基于耦合声场建模的膨胀腔消声器声学性能分析及试验

针对圆柱形膨胀腔消声器三维建模及声学性能分析问题, 提出一种基于切比雪夫变分原理的耦合声场建模方法, 建立三维圆柱形膨胀腔消声器理论模型并搭建试验台架, 传递损失试验结果验证了理论模型的准确性. 将膨胀腔消声器内部声场分解为多个子声场, 基于子声场间压力与质点振速连续性条件, 推导声场耦合变分公式, 构建子声场拉格朗日泛函. 将子声场声压函数展开为切比雪夫-傅里叶级数形式, 通过瑞利-里兹法求解膨胀腔消声器频率、声压响应及传递损失. 计算并对比分析扩张比、扩张腔长度、进出口管偏置对膨胀腔消声器消声性能的影响. 结果表明: 扩张比增大会有效提高消声器在低频段的消声性能, 进出口管的偏置对消声器消声性能影响很小.     

Insights into MWCNTs/polyimide nanocomposites: from synthesis to application as free-standing flexible electrodes in low-cost micro-supercapacitors

In the pursuit to enlarge the library of polyimide materials for energy applications, new polyimide/MWCNTs composite films have been developed by MWCNTs-assisted polycondensation reaction of a hydroxyl and triphenylmethane-containing diamine with benzophenone tetracarboxylic dianhydride targeting to highlight their electrical storage capability as flexible electrodes in micro-supercapacitors (mSCs). The Fourier-transform infrared spectroscopy, proton nuclear magnetic resonance, UV–vis, fluorescence, and Raman spectroscopies were used to demonstrate the evolution of interfacial interactions between MWCNTs and the precursors (diamine monomer and intermediate polyamidic acid) and polyimide matrix that proved to be the origin of MWCNTs homogeneous dispersion. Thus, composite films incorporating 1, 3, 5, and 10 w.t.% MWCNTs were obtained and thoroughly investigated with regard to their morphology, mechanical behavior, thermal stability, and electrical conductivity. The electrochemical performance of these composites was first analyzed in a classical three-electrode cell by cyclic voltammetry and galvanostatic charge-discharge in both aqueous and organic electrolyte systems. By far, the best electrical storage capacity was obtained with the composite polyimide film containing 10% MWCNTs that was further used as both active material and current collector in a flexible symmetric mSC realized by a straightforward and low-cost procedure. In the attempt to better exploit the advantages of this composite film, it was layered with a graphite-containing paint and tested as an electrode in a flexible mSC, which provided satisfactory results. To our knowledge, this is the first report on the electrical charge storage capability of a polyimide/MWCNTs free-standing film as a flexible electrode in mSCs, which do not require time- and resource-consuming processing steps.     

紫色球杆菌视紫红质光谱特性的机器学习研究

近年来，机器学习等人工智能技术被应用于蛋白质工程，其在蛋白质结构、功能预测、催化活性等研究中具有独特优势。在未知蛋白质结构的情况下，将蛋白质序列和功能特性与机器学习相结合，基于序列-活性关系（innovative sequence-activity relationship，ISAR）算法，将蛋白质氨基酸序列数字化，用快速傅里叶变换（fast four transform，FFT）进行预处理，再进行偏最小二乘回归建模，可在数据集较少情况下拟合得到最佳模型。通过机器学习对紫色球杆菌视紫红质（gloeobacter violaceus rhodopsin，GR）的突变体蛋白质氨基酸序列与光谱最大吸收波长进行建模，获得了最佳模型。用最佳索引LEVM760106建模得到的确定系数R² 为0.944，均方误差E为11.64。用小波变换进行的预处理，其R² 虽也约为0.944，但E大于11.64，不及FFT进行的预处理。方法较好地解决了蛋白质序列与功能特性之间的数学建模问题，在蛋白质工程中可为预测更优的突变体提供支持。     

