用于射频滤波器的LN单晶薄膜XBAR的性能研究

基于铌酸锂(LN)薄膜的横向激发体声波谐振器(XBAR)能够兼具大机电耦合系数(K²)和高谐振频率(f)特性，有望满足5G应用的频段要求。然而，常规LN薄膜单层XBAR结构的温度稳定性较差，频率温度系数(TCF)较低。该文提出一种具有SiO₂温度补偿层的SiO₂/LN双层结构XBAR，并建立了精确分析层状结构XBAR的有限元模型。理论分析表明，该双层结构XBAR上激励的主模式是一阶反对称(A1)兰姆波。通过合理优化结构参数配置，能够获得高谐振频率(f~4.75 GHz)和大机电耦合系数(K²~8%)，同时其温度稳定性也得到显著改善(TCF~-36.1×10^-6/℃)，相较于单层XBAR结构提高了近70×10^-6/℃，这为研制温补型高频、大带宽声学滤波器提供了理论基础。     

Tunable structure and intensive upconversion photoluminescence for Ho3+-Yb3+ codoped bismuth titanate composite synthesized by sol-gel-combustion (SGC) method

Ho³⁺ and Yb³⁺ codoped bismuth titanate (BTO) composite powders with infrared to visible upconversion luminescence (UCL) function were prepared by SGC method. The effects of Ho³⁺ and Yb³⁺ doping content on the structure and property were investigated for BTO: xHo, 0.2 Yb (x = 0, 0.02, 0.04, 0.06, 0.08, 0.1) and BTO: 0.02Ho, yYb (y = 0.1, 0.2, 0.3, 0.5, 0.7, 0.9) samples. All the samples include three bismuth titanate phases (Bi₄Ti₃O₁₂, Bi₂Ti₂O₇, and Bi₂₀TiO₃₂), and the phase proportion can be tuned by changing Ho³⁺ and Yb³⁺ doping content. These powders are well crystalized with honeycomb-like microscopic structure, and with good absorption for 233 nm, 310 nm and 975 nm wavelength. The band gap can be tuned from 3.53 eV to 4.03 eV when increasing Yb³⁺ content from y = 0 to y = 0.9. A strong 530–580 nm green emission band and a relative weak 630–690 nm red one corresponding to Ho³⁺: ⁵S₂ → ⁵I₈ and ⁵F₅ → ⁵I₈ transitions appear in the UCL spectra for all the BTO: Ho, Yb samples when pumped at 980 nm. The emission intensities can well be tuned with various Ho³⁺ and Yb³⁺ content. The optimal UCL was obtained in BTO: 0.02Ho, 0.5 Yb for all the prepared samples. The energy transfer mechanism is analyzed by building a two-photon energy transfer model, which is proved by the relationship between emission intensities and pumping power measurement. The concentration quenching of Ho³⁺ is caused by cross relaxation of CR₁ and CR₂ (Ho: ⁵F₄, ⁵S₂ + ⁵I₈ → ⁵I₄ + ⁵I₇) and by CR₃ (Ho: ⁵F₄, ⁵S₂ + Yb: ²F_7/2 → Ho: ⁵I₆ + Yb: ²F_5/2) for Yb³⁺ quenching. The mean luminescence lifetime (τ_m) from Ho: ⁵S₂ decreases monotonously with the increase of Ho³⁺ and Yb³⁺ content.     

Ca2+ doping effects on the structural and electrical properties of Na0.5Bi4.5Ti4O15 piezoceramics

Bismuth layer structured Na_0.5Bi_4.5Ti₄O₁₅ (NBT) ferroelectric is one of the most promising materials for potential applications at high temperature. However, it is challenged to achieve a balance between high Curie temperature piezoelectric coefficient and excellent thermal stability for NBT piezoceramics. Here, through chemical modification at the A site of NBT with Ca²⁺, novel (Na_0.5Bi_0.5)_1-xCa_xBi₄Ti₄O₁₅ piezoceramics with excellent properties fabricated by solid state reaction were studied. After doping of Ca²⁺, the Curie temperature T_C increased from 648 °C to 662 °C while the piezoelectric coefficient d₃₃ increased from 14 pC/N to 22 pC/N which can be attributed to the intrinsic contribution of TiO₆ octahedral lattice distortion (tilting and rotation) and the extrinsic contribution of the increased density of domain walls. The composition of (Na_0.5Bi_0.5)_0.95Ca_0.05Bi₄Ti₄O₁₅ ceramics with x = 0.05 has the optimal performance with high T_C of 655 °C, large d₃₃ of 22 pC/N, high electrical resistivity ρ close to 10⁷ Ω cm at 500 °C and especially excellent thermal stability of d₃₃ only about 5% reduction after being annealed at 625 °C. The work effectively reveals the great potential of CNBT-5 ceramics for high-temperature piezoelectric applications.     

