琼东南盆地深水区构造热演化特征及其影响因素分析

To reveal the tectonic thermal evolution and influence factors on the present heat flow distribution, based on 154 heat flow data, the present heat flow distribution features of the main tectonic units are first analyzed in detail, then the tectonic thermal evolution histories of 20 profiles are reestablished crossing the main deep-water sags with a structural, thermal and sedimentary coupled numerical model. On the basis of the present geothermal features, the Qiongdongnan Basin could be divided into three regions: the northern shelf and upper slope region with a heat flow of 50–70 m W/m2, most of the central depression zone of 70–85 m W/m2, and a NE trending high heat flow zone of 85–105 m W/m2 lying in the eastern basin. Numerical modeling shows that during the syn-rift phase, the heat flow increases generally with time, and is higher in basement high area than in its adjacent sags. At the end of the syn-rift phase, the heat flow in the deepwater sags was in a range of 60–85 m W/m2, while in the basement high area, it was in a range of 75–100 m W/m2. During the post-rift phase, the heat flow decreased gradually, and tended to be more uniform in the basement highs and sags. However, an extensive magmatism, which equivalently happened at around 5 Ma, has greatly increased the heat flow values, and the relict heat still contributes about 10–25 m W/m2 to the present surface heat flow in the central depression zone and the southern uplift zone. Further analyses suggested that the present high heat flow in the deep-water Qiongdongnan Basin is a combined result of the thermal anomaly in the upper mantle, highly thinning of the lithosphere, and the recent extensive magmatism. Other secondary factors might have affected the heat flow distribution features in some local regions. These factors include basement and seafloor topography, sediment heat generation, thermal blanketing, local magmatic injecting and hydrothermal activities related to faulting and overpressure.     

短期风化对混合溢油组成的影响

对两种原油进行混合配比实验,将混合后油样置于人工气候箱进行风化模拟实验,采用气相色谱质谱联用仪(GC-MS)检测风化样品中的生物标志化合物,短期风化作用对混合溢油的油指纹、生物标志化合物诊断指标等的影响。结果表明,混合溢油的正构烷烃总质量变化与单一原油油品的变化规律相近,即前期风化较快,质量减损较多,而后期风化趋缓,不同混合比例的溢油表现差异不明显。常用于短期风化的诊断比值对各混合油样的风化具有指示意义,但难以定性鉴别油品是否发生混合。重复性限法检验只能判定短期风化过程前后的油样为同一油源,而难以反映油样是否为混合油源的特征,各诊断比值的RSD%值较大可能是混合溢油的一个表现。     

走航式海洋多参数剖面测量系统(MVP)温度和电导率滞后效应的一种资料处理方法

走航式海洋多参数剖面测量系统(MVP)是一种集成程度和自动化程度都较高的海洋调查设备,能对多要素进行同时观测,获得高水平空间分辨率的数据资料。MVP由于温度和电导率传感器响应时间的不匹配,下放速度过快(峰值速度4m/s)而造成非常严重的盐度尖峰现象。本研究通过结合F法、GM法和Grose法提出的盐度尖峰订正方案,提出了一种新的方法Match conductivity and temperature response times法,对压力、温度和电导率传感器三者进行响应时间的匹配来减弱盐度尖峰。与SBE-9型CTD资料进行对比发现订正后的资料误差比订正前减小80%,与CTD盐度曲线互相关程度为0.917。对比35N断面修正前后的盐度资料发现订正后温盐跃层处出现的低盐区域消失,与CTD断面资料对比结果显示MVP资料比CTD资料在细结构上更具有优势。     

一种统计技术结合动力释用的沿海海雾预报方法

提出了一种利用历史相似个例查询技术,结合WRF模式输出释用的Gultepe沿海大雾业务预报方法。该方法可以较好的计算近地面大气中的液态水含量(LWC)和云滴粒子数密度(Nd)参数。利用该方法对2011—2013年的3—6月的历史沿海海雾个例进行试报,结果表明,恰当的结合统计方法和动力模式可以一定程度上克服目前数值模式不成熟的缺陷,从而提高WRF模式在大雾业务预报方面的释用水平。     

同步发电机PSS与励磁系统的仿真研究

电力系统稳定器（PSS）是目前抑制电力系统低频振荡最有效的方法之一.在分析了同步发电机励磁控制系统及其模型的基础上,研究了以PSS为辅助控制的同步发电机励磁控制方式,基于Matlab构建了改进的PSS励磁控制系统的仿真模型.在三相短路以及断线故障情况下对单机无穷大系统进行仿真测试,结果表明：加装PSS的励磁控制系统的各方面性能优越于无PSS的励磁控制系统,为励磁系统的设计提供了依据.     

