检索结果-内蒙古大学图书馆

Journal of Network Intelligence 2022年第3期7卷 704-718页

作者： Zhao, Li-Quan Zhou, Si-Ying Teng, Zi-Ming Jiang, Zi-Cong Jia, Yan-Fei Key Laboratory of Modern Power System Simulation and Control & Renewable Energy Technology Ministry of Education Northeast Electric Power University Jilin132012 China College of information and communication Jilin University Changchun130012 China Graduate school of Information Science and Electrical Engineering Kyushu University Fukuoka819-0395 Japan School of Electrical and Information Engineering Beihua University Jilin132013 China

An unsupervised domain adaptation method is proposed based on the step-wise adaptive feature norm method to improve transfer learning accuracy. The feature extraction network consists of channel attention and spatial attention modules, which en-hance feature extraction ability. To reduce the influence of dimension change on channel weight, a convolution layer with the same kernel size is used to replace the fully connected layer in the bottleneck layer used in the common channel attention module. In addition, a smooth loss function is also proposed. The curve of the proposed function is smoother than the L2 function that is the least square error function. The proposed function’s gradient changes more slowly than the L2 function. The proposed function is used as the activate function of classification prediction regression to reduce the risk of gradient explosion. Compared with a selection of existing methods (Resnet50, conditional domain adversarial networks method, deep adaptive network method, joint adaptation network method, domain adaptation neural network method, and stepwise adaptive feature norm method), the proposed transfer learning method has the highest accuracy than other methods on different datasets for all tasks. © 2022 Taiwan Ubiquitous Information CO LTD. All rights reserved.

关键词： Feature extraction

来源：评论

学校读者我要写书评

暂无评论

Wind power Forecasting Considering Data Privacy Protection: A Federated Deep Reinforcement Learning Approach

arXiv

引用

arXiv 2022年

作者： Li, Yang Wang, Ruinong Li, Yuanzheng Zhang, Meng Long, Chao The Key Laboratory of Modern Power System Simulation Control and Renewable Energy Technology Ministry of Education Northeast Electric Power University Jilin132012 China The State Grid Jilin Electric Power Co. Ltd Jilin Power Supply Company Chuanying District Jilin132001 China The School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan430074 China The School of National Key Laboratory of Science and Technology Vessel Integrated Power System Naval University of Engineering Wuhan430033 China The School of Water Energy and Environment Cranfield University United Kingdom

In a modern power system with an increasing proportion of renewable energy, wind power prediction is crucial to the arrangement of power grid dispatching plans due to the volatility of wind power. However, traditional centralized forecasting methods raise concerns regarding data privacy-preserving and data islands problem. To handle the data privacy and openness, we propose a forecasting scheme that combines federated learning and deep reinforcement learning (DRL) for ultra-short-term wind power forecasting, called federated deep reinforcement learning (FedDRL). Firstly, this paper uses the deep deterministic policy gradient (DDPG) algorithm as the basic forecasting model to improve prediction accuracy. Secondly, we integrate the DDPG forecasting model into the framework of federated learning. The designed FedDRL can obtain an accurate prediction model in a decentralized way by sharing model parameters instead of sharing private data which can avoid sensitive privacy issues. The simulation results show that the proposed FedDRL outperforms the traditional prediction methods in terms of forecasting accuracy. More importantly, while ensuring the forecasting performance, FedDRL can effectively protect the data privacy and relieve the communication pressure compared with the traditional centralized forecasting method. In addition, a simulation with different federated learning parameters is conducted to confirm the robustness of the proposed scheme. Copyright © 2022, The Authors. All rights reserved.

关键词： Electric load dispatching

来源：评论

学校读者我要写书评

暂无评论

Corrigendum to ‘Optimised configuration of multi-energy systems considering the adjusting capacity of communication base stations and risk of network congestion’ [energy volume 313 (2024) 133873]

引用

energy 2025年 314卷

作者： Dongfeng Yang Xi Wang Yong Sun Jingying Yang Xiaojun Liu Chao Jiang Key Laboratory of Modern Power System Simulation and Control & Renewable Energy Technology Ministry of Education (Northeast Electric Power University) Jilin 132012 China State Grid Jilin Electric Power Co. Ltd Changchun 30000 China

来源：评论

学校读者我要写书评

暂无评论

Seismic Interpolation Transformer for Consecutively Missing Data: A Case Study in DAS-VSP Data

arXiv

引用

arXiv 2024年

作者： Cheng, Ming Lin, Jun Dong, Xintong Lu, Shaoping Zhong, Tie College of Instrumentation and Electrical Engineering Jilin University Jilin Changchun130026 China School of Earth Sciences and Engineering Sun Yat-Sen University Guangdong Guangzhou510275 China Guangdong Zhuhai519000 China Guangdong Provincial Key Lab of Geodynamics and Geohazards Sun Yat-sen University Guangdong Guangzhou510275 China Key Laboratory of Modern Power System Simulation and Control and Renewable Energy Technology Ministry of Education Jilin132012 China Department of Communication Engineering College of Electric Engineering Northeast Electric Power University Jilin132012 China