Optimal Spectral-Galerkin Methods Using Generalized Jacobi Polynomials

We extend the definition of the classical Jacobi polynomials withindexes α, β>−1 to allow α and/or β to be negative integers. We show that the generalized Jacobi polynomials, with indexes corresponding to the number of boundary conditions in a given partial differential equation, are the natural basis functions for the spectral approximation of this partial differential equation. Moreover, the use of generalized Jacobi polynomials leads to much simplified analysis, more precise error estimates and well conditioned algorithms.Mathematics subject classification 1991. 65N35, 65N22, 65F05, 35J05     

The influence of Johnson–Cook material constants on finite element simulation of machining of AISI 316L steel

In literature, five different sets of work material constants used in the Johnson–Cook's (J–C) constitutive equation are implemented in a numerical model to describe the behaviour of AISI 316L steel. The aim of this research is to study the effects of five different sets of material constants of the J–C constitutive equation in finite-element modelling of orthogonal cutting of AISI 316L on the experimental and predicted cutting forces, chip morphology, temperature distributions and residual stresses. Several experimental equipments were used to estimate the experimental results, such as piezoelectric dynamometer for cutting forces measurements, thermal imaging system for temperature measurements and X-ray diffraction technique for residual stresses determination on the machined surfaces; while an elastic–viscoplastic FEM formulation was implemented to predict the local and global variables involved in this research. It has been observed that all the considered process output and, in particular the residual stresses are very sensitive to the J–C's material constants.     

中国民族乐器的特征值提取和分类

文章处理的是中国民族乐器的识别分类问题.提出了一种基于合适的音频特征值选择方法,该方法在基于MPEG_7标准的声学特征,在特征值的时间特性上进行改造,并加入改进后的特征值.从提取的音频特征值数据集选择K-最近邻算法.特别是对没有加入新特征值和加入新特征值后生成分类器模型的性能进行了比较.实验结果证明新特征值的加入提高了分类器的F1度量值.     

面部表情的深度卷积级联森林识别

面部表情识别广泛应用于各种研究领域,针对面部表情识别使用深度神经网络方法结构复杂、可解释性差和传统机器学习方法特征提取缺乏多样性、识别率低的问题.提出了一种新的深度卷积级联森林(Deep Convolution Cascade Forest,DCCF)方法用于人脸面部表情识别,该方法通过卷积神经网络深度学习人脸面部显著特征,并采用基于随机森林的级联结构森林逐层学习识别出不同的面部表情特征,提高了人脸表情的识别准确率.DCCF在JAFFE、CK+和Fer2013 3个公开面部表情数据集进行了实验,并对面部表情提取的5种特征和7种分类方法进行了比较分析,结果显示DCCF在对比的算法中人脸表情识别性能最好,3个数据集的准确率分别达到91.4%,98.7%,71.6%.     

基于排序损失的ECC多标签代码异味检测方法

代码异味是由糟糕的代码或设计问题引起的一种软件特征,严重影响了软件系统的可靠性和可维护性.在软件系统中,一段代码元素可能同时受到多种代码异味的影响,使得软件质量明显下降.多标签分类适用该情况,将高共现的多个代码异味置于同一标签组,可以更好地考虑代码异味的相关性,但现有的多标签代码异味检测方法未考虑同一段代码元素中多种代码异味检测顺序的影响.对此,提出了一种基于排序损失的集成分类器链(ensemble of classifier chains,ECC)多标签代码异味检测方法,该方法选择随机森林作为基础分类器并采取多次迭代ECC的方式,以排序损失最小化为目标,选择一个较优的标签序列集,优化代码异味检测顺序问题,模拟其生成机理,检测一段代码元素是否同时存在长方法长参数列表、复杂类消息链或消息链过大类这3组代码异味.实验采用9个评价指标,结果表明所提出的检测方法优于现有的多标签代码异味检测方法,F1平均值达97.16%.     

基于Spring MVC及MyBatis的Web应用框架研究

基于EJB等重量级的Web应用框架存在很多问题,如性能差、复杂度高、代码复用率低等,提出了一种B/S结构与C/S结构相结合,采用Spring MVC设计模式和MyBatis为基础的Web应用框架,并对该框架的结构、组成等内容进行分析和研究。以TOPCard信用卡业务系统为应用实例,说明Spring MVC和MyBatis在Web系统中的应用。通过实验结果分析,基于Spring MVC及MyBatis的Web应用框架研究,可以解决性能差、复杂度高、代码复用率低等问题。