课程文献中心 更多>>

关键词： 变压器   缺陷识别   油中溶解气体分析   不平衡性   动态多种群粒子群优化算法   极限学习机  

摘要： 针对变压器缺陷样本存在不平衡性，现有诊断方法适用性不足的问题，提出一种面向油中溶解气体分析（Dissolved Gas Analysis,DGA）不平衡数据的变压器缺陷识别方法。首先，基于内蒙古西部地区部分电厂送检油样的DGA分析，重新划分5种变压器状态类型，并通过增加状态特征增大其差距；然后，构建综合考虑检测气体含量、无编码比值、三比值的输入特征，对智能诊断模型进行训练和测试；最后，选用改进动态多种群粒子群优化（Particle Swarm Optimization,PSO）算法优化极限学习机（Extreme Learning Machine, ELM）神经网络参数，通过对未定义类型样本进行分析，验证方法的有效性。结果表明：本方法适用于DGA不平衡数据集，能够对变压器运行状态及缺陷进行准确识别。

关键词： 光伏功率预测   生成对抗网络   北方苍鹰算法   双向长短期记忆网络  

摘要： 鉴于诸多光伏发电站存在的数据完备性不足，由此引发的数据样本质量不佳进而导致光伏功率预测准确性受限.本文提出了一种光伏功率超短期组合预测模型，该模型融合了基于梯度惩罚Wasserstein生成对抗网络（Wasserstein generative adversarial network with gradient penalty，WGAN-GP）模型及双向长短期记忆（Bidirectional Long Short-Term Memory Network，BiLSTM）模型的技术特点.首先对光伏发电功率与不同天气变量间的相关性进行分析，选取辐射度、温度和风速这三个对光伏发电功率影响较大的天气变量作为特征输入，以提升模型的预测精度；其次，采用WGAN-GP模型来挖掘光伏出力实际数据集的内在结构及与其密切相关的环境和气象因素之间的深层次关联，利用生成的高质量数据样本扩充原有数据集，有效提升预测模型的泛化能力；最后，采用北方苍鹰优化算法（Northern Goshawk Optimization， NGO）对BiLSTM预测模型进行参数寻优可进一步提高光伏发电功率的预测精度.实验结果表明，经过数据增强后NGO-BiLSTM模型的预测精度可有效提升.

摘要： The theoretical limit of subthreshold swing in MOSFETs is extended down to sub-Kelvin temperatures. By including the band and band-tail states, the existence of a new phase 4 is derived from the theory. In phase 4, the subthreshold swing plateaus again when the Fermi level saturates inside the band. The model for subthreshold swing in phase 4 is proportional to a Fermi voltage instead of the thermal voltage. A short formula for the slope factor in phase 3 is also derived which allows to extract the subband density-of-states and effective mass. The phase-4 model allows to extract the Fermi level position. The proposed theoretical model is validated against experimental data from the recent literature. Even for a sharp band edge, the theory predicts that the subthreshold swing deviates from the Boltzmann thermal limit at sub-Kelvin temperatures.

关键词： Fermi level  

摘要： The theoretical limit of subthreshold swing in MOSFETs is extended down to sub-Kelvin temperatures. By including the band and band-tail states, the existence of a new phase 4 is derived from the theory. In phase 4, the subthreshold swing plateaus again when the Fermi level saturates inside the band. The model for subthreshold swing in phase 4 is proportional to a Fermi voltage instead of the thermal voltage. A short formula for the slope factor in phase 3 is also derived which allows to extract the subband density-of-states and effective mass. The phase-4 model allows to extract the Fermi level position. The proposed theoretical model is validated against experimental data from the recent literature. Even for a sharp band edge, the theory predicts that the subthreshold swing deviates from the Boltzmann thermal limit at sub-Kelvin temperatures. Copyright © 2024, The Authors. All rights reserved.

关键词： 脉冲激光器   全光纤   掺镱光纤激光器   种子源调制   主振荡功率放大器  

摘要： 本文以电路直接调制的半导体激光器作为种子源，采用三级MOPA结构，设计并实现了一种输出功率、重复频率与脉冲宽度均可宽调节的高峰值功率高光束质量的全光纤MOPA脉冲激光器。在一二级泵浦功率分别为2.2W和34.3W条件下，获得了平均输出功率为20.85W，最大峰值功率为12.9kW，中心波长为1064.5nm，线宽为3.75nm，M2为1.345，功率稳定性≥95%的多参数可调节的纳秒脉冲激光输出。研究结果表明，改变种子源的重复频率和脉冲宽度能够得到相应参数可调的激光输出。通过优化有源光纤的弯曲半径和光路结构，显著改善了光束质量M2和抗高反能力。该研究方案对研制更高功率和更高峰值功率的MOPA全光纤激光器具有参考意义。

