Ensemble relation network with multi-level measure

Fine-grained few-shot learning is a difficult task in image classification. The reason is that the discriminative features of fine-grained images are often located in local areas of the image, while most of the existing few-shotlearning image classification methods only use top-level features and adopt a single measure. In that way, the localfeatures of the sample cannot be learned well. In response to this problem, ensemble relation network with multi-level measure (ERN-MM) is proposed in this paper. It adds the relation modules in the shallow feature space tocompare the similarity between the samples in the local features, and finally integrates the similarity scores from thefeature spaces to assign the label of the query samples. So the proposed method ERN-MM can use local details andglobal information of different grains. Experimental results on different fine-grained datasets show that the proposedmethod achieves good classification performance and also proves its rationality.     

TEM method for rapid and precise evaluation of continuous recrystallization

With the aid of a double-tilt holder in a transmission electron microscope, a new method for a rapid and precise determination of the misorientations of a large number of subgrain or grain boundaries is given in this article. By use of the method, the continuous recrystallization can be rapidly and precisely evaluated when compared to the other conventional methods.     

Effects of Electric Potentials on Tribological Properties of SiC/HT200 Pairs in NaOH Solutions

On a self-developed circular-translation polishing machine, the influence of different electric potentials on the friction coefficient and wear properties of SiC/HT200 friction pair, which contacted as one electrode, were studied in NaOH solutions. The results show that the friction coefficient of SiC/HT200 pair is 0.25 under no externally applied potential; but it decreases to 0.23 while applying +3 V polarization potential on the pair; in contrast, it increases to 0.30 when applying -3 V polarization potential on the pair. And the polished SiC surface under +3 V potential is much smoother, while there are clear ploughing marks and adhesive region on the polished SiC surface under zero and -3 V potential, respectively. The mixed tribo-electrochemical friction model has been proposed and the mechanisms of the electric effects on the tribological properties of SiC/HT200 friction pair in NaOH electrolyte are addressed. Results and discussion show that the anodic passivation technique may be used for hastening the running-in process of SiC/Fe friction pairs, as well as for polishing SiC to increase its materials removal efficiency while maintaining good surface quality.     

薄板坯连铸中碳钢SS400纵裂原因分析及控制措施

通过对唐钢薄板坯连铸中碳钢SS400纵裂产生原因进行分析,提出了优化保护渣性能、结晶器水量和铜板厚度相匹配等纵裂控制措施.通过采取这些技术措施,使薄板坯纵裂缺陷得到了有效控制.     

基于“项目驱动法”的数字电路课程教学的探索与实践

针对课程教学中普遍存在的问题,结合多年的教学与科研经验,提出以学生为中心,以《数字电路》应用能力为核心,将项目驱动法引入课程教学的全过程。实践证明,此方法切实可行,效果显著。     

基于中值滤波的 OFDR 二维形状感知精度提升研究

基于光频域反射(OFDR)的二维形状传感技术存在相干噪声和光谱匹配错位,导致瑞利散射波长偏移曲线出现乱峰、 杂峰,严重影响形状重构精度。 为此,提出一种使用中值滤波的二维形状传感精度提升方法,分析了 Frenet-Serret 标架下的 OFDR 二维形状重构原理和利用中值滤波提升传感精度的基本原理,搭建了 OFDR 二维形状传感系统,通过多组静态应变实验 标定了应变-波长偏移系数,利用中值滤波方法对 3 种预设二维形状的瑞利散射波长偏移曲线进行降噪处理。 实验结果表明, 应变-波长偏移系数标定为 1. 20 με / pm;在 0. 5 m 形状传感长度上,3 种形状传感末端误差分别从 3. 08%、0. 94%、0. 82%降低至 0. 80%、0. 66%、0. 48%。 研究表明,利用中值滤波降噪可以实现 OFDR 二维形状传感系统的感知精度提升。     

基于改进样板机法的风光互补新能源电站容量配比优化

风光等新能源电站出力具有间歇性和波动性,合理的风光容量配比可以充分实现二者的互补。不准确的电站理论功率计算会影响风光真实特征的提取,进而导致较大的容量配比误差。本文在新能源电站理论功率计算所通常采用的样板机法基础上进行改进,首先对异常数据进行识别及重构,然后识别新能源电站中的异常样板机并更新样板机集合,进一步根据非样板机的实际运行情况选取动态信息窗,利用动态信息窗内样板机和非样板机的实测功率,对非样板机分组并动态识别每组非样板机的比例系数,从而计算新能源电站的理论功率。基于多年历史理论功率对新能源电站进行特征分析,模拟随机出力场景,并进行场景筛选,建立了基于源荷不匹配风险的风光配比优化方法。通过算例验证了改进样板机法的准确性,利用该方法得到西北某地区电网风、光伏电站多年的改进理论功率数据,并优化得到该地区的风光最优配比。     

L-半胱氨酸辅助反相微乳法制备Ag_2S纳米晶

选用TX-100/环己烷/正己醇/水反相微乳体系,以Ag NO_3为Ag源、Na_2S·9H2O为S源、L-半胱氨酸为结构导向剂,室温下制备了高晶化度、直径为10—50 nm的Ag_2S纳米晶,并考察了影响其粒径大小的因素。紫外-可见吸收光谱表明,Ag_2S纳米晶在290 nm处呈现强吸收,与体相Ag2S(1 240 nm)相比,其吸收边发生了明显的蓝移。通过调控微乳体系中ω0(水与表面活性剂物质的量比)、Ag+与L-半胱氨酸的物质的量比、反应物浓度等可以实现对产物尺寸和形貌的调控。     

Composition-Controllable Syntheses and Property Modulations from 2D Ferromagnetic Fe5Se8 to Metallic Fe3Se4 Nanosheets

Exploring new-type 2D magnetic materials with high magnetic transition temperature and robust air stability has attracted wide attention for developing innovative spintronic devices. Recently, intercalation of native metal atoms into the van der Waals gaps of 2D layered transition metal dichalcogenides (TMDs) has been developed to form 2D non-layered magnetic TMDs, while only succeeded in limited systems (e.g., Cr₂S₃, Cr₅Te₈). Herein, composition-controllable syntheses of 2D non-layered iron selenide nanosheets (25% Fe-intercalated triclinic Fe₅Se₈ and 50% Fe-intercalated monoclinic Fe₃Se₄) are firstly reported, via a robust chemical vapor deposition strategy. Specifically, the 2D Fe₅Se₈ exhibits intrinsic room-temperature ferromagnetic property, which is explained by the change of electron spin states from layered 1T'-FeSe₂ to non-layered Fe-intercalated Fe₅Se₈ based on density functional theory calculations. In contrast, the ultrathin Fe₃Se₄ presents novel metallic features comparable with that of metallic TMDs. This work hereby sheds light on the composition-controllable synthesis and fundamental property exploration of 2D self-intercalation induced novel TMDs compounds, by propelling their application explorations in nanoelectronics and spintronics-related fields.     

功率场效应管在线性功率放大器中的应用

论述了功率场效应管的线性应用。实验结果表明,应用功率场效应管的功率放大器比一般晶体管放大器电路更简单,特性更好。