明代生命损失型地震的时空分布特征及成因分析

利用史料记录建立了明代生命损失型地震的数据库，构建生命损失型地震等级序列、10 a尺度上生命损失型地震的频次序列和县次序列，再现明代生命损失型地震的时空分布特征，探讨了生命损失型地震发生及导致人口死亡的时空差异原因。结果表明：（1）从明初到明末，生命损失型地震等级强度有逐渐扩大的趋势。（2）明代前期处于生命损失型地震发生较少的稳定时期；明代中后期处于生命损失型地震发生较多的波动时期。明代生命损失型地震发生频次存在明显的周期性变化特征。（3）生命损失型地震在明代大部分区域皆有分布。灾害重心随时间变化发生移动，由西南到中部再到华北、华东。（4）明代地震频繁的原因除地球本身因素外，太阳活动通过影响气候变化，进而间接影响地壳活动。气候寒冷与干旱和地震有较好的对应关系。明代后期生命损失型地震导致死亡人数大量增加，与明代经济发展，人口密度逐渐增长有关；不同地区生命损失型地震导致的死亡人数不同，且死亡人数与地震频发地区是否处于板块之间的活动构造带和人口分布密度有关。     

LAMOST光谱J152238.11+333136.1 P-Cygni轮廓分析

LAMOST一期巡天成功获取河外星系光谱超过150 000条,大样本光谱数据为探索奇异、稀有的天体从而完善现有的天体演化理论提供了必要的数据条件;而先进的信息技术为从海量的数据中挖掘这些珍稀样本提供了有效途径。针对采用基于DoPS的数据挖掘方法,从LAMOST DR5星系光谱数据获得的离群数据挖掘结果中,呈现出疑似P-Cygni轮廓特征的光谱J152238.11+333136.1进行了深入讨论。首先针对该光谱的基本信息、疑似P-Cygni轮廓特征以及相应的离群数据挖掘方法进行了简要表述,光谱在Hβ和[OⅢ]λ4860处呈现P-Cygni轮廓,在NeⅢλ3869和HeⅠλ5874处呈现反P-Cygni轮廓;其次,对该特征的真实性及其生成机制从以下4个角度展开讨论。(1)交叉同源观测。Sloan巡天2004年(相差11年)的同源观测,其光谱上并未呈现对应的特征,据推测可能是正在进行的演化活动或者光纤定位误差所致;(2)通过分析光谱质量、减天光残差等方法,分析P-Cygni特征是否为观测或数据处理所致。NeⅢλ3869和HeⅠλ5874处呈现反P-Cygni轮廓可信度较低;同时,通过比较目标光谱与超级天光,以及相邻光纤观测到的光谱在对应波长处的光谱特征,说明存在P-Cygni轮廓为减天光过程导致的可能性;(3)光谱子型差异。IRAS和WISE等近红外同源观测,显示其为Seyfert 2型星系,光学波段发射线强比[NⅡ]/Hα,[OⅢ]/Hβ显示其为HⅡ区,结合光学、红外测光图像特征,推测目标可能是两个星系进行并合活动;(4)从导致P-Cygni轮廓的物理机制的角度,分析了由星系并合触发外流、由恒星形成(爆发)电离气体触发的外流以及由Wolf-Rayet特征星系的超星风等原因引起的可能性。     

Time-averaged flow characteristics and mixing properties of double-concentric jets

We examined the flow behaviors and mixing characteristics of double-concentric jets using laser-assisted smoke flow visualization method to analyze typical flow patterns and binary boundary detection technique to investigate jet spread width. Time-averaged velocity vectors, streamline patterns, velocity distributions, turbulence properties, and vorticity contours were analyzed using Particle Image Velocimetry (PIV). Topological flow patterns were analyzed to interpret the vortical flow structures. Mixing properties were investigated using a tracer-gas concentration detection method. Four characteristic modes were observed: annular flow dominated mode, transition mode, central jet dominated mode-low shear, and central jet dominated mode-high shear. The jets’ mixing properties were enhanced by two major phenomena: the merging of annular flow and central jet at the centerline and the large turbulence fluctuations produced in the flow field. The merging of the jets induced stagnation points on the central axis in the annular flow dominated mode, which caused reverse flow on the central axis and drastic turbulence fluctuations of the near field region. When the central jet penetrated the recirculation region in the other three modes, the stagnation points on the central axis and the reverse flow vanished. Therefore, the mixing behaviors were prominently enhanced in the annular flow dominated mode.     

