Estimation of copepod production in a flooded savanna of Venezuela

This is the first contribution to the copepod production in Venezuelan tropical savannas. Total abundance, biomass, production and mean P/B ratio of nauplii, copepodids and adults were determined in interdaily samples from a flooded, embanked savanna during February and March (end of dry season), and May (beginning of rainy season). Highest values of biomass and total production were recorded during dry season (61.5 mg m⁻³, 153.8 mg m⁻³ day⁻¹, respectively), compared to the rainy season (5.6 mg m⁻³, 45.9 mg m⁻³ day⁻¹). The last values are related to a low population density found during rainy season. Highest production were observed in copepodids at the end of the dry season. Significant differences of production between nauplii and copepodid stages, as well as between nauplii and adults, were also found.     

舱容计算中的插值误差分析

本文介绍了在舱容计算过程中用到的插值处理的误差分析;通过计算实例分析误差产生的原因,并提出了减小插值误差的方法。     

应用BP网络测算造船作业任务包定额工时

分析分段各组立阶段所对应中间产品之间的关系，得出它们之间有特定的偏序关系。产品导向型作业分解是依据船厂的资源条件分解一个个任务包，对应中间产品的一个个任务包与作业工时又有非线性的函数关系。以外高桥造船公司为背景提出了中心制造船模式下工时定额和信息反馈的设想，并用网络方法估算作业工时。分析了BP网络的特性，在建立工时测算模型的基础上提出了BP网络的工时测算模型，经过编程试算，说明本方法可行。     

基于时空序列的铁路客运量数据挖掘预测

在分析铁路客运量数据的时空复杂性特征的基础上,以铁路假日运输管理系统中春运期间的客运量数据为依据,采用BP神经网络的数据挖掘方法进行铁路客运量数据挖掘预测研究。通过BP神经网络的建模方法把客运量的空间属性、数据属性和时间属性有机地结合起来,将数据的建模含于网络的数值当中。网络在学习过程中系统误差始终保持持续稳定的下降趋势,没有产生局部振荡和陷入极小现象,整个学习过程中系统稳定性较好。各样本之间的期望输出和实际输出之间吻合较好,从而证明所采取的数据处理方法的有效性和网络学习参数的合理性。根据BP神经网络得到的预测模型在仿真试验中的期望输出和实际输出之间吻合较好,预测的客运量和实际客运量数值非常接近。     

基于分形理论的铁路货运量分析

论证了分形理论用于铁路货运量分析的可行性;计算出铁路年货运量时间序列集的分维,并结合有关的数学、物理理论分析了该分维的意义;用定性分析与基于分形理论的定量计算预测了2001～2010年中国铁路货运量的发展水平。     

Siliceous plankton dominate primary and new productivity during the onset of El Niño conditions in the Santa Barbara Basin, California

The potential for carbon export and the role of siliceous plankton in the cycling of C and N was assessed in natural plankton assemblages in the Santa Barbara Basin, California, by examining uptake rates of inorganic carbon, nitrate and silicic acid. In April–August 1997, the concentrations of chlorophyll a, particulate organic carbon, particulate organic nitrogen and biogenic silica were measured twice monthly, and results revealed the occurrence of at least three blooms, the largest in June. Particulate elemental ratios of C, N and Si were similar to ratios of nutrient-replete diatoms, suggesting that they dominated this bloom. Mean integrated rates of carbon, nitrate and silicon uptake during the 4-month study period are similar to other productive coastal and upwelling regions (103, 8.3 and 13 mmol m⁻² day⁻¹, respectively). New production rates were twice as high as previously reported in this region and indicate that high rates of new production along eastern boundary currents are not confined to the major coastal upwelling regions. C/NO₃⁻, Si/NO₃⁻ and Si/C uptake ratios varied widely, and mean integrated ratios were 14±5.4, 1.6±1.0 and 0.12±0.07 (S.D.), respectively. That mean C/NO₃⁻ uptake ratio corresponds to an f-ratio of about 0.5 indicating a large potential for particulate export. Based on the average Si/NO₃⁻ and Si/C uptake ratios, diatoms could perform all of the primary production and nitrate uptake that occurred during the study; these rates also suggest that export is controlled by diatoms in this system. The mean Si/C biomass ratio was lower than the mean Si/C uptake ratio, consistent with the preferential export of Si relative to C observed in sediment traps in the basin. The study took place during a period of surface-water warming, with nitrate and silicic acid concentrations decreasing throughout the onset of the 1997–1998 El Niño conditions. Although diatoms contributed less to particulate biomass during the low nutrient conditions, high f-ratios (0.33–0.66) were maintained.     

基于径向基神经网络的铁路货运量预测

货运量预测是铁路运输部门一项重要工作,因此,关于铁路货运量预测理论和方法的研究一直是一个热点。但是,铁路货运量受多种因素影响,且各因素的作用机制通常不能或无法用精确的数学语言来准确描述。本文采用径向基函数(RBF)神经网络对货运量进行分析及预测。通过对1989~2002年全国铁路货运量的历史数据分析处理后,得到铁路货运量增长量的时间序列,将时间序列视为一个从输入到输出的非线性映射,引入RBF神经网络来进行非线性映射的逼近。对网络进行学习与训练仿真实验后,用2003~2004年的增长量进行模型检验,并与BP神经网络、灰色预测模型预测结果进行对比,结果表明,应用RBF神经网络对铁路货运量进行短期预测预测精度更高、效果更好。     

车辆段生产支持系统

车辆段生产支持系统的研制以安全、质量、生产进度和人员教培管理为核心,建立了生产管理信息平台和通用分析平台,在对现有的HMIS、材料系统、调车系统、YMIS等系统进行有效的数据资源整合的基础上,重点着眼于各类数据的深层次分析利用,为车辆段的生产组织、成本控制、生产过程控制、安全管理和质量管理等提供决策支持。     

70%低地板轻轨列车车内声学计算与分析

通过对70%低地板车辆进行动力学建模,仿真得出车辆正常运行时二系结构与车体连接处的力和两处下铰力,以此作为激励对车体结构有限元模型进行频率响应计算。用计算得到的车体位移激励车体声学有限元模型,得到车内声学模态、声场分布和ISO标准场点响应。结果显示,车体在几个特定频率下的声压级超出了车内噪声指标。通过对几个特定频率下的各板件声场贡献量计算,得到车顶正贡献量较大,可提供给厂方进行结构优化。同时,根据结构模态、声学模态计算结果对车体下吊设备频率提出了建议。     

浅谈高速公路绿化施工及管理

随着高速公路在我国的跨越式飞速发展,高速公路绿化的模式和建设模式都发生了深刻的变化。讲述了高速公路绿化的要点,并提出了相应的公路绿化施工及管理的措施。