Distributed optical fiber acoustic sensing (DAS) is a rapidly-developed seismic acquisition technology with advantages of low cost, high resolution, high sensitivity, and small interval, etc. Nonetheless, consecutively missing cases often appear in real seismic data acquired by DAS system due to some factors, including optical fiber damage and inferior coupling between cable and well. Recently, some deep-learning seismic interpolation methods based on convolutional neural network (CNN) have shown impressive performance in regular and random missing cases but still remain the consecutively missing case as a challenging task. The main reason is that the weight sharing makes it difficult for CNN to capture enough comprehensive features. In this paper, we propose a transformer-based interpolation method, called seismic interpolation transformer (SIT), to deal with the consecutively missing case. This proposed SIT is an encoder-decoder structure connected by some U-shaped swin-transformer blocks. In encoder and decoder part, the multi-head self-attention (MSA) mechanism is used to capture global features which is essential for the reconstruction of consecutively missing traces. The U-shaped swin-transformer blocks are utilized to perform feature extraction operations on feature maps with different resolutions. Moreover, we combine the loss based on structural similarity index (SSIM) and L1 norm to propose a novel loss function for SIT. In experiments, this proposed SIT outperforms U-Net and swin-transformer. Moreover, ablation studies also demonstrate the advantages of new network architecture and loss function. Copyright © 2024, The Authors. All rights reserved.

关键词： Optical fibers

来源：评论

学校读者我要写书评

暂无评论

Ultrathin microcrystalline hydrogenated Si/Ge alloyed tandem solar cells towards full solar spectrum conversion

引用

Frontiers of Chemical Science and Engineering 2020年第6期14卷 997-1005页

作者： Yu Cao Xinyun Zhu Xingyu Tong Jing Zhou Jian Ni Jianjun Zhang Jinbo Pang Key Laboratory of Modem Power System Simulation and Control&Renewable Energy Technology Northeast Electric Power UniversityJilin 132012China School of Electrical Engineering Northeast Electric Power UniversityJilin 132012China School of Chemical Engineering Northeast Electric Power UniversityJilin 132012China Key Laboratory of Photo-electronics Thin Film Devices and Technique of Tianjin Institute of Photo-electronic Thin Film Devices and TechnologyCollege of Electronic Information and Optical EngineeringNankai UniversityTianjin 300350China Collaborative Innovation Center of Technology and Equipment for Biological Diagnosis and Therapy in Universities of Shandong Institute for Advanced Interdisciplinary Research(iAIR)University of JinanJinan 250022China

Thin film solar cells have been proved the next generation photovoltaic devices due to their low cost,less material consumption and easy mass *** them,micro-crystalline Si and Ge based thin film solar cells have advantages of high efficiency and ultrathin absorber *** individual junction devices are limited in photoelectric conversion efficiency because of the restricted solar spectrum range for its specific *** this work,w e designed a nd simulated a multi-junction solar cell with its four sub-cells selectively absorbing the full solar spectrum including the ultraviolet,green,red as well as near infrared range,respectively.B y tuning the G e content,the record efficiency of 24.80%has been realized with the typical quadruple junction structure of a-Si:H/a-Si0.9Ge0.1:H/μc-Si:H/μc-Si0.5Ge0.5:*** further reduce the material cost,thickness dependent device performances have been *** can be found that the design of total thickness of 4μm is the optimal device design in balancing the thickness a nd the *** the design of ultrathin quadruple junction device with total thickness of 2μm is the optimized device design regarding cost and long-term stability with a little bit more reduction in *** results indicated that our solar cells combine the advantages of low cost and high *** work may provide a general guidance rule of utilizing the full solar spectrum for developing high efficiency and ultrathin multi-junction solar cells.

关键词： thin films solar cells quadruple junction solar cell amorphous silicon silicon germanium alloy quantum efficiency

来源：评论

学校读者我要写书评

暂无评论

Fault Detection Method for Insulators Using Improved YOLOv4

Journal of Network Intelligence

引用

Journal of Network Intelligence 2022年第4期7卷 818-834页

作者： Zhao, Li-Quan Jiang, Zi-Cong Teng, Zi-Ming Jia, Yan-Fei Key Laboratory of Modern Power System Simulation and Control & Renewable Energy Technology Ministry of Education Northeast Electric Power University 168 Changchun Road Jilin132012 China Graduate school of Information Science and Electrical Engineering Kyushu University Fukuoka819-0395 Japan College of information and communication Jilin University Changchun130012 China School of Electrical and Information Engineering Beihua University Jilin132013 China