空中目标多波段红外辐射特性描述与仿真分析

针对空中目标在复杂背景下的探测需求,根据实际目标的运动特性,分析目标在飞行高度、飞行姿态角改变时的辐射特点,基于MODTRAN计算得到大气辐射和衰减数据,建立目标的三维模型、热辐射和反射模型,搭建空中目标的红外成像仿真系统.分析和仿真结果表明:在中波波段,目标尾焰的红外辐射比蒙皮强很多,在长波波段,蒙皮的红外辐射比较强,仿真图像的细节比较多,尾焰的红外辐射虽然有所减弱,红外成像效果依旧很好;相同探测条件下,由于位置越高大气越稀薄,探测器的可探测距离会变得比较远.目标红外辐射特性的分析和红外仿真系统的搭建对缩短红外探测器的研制周期和进一步确定探测器波段和系统分辨率等指标提供了参考依据.     

Prediction of branch on branch and topological characteristics of low‐density polyethylene polymerization by a novel stochastic approach

A novel approach using Monte Carlo method applied to simulation of low‐density polyethylene (LDPE) polymerization in tubular reactor showing topological characteristics, and the comprehensive kinetic mechanism has been taken into consideration. The results show the precise details of the structure of a chain in the three levels of the backbone, the main branches, and branches on branch. The chain types include dead polymer, dead polymer with unsaturated end, and live polymer with primary radical, secondary radical, and tertiary radical. In this work, the branches on branch were identified in terms of number, length, and position of the branch. Sixty percent of branches on branch are 1 to 5 carbons long, and the longest branch on branch is about 50 carbons. Thus, this study provides a tool for more accurately mapping the polymer chains architecture, superior to determine the number, and position of long‐ and short‐chain branches in past researches. Finally, this approach will advance the prediction of microstructure‐related properties of polymer one step further.     

贝叶斯优化卷积神经网络公共场所异常声识别*

针对公共场所异常声的感知和识别问题,提出一种基于贝叶斯优化卷积神经网络的识别方法。提取声信号的Gammatone倒谱系数、倍频程功率谱、短时能量和谱质心,组合成声信号的特征图。构建卷积神经网络作为分类器,利用递增的卷积核设置和池化操作处理不同尺度的特征。基于贝叶斯优化算法优化卷积神经网络的模型参数,对包括火苗噼啪声、婴儿啼哭声、烟花燃放声、玻璃破碎声和警报声的5种公共场所异常声进行识别。该方法的识别结果与基于不同的特征提取和分类器方案得到的识别结果进行比较,结果表明该方法的识别效果优于其他特征提取和分类器方案的识别效果。最后分析了该方法在不同信噪比噪声干扰下的识别结果,验证了该方法的有效性。     

氩气感应耦合等离子体速度与温度数值模拟

为了获得用于研究再入飞行器热防护系统的感应耦合等离子体风洞流场数据，基于流场、电磁场和化学场的多场耦合建立了非平衡态感应耦合等离子体数值模型。利用该模型对不同入口质量流率和不同工作压力下的感应耦合等离子体进行了数值模拟，得到了相应工作参数下感应耦合等离子体温度与速度的分布特性。计算结果表明：等离子体中心线上的速度随着入口质量流率的增大而增大，而随着工作压力的增大而减小；同时，等离子体中心线上的温度随着入口质量流率的增大而减小，而随着压力的增大先减小后增大。这些结果可为感应耦合等离子体风洞优化设计及其工业应用提供理论指导。