课程文献中心 更多>>

关键词： 激光器   固体激光   直接液冷   Zig-Zag光路   薄片激光器  

摘要： 设计了一种分布反射式直接液冷激光放大构型，其采用数十片Nd∶YAG增益介质密集堆叠构成激光增益模块，通过特种冷却液在增益介质间平板微流道内的流动实现直接冷却，同时在增益模块中设计了类似Zig-Zag的放大路径，以实现激光放大输出。本设计效结合了直接液冷激光器高效冷却、无热沉焊接的特点以及Zig-Zag光路像差自补偿及高效提取激光的优势。对激光放大构型、增益介质参数、激光增益、热致波前等开展了理论仿真和分析。搭建了分布反射式直接液冷激光放大实验系统，实现了输出功率为21.2 kW、光光效率为36.9%、光束质量参数β为4.8倍衍射极限的准连续（QCW）线偏振激光输出。本研究验证了分布反射式直接液冷激光放大技术路线的可行性。

关键词： 风电波动   混合储能   优化调度   经验模态分解   滤波器  

摘要： 针对风功率波动平抑场景，以风储联合发电系统作为研究对象，提出一种FIR滤波器结合降幅环节的风电波动平抑策略，同时通过对被滤除波动分量进行基于EMD的自适应重构，从而实现对混合储能的优化调度。该策略选取了一种可以动态调节滤波阶数以及截止频率组合的FIR滤波器进行滤波，并加入FRA降幅环节处理FIR滤波器无法很好平滑的高幅值数据点，从而对风电波动进行平抑；将平抑后滤除的风电波动的作为混合储能功率进行EMD分解，得到的IMF分量通过算法自适应重构为高-低频两组功率分别配给超级电容器-蓄电池，进而得到最优调度参数。通过算例验证分析，加入FRA环节的FIR滤波能够适应不同频率、幅值的功率特性，同时能有效降低1 min、10 min尺度的功率波动，避免了平滑过度导致储能负担加重问题；EMD分解后对IMF分量的自适应重构，能够得到用于补偿不同充放电特性的蓄电池、超级电容器功率，显著降低了储能出力负担同时提高调度经济性。

关键词： Topology   Switches   Electrochemical processes   Stress   Capacitors   Electrodes   Semiconductor diodes   MOSFET   Inductors   Indium phosphide   Green hydrogen   nonisolated converters   ultra-high gain   pulsed electrolysis  

摘要： This article introduces a family of nonisolated dc-dc converters designed to amplify lower dc voltages to higher levels. These converters are particularly suitable for use in the dc link of pulse converters in hydrogen electrolyzers. The primary structure serves as a base for additional configurations with greater voltage amplification and reduced component stress. Two supplementary categories are examined based on the main structure, offering improved functionalities, and decreased pressures. The proposed topologies' specific parameters are evaluated in both ideal and nonideal operating scenarios. In addition, thorough comparisons are made among the converters from various perspectives. These topologies feature a single switch and minimal voltage stress, which provides a significant advantage in high-voltage scenarios by limiting the number of switches and their associated stresses. This study achieves a notable increase in voltage amplification without relying on high-frequency transformers, resulting in increased power density and reduced volume and mass. To validate the proposed designs, experimental findings are provided and compared with theoretical predictions.

关键词： Rectifiers   MOSFET   Silicon   Reliability   Gallium nitride   Silicon carbide   Performance evaluation   Capacitors   Inductors   Topology   Electric vehicles   gallium nitride transistors   silicon   silicon carbide   Vienna rectifier   wide bandgap  

摘要： This study examines the reliability of several electronic components in a Vienna Rectifier configuration, which is a critical topology for power conversion systems. Component selection has become increasingly important over the past few years since the power electronics of today demand more efficiency, power density, and operational reliability. Extreme reliability testing such as temperature cycling, electrical overload, and long duration of high-frequency operation was a part of the study. GaN MOSFETs had an edge over Si and SiC MOSFETs in several aspects, such as decreased conduction and switching losses, better thermal management, and more consistent performance with time. While GaN MOSFETs performed better in general and especially at high frequencies and temperatures, SiC MOSFETs showed some improvements over the conventional Si devices. Capacitors, diodes, MOSFETs, and inductors are put to test for reliability under different stress conditions. The combination of diodes and GaN MOSFETs showed a synergistic effect in improving system dependability and reducing temperature-induced degradation. This is a significant result. The combined effects allowed active and passive parts to last longer and function more reliably. These results open insights into selecting components for systems in the automotive and aerospace industries, which mostly rely on reliability.