A defective insulator detection method based on improved You Only Look Once version 4 (YOLOv4) is proposed to improve the precision of defective insulator detection. This method designs a dual-branch attention block by simulating the attention process of the human eye, and introduces it into the backbone of YOLOv4 to extract richer information on defective insulator features. To make the output feature map of shallow layer network fuse more different scale features, a path aggregation network has been designed. To improve detection accuracy, the spatial pyramid pooling has been replaced by atrous spatial pyramid pooling. A dynamic non-maximum suppression method is proposed to improve defective insulator detection accuracy. It employs an adaptive threshold control strategy and weighted operation to full use of the predicted box. We compare our method with defective insulator detection methods based on YOLO and R-CNN. Our simulation results show that the proposed method is more accurate than defective insulator detection methods based on YOLO, and has a faster detection speed than methods based on R-CNN. Compared with the defective insulator detection method based on R-CNN that has the highest mean average precision, the mean average precision of the proposed method is lower, but its detection speed is faster. Compared with YOLOv4, its detection speed is slower, but the mean average precision is higher. The proposed method strikes a satisfactory tradeoff between detection speed and mean average precision. © 2022, Taiwan Ubiquitous Information CO LTD. All rights reserved.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Selective Preferred Orientation for High-Performance Antimony Selenide Thin-Film Solar Cells Via Substrate Surface Modulation

SSRN

引用

SSRN 2022年

作者： Zhou, Jing Zhu, Jiawei He, Weilan Cao, Yu Pang, Jinbo Ni, Jian Zhang, Jianjun School of Chemical Engineering Northeast Electric Power University Jilin132012 China Key Laboratory of Modern Power System Simulation and Control & Renewable Energy Technology Ministry of Education School of Electrical Engineering Northeast Electric Power University Jilin132012 China University of Jinan Jinan250022 China College of Electronic Information and Optical Engineering Nankai University Tianjin300350 China

Antimony selenide (Sb2Se3) has rapidly emerged in the field of photovoltaics using one-dimensional crystal structure. However, the power conversion efficiency (PCE) of Sb2Se3 solar cells is influenced by the orientation of the Sb2Se3 photoactive layers. In this study, we investigate the influence of substrate properties on the preferred orientation and microstructure of Sb2Se3 films systematically. With the rise of substrate surface temperature and the decline of substrate surface bonding energy, the preferred orientation of the Sb2Se3 films altered from (hk0) to (hk1) and finally to (002). After summarizing the performance of different preferentially oriented Sb2Se3 solar cells, the results showed that the (hk0)-preferentially oriented devices exhibited a poor performance of less than 1.1%. As the preferred orientation of Sb2Se3 absorbers changed from (hk0) to (hk1), the carrier transport changed from inter-ribbon charge hopping to transportation along the (Sb4Se6)n ribbons. This enhanced the carrier transport efficiency, thereby leading to the enhancement of the PCE to 2%–3%. Furthermore, when the (002) texture coefficient (TC) of the Sb2Se3 films was between 1.8 and 3.6, the PCE of the devices was further enhanced to above 3.5%. A maximum PCE of 4.79% can be obtained when the (002) TC was 3.24. This study reveals the relationship between the Sb2Se3 film preferred orientation and device performance, and demonstrates an effective way to precisely control the Sb2Se3 film microstructure to obtain a high PCE. © 2022, The Authors. All rights reserved.

关键词： Selenium compounds

来源：评论

学校读者我要写书评

暂无评论

Conditional Adversarial Domain Adaption based on Self-attention

Journal of Network Intelligence

引用

Journal of Network Intelligence 2022年第1期7卷 175-188页

作者： Zhao, Li-Quan Zhang, Yu-Peng Teng, Zi-Ming Jiang, Zi-Cong Cui, Ying Kan, Zhong-Feng Key Laboratory of Modern Power System Simulation and Control & Renewable Energy Technology Ministry of Education Northeast Electric Power University Jilin132012 China College of information and communication Jilin University Changchun130012 China Graduate school of Information Science and Electrical Engineering Kyushu University Fukuoka819-0395 Japan Zhuhai Power Supply Bureau Guangdong Electric Power Corporation Zhuhai519000 China Jilin Power Supply Bureau Jilin Electric Power Corporation Jilin132012 China

