含La或Sm的BaNd_(2-y)R_yTi_(4+x)O_(12+2x)瓷的结构与性能

ＢａＯ－Ｎｄ２Ｏ３－ＴｉＯ２系富钛区含Ｌａ或Ｓｍ的陶瓷,ＸＲＤ、ＳＥＭ显微结构分析表明,其主晶相为Ｒ（Ｒ＝Ｌａ或Ｓｍ）取代Ｎｄ固溶体ＢａＮｄ２－ｙＲｙＴｉ５Ｏ１４和ＢａＮｄ２－ｙＲｙＴｉ４Ｏ１２,均属斜方晶系,并存在少量的次晶相Ｒ２Ｔｉ２Ｏ７、ＢａＴｉ４Ｏ９、Ｂａ２Ｔｉ９Ｏ２０等。富钛区ＢａＮｄ２Ｔｉ４＋ｘＯ１２＋２ｘ（ｘ＝０～１．４００）组分和含Ｌａ或Ｓｍ的ＢａＮｄ２－ｙＲｙＴｉ５Ｏ１４组分所组成的瓷料,介质损耗显著地降低,介电常数和电容温度系数遵从李赫涅德凯对数混合定律。     

基于树结构分层随机森林在非约束环境下的头部姿态估计

头部姿态估计是人类行为和注意力的关键,受到光照、噪声、身份、遮挡等许多因素的影响。为了提高非约束环境下的估计准确率和鲁棒性,该论文提出了树结构分层随机森林在非约束环境下的多类头部姿态估计。首先,为了消除不同环境的噪声影响,提取人脸区域的组合纹理特征,对人脸区域进行积极人脸子区域的分类,分类结果作为树结构分层随机森林的先验知识输入;其次,提出了一种树结构分层随机森林算法,分层估计多自由度下的头部姿态;再次,为了增强算法的分类能力,使用自适应高斯混合模型作为多层次子森林叶子节点的投票模型。在多个公共数据集上的多种非约束实验环境下进行头部姿态估计,最终实验结果表明所提算法在不同质量的图像上都有很好的估计准确率和鲁棒性。     

高功率CO2激光器换热性能的研究

高功率CO2气体激光器的换热性能好坏是器件长期稳定运行的关键环节之一。本文以板翅式热交换器为例阐明了热交换器的计算和设计方法。     

水泥配料控制中控制算法的研究

配料是水泥生产过程的重要环节,在一些连续称量配料工艺过程中,以往大多通过控制每种物料的瞬时流量来满足工艺要求,但往往误差较大,给生产带来不良影响。为提高控制精度,提出了一种偏差加权累积量控制PID算法,并在水泥生料配料控制中进行了实施,实验结果表明该算法能有效地抑制和消除系统过量超调和不停振荡,且对于某些不要求每个瞬时值都控制给定值的量,该算法可以避免调节阀的频繁动作并大大提高调节品质。偏差加权累积的PID控制算法具有简单、方便、易于实现的特点,且在水泥生料配料控制中取得了理想的效果,有很好的推广前景。     

高浓度固液混合流的湍流模型

建立了高浓度固液混合流的湍流模型，分析了混合流的应力（包括颗粒碰撞应力、粘性应力、湍流应力及Ｃｏｕｌｏｍｂｉｃ摩擦应力）。根据高浓度混合流中颗粒的受力，得出了颗粒运动的力平衡方程。由于引入颗粒碰撞应力和Ｃｏｕｌｏｍｂｉｃ摩擦应力，本模型较其它模型更适合具有高浓度和大颗粒的流场。利用本模型模拟的高浓度含沙水流在水平管道中的流动特性，其数值结果与前人所做的实验结果较一致。     

利用性别判识技术的说话人识别系统的研究

本文在基于GMM的与文本无关说话人识别系统中考虑性别差异,提出了把不同性别的说话人特征参数归成不同的两个子集,识别时先判断待识别说话人特征是属于哪个性别子集后,再在相应的子集中进行识别的方法。实验结果表明该方法能够改善基于GMM的与文本无关说话人识别系统的识别率。     

Novel photoresist stripping technology using steam-water mixture

A novel wet vapor photoresist stripping technology is developed as an alternative to dry plasma ashing and wet stripping.Experiments using this technology to strip hard baked SU-8 photoresist,aurum and chromium film are carried out.Then the images of stripping results are shown and the mechanism is analyzed and discussed. The most striking result of this experiment is that the spraying mixture of steam and water droplets can strip photoresist and even metal film with ease.     

Complex‐Shaped Cellulose Composites Made by Wet Densification of 3D Printed Scaffolds

Cellulose is an attractive material resource for the fabrication of sustainable functional products, but its processing into structures with complex architecture and high cellulose content remains challenging. Such limitation has prevented cellulose‐based synthetic materials from reaching the level of structural control and mechanical properties observed in their biological counterparts, such as wood and plant tissues. To address this issue, a simple approach is reported to manufacture complex‐shaped cellulose‐based composites, in which the shaping capabilities of 3D printing technologies are combined with a wet densification process that increases the concentration of cellulose in the final printed material. Densification is achieved by exchanging the liquid of the wet printed material with a poor solvent mixture that induces attractive interactions between cellulose particles. The effect of the solvent mixture on the final cellulose concentration is rationalized using solubility parameters that quantify the attractive interparticle interactions. Using X‐ray diffraction analysis and mechanical tests, 3D printed composites obtained through this process are shown to exhibit highly aligned microstructures and mechanical properties significantly higher than those obtained by earlier additively manufactured cellulose‐based materials. These features enable the fabrication of cellulose‐rich synthetic structures that more closely resemble the exquisite designs found in biological materials grown by plants in nature.     

固液混合物体热成像图的识别处理方式研究

热成像图能够反映物体辐射的性能,可在图像处理后应用于固体与液体材料的识别,而固液混合物热成像图的图像处理是技术难点。文中旨在从固液混合物的热成像图中提取内部特征及轮廓特征,先进行包含线性灰度变换、直方图修正、PCA分析的不同顺序及组合方式的图像修正处理,后进行包含MSER特征点与Prewitt特征边界的提取,最终对处理后的图像进行比较分析,得到三条图像处理路径,分别可用于固体成分分析、液体成分分析和物体外部轮廓分析。该处理方式可用于垃圾分类与工业物料识别,具有广阔前景。