关键词： gallium oxide   thermal management   ion-cutting   bottom packaging   MOSFET   heterogeneous integration   TTR   Raman thermography  

摘要： Gallium oxide (Ga2O3) emerges as a promising ultrawide bandgap semiconductor, which is expected to surpass the performance of current wide bandgap materials, like GaN and SiC, in electronic devices. However, widespread application of Ga2O3 is hindered by its extremely low thermal conductivity and lack of effective device-level thermal management strategies. In this work, Ga2O3 metal-oxide-semiconductor field-effect transistors (MOSFETs) are fabricated by conducting co-integrated design of substrate engineering with layer transferring and device packaging. 3D Raman thermography is introduced as a novel method to analyze the temperature distribution within the device, which provides valuable insights into their thermal performances. A high-quality Ga2O3-SiC heterogeneous integrated material is successfully fabricated with an extremely low interface thermal resistance of 6.67 +/- 2 m(2)K/GW. Compared to the homoepitaxial Ga2O3 MOSFETs, the degradation of I-on/I-off in Ga2O3-SiC MOSFETs is decreased by 1.5 orders of magnitude, and that of R-on is decreased by 31%, showing the great thermal stability of Ga2O3-SiC MOSFETs. With the additional device packaging, a significant one order-of-magnitude reduction in the thermal resistance of the Ga2O3-SiC MOSFET is achieved, reaching a record-low value of 4.45 Kmm/W in the reported Ga2O3 MOSFETs. This work demonstrates an efficient strategy for device-level thermal management in next-generation Ga2O3 power and RF applications.

关键词： 虚拟电厂   紧急频率控制   频率响应模型   参数估计   多灵活资源  

摘要： 为适应直流受端区域电网的双高化发展趋势，针对虚拟电厂（VPP）辅助大电网进行紧急频率控制的问题，提出了一种考虑多灵活资源的VPP频率响应特性建模方法。首先，设计了VPP参与紧急频率控制的“集中-分散式”分层控制架构。其次，分析了多类型灵活资源的频率响应特性，构建VPP的系统频率响应模型及等值模型。然后，基于逐步线性规划算法，提出一种VPP频率响应模型参数估计方法，以兼顾模型的准确性和计算效率。最后，基于PSD-BPA通过对三区域直流受端区域电网仿真测试验证了所提方法的准确性，且参数估计结果具有保守性，能够防止在紧急频率控制决策中系统误判。

关键词： 半导体器件  

ISBN: (纸本)9787111765387

摘要： 本书详细地阐述了宽禁带功率半导体器件的发展现状、电热行为模型建模方法与模型参数提取优化算法、开通和关断过电压问题分析和抑制方法、串扰导通问题机理与抑制方法。通过LLC变换器展示了如何借助宽禁带器件电热行为模型完成功率变换器硬件优化设计和控制算法的仿真验证，并分析了平面磁集成矩阵变压器的优化设计方法，建立了LLC变换器小信号模型，提出并验证了基于LLC变换器小信号模型的输出电流纹波抑制方法。附录提供了基于AnsysQ3D的电路板寄生参数提取方法、宽禁带器件电热行为模型建模方法、遗传算法和列文伯格-麦夸尔特算法组成的复合优化算法实现、LLC变换器小信号模型建模方法，方便读者学习参考。

关键词： 钢铁企业生产短流程   机理数据混合驱动   物质流   能量流   功率特性  

摘要： 钢铁企业用电负荷受企业生产工况和物理过程参数影响较大，现有研究忽略了工序间物质流与能量流的耦合影响，导致仿真精度有所欠缺。针对上述问题，本文提出一种机理数据混合驱动的钢铁企业生产短流程功率模型。首先，基于物质流与能量流在输送过程中的耦合特性，推导得到前后相连工序间物质量和温度值等物理参数的衰减规律；其次，根据设备运行性质构建功率函数通用表达式，并采用核极限学习机拟合物理量和功率函数特性参数；最后，将不同工序功率进行时域叠加，得到短流程钢铁企业总功率曲线。以国内某钢铁企业的实测数据进行算例分析，仿真结果表明，所提模型能更精确地反映其功率特性。

摘要： 英飞凌宣布,已成功开发出全球首项300 mm氮化镓(GaN)功率半导体晶圆技术.英飞凌是全球首家在现有且可扩展的大规模生产环境中掌握这一突破性技术的企业.这项突破将极大地推动GaN功率半导体市场的发展.相较于200 mm晶圆,300 mm晶圆芯片生产不仅在技术上更先进,也因为晶圆直径的扩大,每片晶圆上的芯片数量增加了2.3倍,效率也显著提高.