通过软硬变化检测识别冬小麦

提出一种软硬变化检测的作物识别方法 SHLUCD(Soft and Hard Land Use/Cover Change Detection Method)。该方法利用多期遥感影像能够有效表达作物的生长物候特征,以达到在离散变化区(即纯净像元区,包括完全转换成作物的突变区域和非作物区域)和连续变化区(即渐变区,混合像元区,是部分转化为作物的区域)准确进行作物的识别。在北京市选择一个研究区,以冬小麦为研究对象,选用2011年10月6日(播种期)和2012年4月16日(拔节期)两期环境减灾1号卫星影像,分别采用硬变化检测方法 HLUCD(Hard Land Use/Cover Change Detection Method)、软变化检测方法 SLUCD(Soft Land Use/Cover Change Detection Method)和SHLUCD进行冬小麦的识别。实验结果表明:在不同尺度窗口下,SHLUCD较传统方法表现出较明显的优势,具有更低的均方根误差RMSE(SHLUCD为[0.14,0.07],HLUCD为[0.15,0.07],SLUCD为[0.16,0.08])和偏差bias(SHLUCD为-0.0008,HLUCD为-0.007,SLUCD为0.014)和更高的决定系数R2(SHLUCD为[0.68,0.86],HLUCD为[0.62,0.86],SLUCD为[0.60,0.86])。针对冬小麦突变区域、冬小麦渐变区域和非冬小麦区域分别进行评价,表明SHLUCD识别精度接近各区最佳的识别方法,进一步验证了SHLUCD的灵活性和适用性。SHLUCD方法在离散变化区能够通过土地覆盖类型状态变化来有效地识别出冬小麦,在连续变化区可识别出土地覆盖的状态变化程度定量表达冬小麦的丰度,是其他作物多时相遥感变化检测的前期实验基础。     

Progress in Studies of the Precipitation Diurnal Variation over Contiguous China

This paper summarizes the recent progress in studies of the diurnal variation of precipitation over con- tiguous China. The main results are as follows. （1） The rainfall diurnal variation over contiguous China presents distinct regional features. In summer, precipitation peaks in the late afternoon over the south- ern inland China and northeastern China, while it peaks around midnight over southwestern China. In the upper and middle reaches of Yangtze River valley, precipitation occurs mostly in the early morning. Summer precipitation over the central eastern China （most regions of the Tibetan Plateau） has two diurnal peaks, i.e., one in the early morning （midnight） and the other in the late afternoon. （2） The rainfall diurnal variation experiences obvious seasonal and sub-seasonal evolutions. In cold seasons, the regional contrast of rainfall diurnal peaks decreases, with an early morning maximum over most of the southern China. Over the central eastern China, diurnal monsoon rainfall shows sub-seasonal variations with the movement of summer monsoon systems. The rainfall peak mainly occurs in the early morning （late afternoon） during the active （break） monsoon period. （3） Cloud properties and occurrence time of rainfall diurnal peaks are different for long- and short-duration rainfall events. Long-duration rainfall events are dominated by strat- iform precipitation, with the maximum surface rain rate and the highest profile occurring in the late night to early morning, while short-duration rainfall events are more related to convective precipitation, with the maximum surface rain rate and the highest profile occurring between the late afternoon and early night. （4） The rainfall diurnal variation is influenced by multi-scale mountain-valley and land-sea breezes as well as large-scale atmospheric circulation, and involves complicated formation and evolution of cloud and rainfall systems. The diurnal cycle of winds in the lower troposphere also contributes to the regional differences     

基于伪语义相似度模型的中文地址匹配方法

地址匹配中,由于传统相似度模型受字符重叠数影响大,在处理简写、缩写地址要素单元时,错误匹配问题突出;深度学习方法需要大量样本支撑,但庞大的数据量和多样的形式,导致生成样本的成本过高。为解决上述问题,本文首先应用基于条件随机场和双向长短时记忆神经网络的模型,对地址进行分词;然后通过建立一种伪语义相似度,对地址要素进行分级匹配。通过对公安业务中地址数据进行测试,在对缩写、简写等不规范地址描述方面,本文模型能较理想地完成任务,各参考指标均高于0.9。     

基于空间变异修正普通克里金的矿区复杂地形高精度建模

普通克里金法是构建矿区三维地形模型并揭示地表沉陷形变场动态变化规律的有效方法。但普通克里金法存在平滑效应的问题,导致估计值的空间变异程度小于实际,无法真实反映矿区复杂地表形态的空间变化特征。本文结合矿区实际地形采样数据,提出了一种空间变异修正的普通克里金法,并将其应用于矿区复杂地形的高精度建模,并与已有方法在建模精度和空间变异性复现方面进行了对比分析。研究结果表明,空间变异修正的普通克里金法能够很好地处理平滑效应对建模的影响,具有更高的建模精度和空间变异性复现能力,在矿区复杂地形高精度建模应用中具有较强的适用性,可以作为一种可靠的建模工具用于矿区复杂地形沉陷形变场动态变化规律分析。     

一种基于改进CoHOG的视觉SLAM算法

移动机器人在SLAM的闭环检测环节计算量大、运行时间长、匹配误差大,从而导致闭环检测精度较低。针对该问题,本文在CoHOG闭环检测算法的基础上进行改进,将算法中的HOG描述符改进为GDF-HOG描述符,以增强图像特征表现,提高图像特征提取效率;在匹配环节前添加GDF-HOG全局粗匹配,以减少视觉模板的数量,提高算法的计算效率;在匹配环节后添加感兴趣区域（ROI）位置匹配进行检验,以减少闭环检测的假阳性,提高准确率。将本文闭环检测算法与RatSLAM相结合,在公开数据集与真实环境中进行测试,测试结果表明,本文算法在闭环检测环节的准确率较高,且对环境的适应能力较强。