碳酸盐型磷矿的反浮选工艺研究

针对某碳酸盐型磷矿,采用反浮选工艺流程进行试验研究,获得了低镁优质磷精矿。反浮选工艺具有药剂种类少,分选效率高,选矿技术指标先进,操作控制容易等优点,是选别碳酸盐型磷矿的先进浮选工艺。     

塔里木盆地西南地区早中寒武世岩相古地理

塔里木盆地西南地区早中寒武世岩相古地理存在一定争议,这制约了该区的勘探评价。综合利用露头、钻井和地震资料,开展沉积相、地震相分析,并结合区域构造背景,恢复了岩相古地理格局,提出了相应的沉积模式。中下寒武统发育陆棚相、开阔台地相、局限台地相、蒸发台地相和缓坡相。震旦纪末期麦盖提斜坡南部形成水下低隆起,向南在西南坳陷区残留部分被动陆缘,主要发育缓坡相、陆棚相;麦盖提斜坡区地势较高,地层较薄,主要发育局限台地和蒸发台地潮坪;巴楚隆起区水体局限,在干燥炎热气候下形成了厚度较大的蒸发台地膏盐湖沉积。巴楚隆起区和麦盖提斜坡区中下寒武统膏盐岩分布广泛,与下伏碳酸盐岩可构成良好的储盖组合,西南坳陷区残留部分被动陆缘,烃源岩发育条件较好,这表明塔西南具有较好的油气前景。     

多隐层BP神经网络在模式预报中的简化应用

EC细网格预报效果好,基本满足业务需要,在工作中被广泛应用。为进一步提升预报准确性,做好迪士尼园区的气象服务保障,选取2016年7月至2017年6月1年的2 m温度预报场,24 h预报时效的时间分辨率为3 h,72 h预报时效的时间分辨率为24 h,分别用回归分析法、S型和简化Line型BP神经网络法进行模式释用,与迪士尼气象站观测数据对比。结果表明:阈值为1℃时,对模式结果释用后,均方根误差减少了0.5℃到1.0℃,3—9 h和21—72 h预报时效的准确率由原来的50%和30%分别上升到70%和50%。采用S型多隐层BP神经网络误差最小,不同预报时效释用稳定性最高,同时该释用方法对t_(min)的预报特征把握更精准,释用效果明显优于对t_(max)的预报释用,但迭代计算耗费时间大幅增多,与预报效果的提升不成正比。简化Line型的BP神经网络通过8个半月的数据量和简单的网络模式,捕获了EC预报的特征,不但减小了计算量,大幅缩短了计算时间,而且预报结果也有显著提升,预报稳定性较好,具有广泛的业务应用空间。     

Post-Aswan dam sedimentation rate of lagoons of the Nile Delta,Egypt

This study uses radiometric analysis (²¹⁰Pb and ¹³⁷Cs) of short sediment cores with high-resolution sampling (1-cm interval) to trace sedimentation rates in the Nile Delta lagoons, particularly since completion of the Aswan High Dam in 1964. A declining trend in ²¹⁰Pb_ex as calculated by the CIC model is clearly identified in about 10 cm of the upper-core sediments from the lagoons of Manzala and Edku, accompanied by two spikes of ¹³⁷Cs in cores from the lagoons of Burullus and Edku. These findings illustrate average post-dam sedimentation rates ranging from 0.22 to 0.27 cm a⁻¹ in the lagoons, in contrast with those found previously based on low-resolution sampling. The lower sedimentation rates in the lagoons are a consequence of a dramatic reduction in riverine sediment load to the coastal area as a result of the damming. Although widespread erosion occurs along the open estuarine coast, the lagoon setting remains calmer than before due to coastal diking and freshwater regulation in the delta plain in the past decades. This provides the possibility of continuously preserved radiometric records in the less-bioturbated lagoon sediments. Dating individual layers using the CRS model has revealed increasing sedimentation rates in Manzala and Burullus since the 1980s, which can largely be explained as a consequence of the reduction in lagoon area due to intensifying reclamation. The post-dam sedimentation in the shrinking lagoons may have some adverse ecological consequences due to finer sediment’s affinity with pollutants. These findings would shed light on the environmental conservation and socioeconomic development in the Nile Delta region.     

南海北部海区温跃层分布特征及成因的初步分析

利用二十一层海温再分析资料，详细分析了我国南海北部海区温跃层的强度、深度及厚度的季节变化特征。结果表明：在南海陆架浅水区域内，存在着随季节变化明显的辐射型温跃层；3－5月是温跃层的成长期：6－8月是该海域温跃层的强盛期；而9－11月温跃层开始减弱，到了冬季（12月到次年2月）温跃层变得最终，趋于消亡。结合本海区温跃层的这种变化特征，分析了该海域净辐射通量的分布状况及随季节的变化特征，证明了净辐射通量是影响该海域温跃层季节性分布特征的最重要因素之一。     

定点潮汐形变观测与GPS大地测量

论述了我国定点潮汐形变观测与GPS在观测周期，物理量，频域，实测精度，监测优势者方面的异同，认为定点潮汐形变观测是测震学与大地测量（含GPS）的必要补充，介绍了我国定点潮汐形变观测技术的发展与台网观测技术，以及它在我国10多次中强震测报中的应用。     

山西辛安岩溶泉形成条件及随机模型

本文详细分析了山西辛安岩溶泉的水文地质条件,并通过对泉域岩溶水的补给、径流、排泄条件的分析,将辛安泉域概化为单输入单输出线性时不变集中参数随机系统的概念模型。利用24年的水文序列资料进行基流分割,获得泉多年月平均流量过程曲线,求得泉年平均流量和年平均补给量。在此基础上,应用滑动平均模型（MA）,自回归—滑动平均模型（AR-MA）及泉流量调和分析—残差自回归等三个模型,对泉域地下水进行模拟,获得了满意的结果,不仅揭示了泉水动态特征,而且对泉流量动态进行了预报,为晋东南区域经济规划提供了有益资料。     

贵州五指山铅锌矿田深部找矿远景分析

贵州五指山铅锌矿田发育北东向、北东东向和北西向断层,分别平行于区域性安顺-贵阳、纳雍-息烽及水城-紫云深大断裂,控制了寒武纪-泥盆纪时期的沉积相和石炭纪前的地层剥蚀程度,以及沉积盆地中成矿流体活动。矿田中的那雍枝矿床的矿体产状、矿石结构构造、围岩蚀变及元素地球化学特征,反映其具有海底喷流沉积矿床成因。矿田深部地球物理场显示,矿田及附近存在北东向、北西向和东西向深大断裂,在几组断裂交汇处及附近的深部存在岩浆岩体及大型岩基,具备深部流体上升的条件。推测那润断层北盘那雍枝矿床深部灯影组内,F9、F10与F1断层相交,灯影组和清虚洞组层位及对应的物化探异常变异区的深部应存在大型层状铅锌矿体,具有超大型铅锌矿床的找矿前景。     

Heavy-Rain expert-system realized by logical language prolog

This article, on the basis of the expert’s analyses and theories of the forecast of heavy-rain, summarizes a variety of the current instruments and methods of forecast and, according to the train of thought of the high-rank forecasters, distills their experiences in forecasting heavy-rain into an inference-tree of 106 junctures from which 101 rule-bases are derived. The logical calculation is automatically carried out with our introduced and developed PROLOG, one of the intelligent languagas by means of micro-computer. This process adopts the uncertain inferential method based on the theory of fuzzy sets, breaks through the limits of two-value logic and is characteristic of the thinking of human brain.