High sensitivity SPR refractive index sensor with high resolution based on anti-resonance fiber

We proposed a surface plasmon resonance (SPR) refractive index sensor based on hollow-core anti-resonance fiber (HC-ARF). Gold was filled in two symmetric cladding tubes of the fiber, while the analyte was filled in central air holes. The sensing performance was investigated by the finite element method (FEM). Results show that two resonance peaks (a crossing point around 1 030 nm and an anti-crossing region around 1 065 nm) appear for x-polarization (x-pol) core mode coupling with the surface plasmon polariton (SPP) mode. Moreover, the sensitivity was also analyzed. The sensitivity increased with the increase of cladding tube thickness t. The sensor with thickness t=1.7 μm gave a wavelength sensitivity of 7 350—14 790 nm/RIU in the refractive index range of 1.33—1.45 with resonance wavelength from 1 900 nm to 450 nm. Meanwhile, the resolution of 10-6 RIU was achieved. Thanks to high sensitivity and resolution, the proposed sensor has potential applications in glucose detection.     

A high sensitivity localized surface plasmon resonance sensor based on D-shaped photonic crystal fiber for low refractive index detection

In this paper, a localized surface plasmon resonance (LSPR) refractive index sensor based on photonic crystal fiber (PCF) is proposed to solve the problem of low refractive index analyte detection. 31 silver nanowires are placed on the surface of the D-shaped PCF, which increases the contact area between the plasma material and the analyte. The simulation results indicate that the maximum sensitivity of the sensor reaches 16 400 nm/RIU, and the refractive index detection range is 1.26—1.33. It is proved that the sensor has a good prospect in low refractive index detection.     

基于小波去噪与随机森林的配电网高阻接地故障 半监督识别方法

针对配电网高阻接地故障识别易受噪声干扰、无标签数据难以利用的难题,提出一种基于小波去噪与随机森林的高阻接地故障半监督识别方法。区别于监督式学习方法仅利用标签数据,基于协同训练方法能够充分利用有标签数据与无标签数据。首先,使用小波阈值去噪算法消除零序电流中的噪声。其次,采用波峰波谷故障启动算法判断线路是否发生故障或扰动事件。运用小波变换提取零序电流的小波系数作为故障特征。最后,基于小波系数故障特征构建两个随机森林作为半监督分类器进行协同训练,从而实现高阻接地故障的检测识别。仿真结果表明,所提配电网高阻接地故障半监督识别方法可以充分挖掘配电网既有的故障案例中无标注数据蕴含的关键特征,从而提高故障分类准确率,具有较强的准确性和灵敏性。     

含风电的双馈抽水蓄能机组协调调频策略

双馈抽水蓄能机组解耦控制使得其转子无法响应系统频率变化。为提高双馈抽水蓄能机组参与系统调频能力,提出一种转子动能与导叶开度协调控制的控制策略。首先建立水泵水轮机和双馈电机的数学模型,分析双馈抽水蓄能机组进行导叶开度寻优的过程。其次在水泵水轮机的导叶开度控制中加入频率控制环节,同时根据机组可用转子动能整定可变调差系数,确保机组转速始终运行在安全范围内。最后结合两种控制的调频优势,依据频率变化协调两种控制参与调频的比例系数,实现二者平滑切换。在含风机的系统中仿真分析,结果表明所提控制策略在抽水蓄能机组发电、电动工况下均能够提高机组频率响应能力和电网的风电消纳能力。     

基于模型识别相关性的光伏发电并网联络线纵联保护

针对集中式光伏并网系统联络线的保护原理存在适应性问题,提出了一种基于模型识别相关性判别的并网联络线纵联保护方法。在分析并网联络线故障特性的基础上,利用差动电压导数与差动电流的相关系数,对并网联络线区内、外故障模型进行区分：当三相线路的相关系数都接近1时,判定为光伏发电并网联络线的区外故障,而当相关系数小于保护整定值时,判定为区内故障;当单相（或多相）的相关系数小于整定值时,则判定为单相（或多相）故障。建立了光伏发电并网系统模型,仿真结果表明,提出的保护方法不受控制暂态响应影响,具有较强的抗过渡电阻能力。     

基于傅里叶变换与最大关联距离相结合的配电网故障选线新方法

针对现有配电网故障选线准确率与可靠性均偏低的问题,提出了一种基于傅里叶变换与关联距离相结合的故障选线新方法。当发生单相接地故障时,利用快速傅里叶变换（fast fourier transformation,FFT）提取暂态零序电流第2个周期内的稳态工频分量,反演后得到暂态零序电流的整体工频分量;剔除暂态零序电流中工频分量后得到非工频自由振荡分量,在此基础上,利用非工频自由振荡分量构建基于关联距离的故障选线新判据;将具备最大关联距离所对应的线路判定为故障线路,与此同时,将其他线路判定为健全线路。仿真结果表明,该方法不受故障电阻、故障相角与故障位置的影响,同时还可抵抗噪声干扰,具有较强的鲁棒性。