基于涂层显微组织图片的拉伸实验有限元数值模拟

针对热障涂层,利用数字图像处理技术与有限元网格模型生成原理相结合的方法,生成与实际ZrO2陶瓷层显微组织图片一致的介观尺度的有限元网格模型.利用LS-DYNA有限元方法进行ZrO2陶瓷层在不同拉应力作用下的裂纹扩展模拟.数值模拟结果表明:在考虑了ZrO2陶瓷层所包含的孔洞、微裂纹等缺陷的基础上,裂纹从已有缺陷处开始迅速扩展,并最终导致ZrO2陶瓷层断裂失效.数值模拟结果与实验结果吻合良好.     

Fine mapping of <Emphasis Type="Italic">qHUS6.1</Emphasis>, a quantitative trait locus for silicon content in rice (<Emphasis Type="Italic">Oryza sativa</Emphasis> L.)

Silicon is essential for optimal growth of rice (Oryza sativa L.). This study was conducted to fine map qHUS6.1, a quantitative trait locus (QTL) for rice hull silicon content previously located in the interval RM510–RM19417 on the short arm of chromosome 6, and to analyze the effect of this QTL on the silicon content in different organs of rice. Selfed progenies of a residual heterozygous line of rice were detected using 13 microsatellite markers in the vicinity of qHUS6.1. Three plants with overlapping heterozygous segments were selected. Three sets of near isogenic lines (NILs) were developed from the selfed progenies of the 3 plants. They were grown in a paddy field and the silicon contents of the hull, flag leaf, and stem were measured at maturity. Based on analyses of the phenotypic distribution and variance among different genotypic groups in the same NIL set, a significant genotypic effect was shown in the NIL set that was heterogenous in the interval RM19410–RM5815, whereas a significant effect was not found in the remaining 2 NIL sets that were heterogenous in either of the intervals RM4923–RM19410 or RM19417–RM204. On comparison among the physical positions of the 3 heterogenous segments, qHUS6.1 was delimited to a 64.2-kb region flanked by RM19410 and RM19417 that contains nine annotated genes according to the genome sequence of Nipponbare. This QTL showed strong effects on all of the three traits tested, and the enhancing alleles were always derived from the paternal line Milyang 46. The present study will facilitate the cloning of qHUS6.1 and the exploration of new genetic resources for QTL fine mapping.     

基于机器视觉的烹饪机器人锅具识别算法

全面介绍了基于机器视觉的烹饪机器人锅具识别算法,用于提取复杂背景中的锅具曲线.算法分为两个阶段:特征点确定、图像匹配.使用曲率作为特征点,给出了数字图像的离散特征点计算方法和加权Hausdorff距离识别锅具轮廓特征点.解决了图像因为旋转、位移、缩放造成图像形变以及由于锅具动作图像部分遮挡、动态模糊造成识别困难问题.实验表明该锅具识别算法能够高效地提取画面中的锅具轮廓,能较好地满足烹饪机器人示教系统的要求.     

一种基于SUSAN算子和相位相关实现图像配准的算法

提出了一种基于SUSAN算子检测角点,然后利用角点为中心的邻域的相位相关性实现图像配准的新算法,并在Matlab环境中模拟验证了该算法的有效性.实验表明,该算法配准精度高,对灰度变化和噪声具有鲁棒性.     

Fast Control of Negative Pressure Response of a Bio-inspired Suction Cup Actuated by Shape Memory Alloy

A bio-inspired suction cup actuated by drape memory alloy (SMA) for miniature wall climbing robots ls developed based on studying characteristics of biological suction apparatus. Some fast control strategies are introduced to improve negative pressure response. model of the suction cup is built,and simulation and experiments results indicate the effectiveness of the fast control strategies. The largest negative pressure of the suction cup can reach 14 000 Pa,and its generating and cancelling just need 5 s. Research results Indicate the suction cup can be used as an adhesion mechanism for miniature wall climbing robots.     

基于改进自适应粒子群算法的混合核函数最小二乘支持向量机大坝变形预测

针对大坝变形影响因素的复杂性以及监测数据的非线性、随机波动大和预测难度大等问题,提出一种改进自适应粒子群(particle swarm,PSO)算法的混合核函数最小二乘支持向量机(least squares support vector machine,LSSVM)模型,实现了大坝水平变形的时间序列预测方法.基于Mercer理论,将多项式核函数和高斯核函数进行线性组合,构建混合核函数,作为LSSVM模型的核函数,并以特征因子与大坝变形间的相互联系为基础,采用动态自适应惯性权重的PSO算法,对混合核函数的LSSVM模型进行参数寻优,以确保建立最佳LSSVM预测模型.将模型应用于丰满大坝,并与传统多项式核函数和传统高斯核函数的LSSVM模型进行对比仿真实验,对所提方法的有效性和准确性进行验证评估.结果表明,该模型在预测精度上有了明显提高,预测性能尤佳.可见改进自适应粒子群的混合核函数LSSVM模型对大坝变形的时间序列预测有良好的实用价值.     

基于凸约束下泰勒估计的主瓣干扰抑制算法

针对真实信号协方差矩阵估计难以直接获取及低快拍条件下传统采样协方差矩阵存在较大误差的问题,提出了基于凸约束下泰勒估计的抗主瓣干扰波束形成算法。该算法首先利用凸约束下的泰勒估计法在低快拍数条件下对信号协方差矩阵进行估计。其次利用多信号分类算法进行波达方向估计,筛选主瓣干扰对应特征矢量。而后利用特征投影矩阵法对主瓣干扰进行抑制。最后,通过添加线性约束获得权值矢量进行波束形成。仿真结果显示,在低快拍数条件下,所提算法对信号协方差矩阵具有更高的估计精度,波束形成性能稳健且具有更高的输出信干噪比。     

基于信息优势的指控系统指标体系构建及建模

随着网络中心战的提出,信息优势在指挥控制系统中的作用日益突出,针对传统指挥控制系统评估指标体系未能很好吸纳信息优势,而导致评估结果与实际结果有偏差的问题,创新性地将信息优势作为一个重要因素纳入评估指标体系,构建了基于信息优势的指挥控制系统评估指标体系,并结合各指标特点建立了典型的、可量化的数学模型。最后,提出了指标体系构建过程中需要注意和把握的要点,并通过实例分析验证了指标体系的可行性及模型的正确性。     

短波信道下直扩信号检测与载频盲估计

针对通信侦察识别中广泛遇到的短波信道下直接序列扩频信号(DSSS)的检测与参数估计问题,提出了一种利用广度优先搜索邻居(BFSN)聚类分析提取未知信号高阶循环累积量循环频率的算法。该算法通过聚类比较截获信号的各类峰值,剔除干扰和奇异类信号,克服了短波信道中常见的多径效应引起的码间串扰(ISI)影响,从而实现了在低信噪比条件下,对直扩信号进行检测和载频的盲估计。对算法进行了理论分析,并通过计算机仿真验证了该算法在低信噪比下具有较好的稳健性。     

基于工艺的自适应数据质量多模型择优预测

针对钢铁企业实际生产过程中,采用单一预测模型进行预测难以把握大规模启停设备用电规律,预测精度较低等问题,根据生产-检修阶段的实际工艺情况,将生产,检修问题采用随机近似贪婪搜索RAGS对复杂特征进行特征选择,建立了一个自适应数据质量的多模型择优预测框架进行建模;将其应用于宝钢电网。仿真结果表明,提出多模型择优预测框架可以准确预测钢铁企业电力日负荷,为实现电力系统能源调度提供决策依据。