared with BILSTM-CRF,Mutil_Attention-Bilstm-Crf,Deep_Neural_Model_NER and BERT_Transformer,achieving precision rates of 75.98% and 98.32% on the two datasets,respectively.Statistical tests demonstrate the statistical significance and effectiveness of the proposed algorithms in this paper.
文摘Ischemic stroke is an important disease leading to death and disability for all human beings, and the key to its treatment lies in the early opening of obstructed vessels and restoration of perfusion to the local infarcted area. Intravenous thrombolysis with tissue plasminogen activator (tPA) is one of the effective therapies to achieve revascularization, but it faces strict indications with a narrow therapeutic time window, and significantly increases the incidence of hemorrhagic transformation, HT, after reperfusion of the infarcted foci, which greatly reduces the incidence of patients with ischemic stroke. which significantly increases the incidence of hemorrhagic transformation (HT) after reperfusion of the infarcted focus, greatly reducing patient utilization and clinical benefit. Since the mechanism of HT has not been fully elucidated, and the related molecular mechanisms are complex and interactive, there is no specific and effective therapy to avoid the occurrence of HT. In this article, we focus on the research progress on the mechanism of HT after tPA intravenous thrombolysis in ischemic stroke patients from the aspects of vascular integrity disruption, oxidative stress, and neuroinflammatory response and the corresponding therapeutic strategies, in order to improve the safety and prognosis of tPA intravenous thrombolysis in the clinic.