Transfer learning is a deep learning method. It transforms knowledge in the source domain and feeds it to the target domain to improve the performance of the target domain model. The conditional adversarial domain adaption method is one of a number of transfer learning methods. It uses Resnet as the feature extraction network;however, Resnet can only determine the relationship between local features, and so its feature extraction accuracy is limited. To overcome this problem, a self-attention layer has been designed in order to determine the relationship among all features. This layer is added between the fourth layer and fifth layer of Resnet. Besides, a cross-entropy function is introduced into loss function, and a sigmoid layer is designed and used as the out layer in conditional adversarial domain adaption. These produce gradient disappearance in the deep network, and influence the discriminator training effect. To solve that problem, a least squares function is used as to design the loss function. The new type of loss function does not require the sigmoid function. Using theoretical analysis, we explain how this new loss function can be used to solve the gradient disappearance problem, thus rendering the network structure more stable. We compare our proposed method with other datasets, including Joint Adaptation Networks, adversarial domain adaptation, conditional adversarial domain adaption, Resnet-50, domain-adversarial neural network on imageCLEF-DA, Office-Home, and Office-31. Our simulation results show that the proposed conditional adversarial domain adaption based on self-attention method has the fastest convergence speed and highest accuracy the among different datasets. © 2022, Taiwan Ubiquitous Information CO LTD. All rights reserved.

关键词： Computer vision

来源：评论

学校读者我要写书评

暂无评论

Capacity and power Allocation Strategy of energy Storage Participating in Grid Peak Shaving

Capacity and Power Allocation Strategy of Energy Storage Par...

引用

IEEE Conference on energy Internet and energy system Integration (EI2)

作者： Tengyun Qi Qiren Wu Jianping Liu Fei Xu Yajing Li He Li Yiming Wang Cuiping Li Junhui Li China Three Gorges Corporation Beijing China China Three Gorges Renewables(Group) Co. Ltd Beijing China State Key Laboratory of Power System and Power Generation Equipment Control and Simulation (Tsinghua University) Beijing China Key Laboratory of Modern Power System Simulation and Control & Renewable Energy Technology Ministry of Education (Northeast Electric Power University) Jilin China

ISBN: (纸本)9781665434263

High penetration wind power grid with energy storage system can effectively improve peak load regulation pressure and increase wind power capacity. In this paper, a capacity allocation method of energy storage system under peak load regulation scenario is proposed. The upper model combines the investment cost, operation cost, arbitrage income, environmental income, and wind power grid benefits during the entire life cycle of the energy storage system, with the goal of maximizing the net income during the energy storage life cycle; the lower model optimizes the charging and discharging power of the energy storage system according to each energy storage configuration scheme in the upper modeling, and its objective is to minimize the standard deviation of system net load and waste wind power. Finally, under the conditions of local grid load and wind power data, verified the effectiveness of capacity allocation strategy.

关键词： Costs system integration Production Wind power generation Lithium batteries Regulation Resource management

来源：评论

学校读者我要写书评

暂无评论

MAE-GAN: A Novel Strategy for Simultaneous Super-resolution Reconstruction and Denoising of Post-stack Seismic Profile

arXiv

引用

arXiv 2024年

作者： Yu, Wenshuo Dong, Shiqi Lu, Shaoping Dong, Xintong The College of Instrumentation and Electrical Engineering Jilin University Jilin Changchun130026 China The Key Laboratory of Modern Power System Simulation and the Control and Renewable Energy Technology Ministry of Education Jilin Jilin City132012 China Department of Communication Engineering College of Electric Engineering Northeast Electric Power University Jilin Jilin City132012 China School of Earth Sciences and Engineering Sun Yat-Sen University Guangdong Guangzhou510275 China Guangdong Zhuhai519000 China Guangdong Provincial Key Lab of Geodynamics and Geohazards Sun Yat-Sen University Guangdong Guangzhou510275 China

Post-stack seismic profiles are images reflecting containing geological structures which provides a critical foundation for understanding the distribution of oil and gas resources. However, due to the limitations of seismic acquisition equipment and data collecting geometry, the post-stack profiles suffer from low resolution and strong noise issues, which severely affects subsequent seismic interpretation. To better enhance the spatial resolution and signal-to-noise ratio of post-seismic profiles, a multi-scale attention encoder-decoder network based on generative adversarial network (MAE-GAN) is proposed. This method improves the resolution of post-stack profiles, and effectively suppresses noises and recovers weak signals as well. A multi-scale residual module is proposed to extract geological features under different receptive fields. At the same time, an attention module is designed to further guide the network to focus on important feature information. Additionally, to better recover the global and local information of post-stack profiles, an adversarial network based on a Markov discriminator is proposed. Finally, by introducing an edge information preservation loss function, the conventional loss function of the Generative Adversarial Network is improved, which enables better recovery of the edge information of the original post-stack profiles. Experimental results on simulated and field post-stack profiles demonstrate that the proposed MAE-GAN method outperforms two advanced convolutional neural network-based methods in noise suppression and weak signal recovery. Furthermore, the profiles reconstructed by the MAE-GAN method preserve more geological structures. Copyright © 2024, The Authors. All rights reserved.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：