关键词： 智能控制算法   光伏发电系统   最大功率点跟踪技术  

摘要： 阐述最大功率点跟踪技术的分类，对比基于扰动观察算法、增量电导算法、模糊逻辑控制算法、人工神经网络算法的最大功率点跟踪技术，分析智能控制算法的特点、性能评估标准。

关键词： Bridge circuits   Sensitivity   Resistors   Sensors   Topology   MOSFET   Voltage measurement   Bridges   Oscilloscopes   Time measurement   AC bridge circuit   capacitive sensor   direct interface resistive sensor   sensitivity analysis   square alternating signal   step function   U(t) signal   wheatstone bridge  

摘要： Many physical variables of interest (movement, force, torque, power, pressure, flow, or temperature) are usually measured with sensors whose output (of an electrical parameter) varies little with respect to those physical variables of interest. Thus, to have adequate accuracies and resolutions, it is usually configured in a differential way through an electrical circuit known as a Wheatstone Bridge. This article presents a new Wheatstone Bridge topology that with square alternating signal excitation increases the traditional sensitivity in voltage difference in the two branches of the Bridge with a sensor element ( epsilon/4) to a sensitivity of ( epsilon /e), where ( epsilon ) is the difference in the sensor (as per one) caused by the physical variable of interest. It is analyzed analytically, then simulated with calculation, and finally measured in an ad hoc set-up.

关键词： 海上风能   深度学习   功率预测   自注意力   敏感度分析   小波分解  

摘要： 为了提高海上风电功率预测的精度以及其预测结果的可信度，提出了一种融合可学习小波的自注意力模型，有效提升了海上风电功率预测的精度，并且具备一定的预测过程分析能力。首先，将小波分解与深度学习模型融合，使模型具备从不同频域提取特征的能力。其次，构建稀疏自注意力预测网络，实现对全局信息特征的有效提取，提高模型的预测性能。接着，提出一种时序“敏感度”量化分析方法，在多维度下对海上风电参量输入进行重要性评估，在一定程度上对预测机理进行合理的分析。最后，基于风场实际运行数据进行相关实验。实验结果表明，相较对比模型，所提模型在海上风电超短期预测任务上具有更高的预测精度。

关键词： 变压器顶层油温   极限学习机   热模型   融合预测   负荷形态聚类  

摘要： 提出一种采用极限学习机对IEEE导则中的变压器顶层油温热模型计算偏差进行预测和修正，从而实现对变压器顶层油温精确预测的融合预测方法。首先，介绍了变压器顶层油温热模型和极限学习机预测模型各自的特性。其次，为避免采用两级智能预测导致运算速度慢的问题，采用加权多点外推法结合负荷形态聚类算法预测变压器未来时段负载系数，作为模型的负荷预测级。最后，利用变压器顶层油温热模型获取相应的油温计算值，并利用极限学习机对计算值与实测值间的偏差进行预测，最终得到变压器顶层油温的精确预测值。搭建了仿真模型对所提方法进行了验证，仿真结果表明，所提预测方法平均预测误差率仅为0.59%，均方根误差仅为0.47℃，相比其他3种方法有更高的预测精度和稳定性，模型训练和预测时间分别只有1.21ms和0.39 ms，证明了所提出和建立的融合预测模型具有较高的预测精度、稳定性和运算速度。

关键词： 变压器   局部放电   瞬时能量特征   排列熵   最近邻图算法  

摘要： 局部放电(PD)定位是一种能早期预测变压器绕组即将发生的可靠和必要的技术。不同的PD源会受到绝缘的影响，在不同位置测量的PD表现出独特的能量特征。提出一种基于能量特征的绕组内部PD定位方法，为了揭示PD信号的多尺度本征特征，使用集成经验模式分解(EEMD)提取本征模函数(IMF),计算排列在前几个IMF的瞬时能量分布值以捕获PD能量，采用快速主成分(F-PCA)进行特征降维，提出一种基于自适应密度的空间聚类-最近邻图(ADBSCAN-NNG)算法识别局部高密度能量分量，并采用高频电流传感器(HFCT)对真空铸造的变压器绕组验证所提算法的可行性。定位结果表明，算法能够实现对变压器绕组PD信号的定位，且算法的复杂度大大降低，有效提高定位的精度和速度，具有一定的工程实践指导意义。

关键词： 光纤传感技术   光频域反射计(OFDR)   现场可编程门阵列(FPGA)   实时处理   可变系数滤波器   噪声消除  

摘要： 随着光频域反射计（OFDR）探测精度（<-100 dB）的提高，谐波噪声对探测的干扰逐渐增大，可能会导致OFDR系统出现传感误差或者误判。在传统的信号处理中，可以采用滤波器对谐波噪声进行消除；但是OFDR信号具有时变特性，传统滤波器无法完全消除谐波噪声。鉴于此，本课题组提出了一种实时短时频率-可变系数有限长脉冲响应（STF-VCFIR）滤波器系统。该系统通过测量辅助干涉仪信号的短时频率、动态调节滤波器系数、对探测信号进行实时滤波处理，改善了传统滤波器导致的谐波泄漏的缺点，大幅降低了噪声的影响。最终通过实验证明了该系统可在实时处理的同时实现对谐波噪声的消除。与无滤波系统相比，该系统对谐波噪声的最大消除可达8 dB以上，使得谐波噪声降低到-102.7 dB以下，达到与噪声均值相同的水平，减小了将谐波噪声误判为反射点的概率，提高了系统判别的准确性。