物理精炼菜籽油的前处理——菜籽油的物理精炼研究(Ⅱ)

本课题采用正交试验方法,对脱胶添加剂和国产脱色白土的种类、用量及其他操作条件进行筛选试验研究,将试验结果对贵州产二级毛菜籽油进行前处理,经水蒸汽精炼后可以制得合符标准的外销色拉油。     

对体育教育专业篮球普修课教学现状的思考

通过对体育教育专业篮球普修课教学目标、教学内容、教学模式以及对学生的成绩评价进行思考和令尊，结合《全国普通高校学校体育教育本科专业课程方案》对培养目标的规定，提出了对体育教育专业篮球普修课的培养目标进行重新定位，优化教学内容、优化教学模式、优化教学手段、优化评价体系，培养出合格的21世纪体育教师。     

健康第一原则与学校体育教育改革的思考

学校体育立“素质教育”,树“健康第一”的指导思想是新时期学校体育发展的必然走向。但学校体育的健康教育还停留于形式上,还没有完全从原有的模式跳出来。学校体育改革就是要我们去做一种与以往不同但却可以达到以增进健康为主要目的的体育活动。从大众体育健康的角度,忽视或不要去刻意追求过去那种生理“波峰”的传统模式。     

G-X5CrNi13.4不锈钢水斗现场检测方法

针对国内水头最高的水电站——西藏羊卓雍湖抽水蓄能电站4^#机水斗断裂事故，开展了在役机组水斗检测方法的研究，对照各种检测方法的差异，找出了最为有效的检测手段。     

锻炼对老年人肺通气功能影响的研究

本文采用电子肺活量计对218名老人进行了肺通气功能的测试。测试结果表明:体育锻炼对老年人肺通气功能有良好影响。本文为老年人肺通气功能的研究、评价及老年人体育运动的研究提供了实验性依据。     

Evolutionary and ecological implications of cardenolide sequestration in the monarch butterfly

Summary Monarch butterflies sequester cardenolides from their larval host plants in the milkweed genusAsclepias for use in defense against predation. Of 108Asclepias species in North America, monarchs are known to feed as larvae on 27. Research on 11 of these has shown that monarchs sequester cardenolides most effectively, to an asymptote of approximately 350 g/0.1 g dry butterfly, from plants with intermediate cardenolide contents rather than from those with very high or very low cardenolide contents. SinceAsclepias host plant species are distributed widely in space and time across the continent, monarchs exploit them by migration between breeding and overwintering areas. After overwintering in central Mexico, spring migrants east of the Rocky Mountains exploit three predominantAsclepias species in the southern USA that have moderately high cardenolide contents. Monarchs sequester cardenolides very effectively from these species. First generation butterflies are thus well protected against predators and continue the migration north. Across the northern USA and southern Canada most summer breeding occurs on a fourthAsclepias species and in autumn most of these monarchs migrate back to Mexican overwintering sites. The ecological implications of this cycle of cardenolide sequestration for the evolution of monarch migration are discussed.     

Chemical ecology of oviposition in phytophagous insects

Summary Selection of a suitable site for oviposition by phytophagous insects is critical for successful development of the offspring. The behavioral events leading to oviposition are mediated to a large extent by chemical cues associated with potential host plants. Orientation and landing are primarily guided by volatile constituents of a plant, whereas assessment of a leaf surface depends on contact stimuli. Chemical mechanisms that ensure adequate spacing of progeny on limited resources include the production of oviposition-deterring pheromones as well as recognition of plant constituents released as a result of previous damage. Perception of chemical cues that affects oviposition involves receptors on antennae, tarsi, mouthparts or the ovipositor. Complex behavior such as tarsal drumming or stem runs may serve to provide increased receptor contact with chemical stimuli. Abiotic and biotic environmental factors often influence the production or release of behavior-modifying chemicals by a plant, and therefore affect oviposition preferences. Plant chemistry may be involved in associative learning, but may also lead to mistakes. Thus a clear correlation between oviposition preference and offspring success does not always exist.     

关于最大功原理的应用条件

本文从热力学第二定律的基本原理出发,概要地总结了论证“最大功原理”的方法,进而讨论了热力学中的一些基本问题。其基本思想是:分析热力学问题切不可忘了环境。由此得出了某些有意义的结论,澄清了在学习热力学过程中人们常产生的一些疑惑。     

基于降维观测器与多胞体方法的执行器故障检测

针对具有执行器故障且状态方程和输出方程同时受到外部扰动影响的离散系统,研究了降维观测器设计方法和基于多胞体理论的区间估计方法,并在此基础上提出了基于区间估计的执行器故障检测方法。首先,通过适当的状态等价变换,将原系统分解成2个降维子系统,使得一个子系统可以直接解耦干扰。然后,在无执行器故障发生的前提下,基于对干扰解耦的子系统,提出了一种降维观测器设计方法,实现了原系统状态的渐近收敛估计。其次,针对具有干扰的子系统,设计了Luenberger型观测器,并结合多胞体理论,提出了一种对可测输出的区间估计方法。之后,基于可测输出的区间估计,提出了一种残差的构造方法,以此提出了一种执行器故障检测新方法。最后,给出了仿真设计和分析,验证了方法的有效性。     

基于层次注意力机制的垃圾用户检测模型

目前基于网络的垃圾用户检测方法只考虑了简单社会关系,缺乏对更多复杂社会语义关系的利用,难以达到最优性能.针对这一挑战,提出一种基于层次注意力机制的垃圾用户检测模型(HAM-SD).模型首先使用异质信息网络对社交媒体进行建模,挖掘丰富的语义与结构信息,接着利用节点级注意力层聚合元路径邻居增强节点表示,同时利用自适应层级聚合模块选择不同层级特征提升表征能力,然后通过语义级注意力层融合不同元路径下的节点表示,最后带入分类检测模块实现垃圾用户检测.在公开数据集上的实验结果表明该模型能够有效检测垃圾用户,并在不平衡数据分布时保持较强的稳定性.