动量BP算法在路基沉降预测中的应用

提出一种采用动量BP算法来预测路基沉降的方法,结合具体的工程实例,构建了预测路基沉降的具体BP神经网络模型。预测结果表明,该模型有较高的预测精度,可作为预测路基沉降的一种新方法。     

地地震多参数BP神经网络预测煤层厚度

依据煤层反射波运动学和动力学特征，提取出了波峰波谷振幅A1、平均频率Fa、主频带能量Qf1、低频带宽能量Qf和峰值频率Fmain等5个地地震特征参数。选取8组学习样本，利用4层BP（Back Propagation)人工神经网络模型，采用动量法和自适应调整的改进算法，训练BP网络，用训练好的BP网络预测煤层厚度。经实例验证，地震多参数BP网络预测煤层厚度精度高，是一种有效的煤厚预测方法。     

应用遗传-神经网络方法预测软土路基沉降

理论方法预测软土地基沉降与实际存在较大的差距,使得预测结果很难达到设计要求,不利于指导施工。将现有的理论方法同现场观测信息相结合,对软土地基变形作出更为准确的预测,有利于指导和控制工程施工。采用遗传算法和BP最优化法相结合的算法来训练网络,用遗传算法来优化BP神经网络中权值;用龚帕斯曲线来分解沉降时序,通过沉降趋势线偏移量来训练网络。采用这种方法预测软土路基沉降取得了较好的应用效果。     

基于改进BP算法的Elman网络在软基沉降预测中的应用

E lm an网络具有动态特性好,逼近速度快,精度高等特点,本文结合软基沉降的基本特征,建立了E lm an网络软基沉降预测模型。考虑到经典BP算法的缺陷,采用改进BP算法对网络进行训练学习。实例分析表明,本文所建立的E lm an网络模型具有一定的可靠性和实用性。     

神经网络技术在路堤沉降预测中应用探讨

针对传统的BP网络模型的不足，应用了改进的BP神经网络模型，把它应用到软基高速公路的路堤沉降预测中，提出了两种构造神经网络训练样本的思路，并分别进行了计算和对比，指出了各自的优、缺点。结果表明改进的BP网络模型比较稳健、收敛快，而且根据时间与对应的沉降量形成的样本训练的网络预测出的工后沉降误差小、精度高。     

人工神经网络在天津市区地面沉降预测中的应用

在分析天津市区地面沉降特点的基础上，结合人工神经网络原理，选择1961-1980年的天津市区降水量、地下水开采量、前年沉降量、固结度作为训练样本的输入量，以这20年的地面沉降量作为输出量，用贝叶斯正则化算法训练BP网络，得到沉降的仿真模型。并把1981-1993年的资料用来进行预测检验，结果表明这是一种比较理想的地面沉降预测方法。最后在不同的降水量保证率下，预测了到2010年天津市区地面沉降的情况。     

基于因子分析的LM-BP网络的盾构施工地表沉降预测

盾构施工引起的地表沉降量的预测与评价具有很强的随机性和不确定性,用常规的数学模型难以准确表达。本文采用因子分析法对影响盾构施工引起地表沉降的9个因素进行了主因子的提取,得到4个主因子。以此数据集对BP神经网络进行训练,并用Levenberg-Marquardt(LM)算法对BP神经网络进行了优化,建立了基于因子分析的LM-BP神经网络预测模型。结合工程实例将预测结果与LM-BP神经网络模型的预测结果进行比较分析,表明本预测模型具有较高的准确性,符合实际工程的需要。     

基于BP神经网络的太原市地面沉降预测及分析

通过对太原市地面沉降资料、气象水文和地质条件以及地下水等大量资料的分析与整理,考虑到太原市各个沉降中心的沉降趋势不尽相同,对太原市吴家堡、西张、万柏林和下元4个沉降中心分别建立BP神经网络模型,并基于训练好的BP神经网络模型,在太原市地下水开采量的规划方案下,预测了在不同降水保证率下2009-2015年地面沉降的趋势,...     

基于DEBP的GPS高程拟合在济宁矿区沉降精细化监测的应用

为减少济宁矿区开采过程存在的安全隐患,拟对济宁矿区沉降进行精细化变形监测,并建立预测模型。鉴于传统的BP神经网络算法存在的局部最优点、收敛困难、反演精度难以确定的问题,模型采用DEBP算法,结合DE算法和BP神经网络算法的优势,在收敛速度、精度以及稳定性方面都有提高。对比发现:基于DEBP的GPS高程拟合的预测模型,在济宁矿区沉降监测应用中能够有效提高监测精度,高达厘米级;结合灰度模型进行矿区沉降预测,通过监测结果可知,模型预测值与实测值一致性较好,表明该模型可用于矿区沉降预测,对于保护矿区具有重要的使用价值。     

RPROP算法在测井岩性识别中的应用

为了更好地解决测井岩性识别问题,引入一种快速实用的BP算法--Resilient Backpropagation (RPROP)算法。在说明RPROP算法的基础上,结合某地的实际测井资料,建立基于RPROP算法的BP网络岩性识别模型,进行岩性识别的应用研究。结果表明,应用RPROP算法进行测井资料岩性识别,识别的准确率较高,与基本BP算法及其一些改进算法相比,训练速度快,具有很好的应用前景。     

BP神经网络识别塔北低阻油气层

简要介绍了塔北低阻油气层岩性剖面、低阻油气层地球物理测井曲线特征,分析了塔北地区低阻油气储层成因,重点论述BP人工神经网络识别油气层、油水同层、水层和干层的方法原理。识别实例表明,BP人工神经网络识别低阻油(气)、水层的结果与实际相符,明显地提高了测井的解释精度。     

应用两总体判别分析法对试油层位进行优选

把油水同层作为有利试油层位含油气性的下限，通过提取试油层位的4种地质参数，利用两总体判别分析法，建立有利试油层位和不利试油层位的判别函数，根据判别函数并结合测井解释资料对试油层位进行优选．     

Microfacies of detrital event layers deposited in Quaternary varved lake sediments of the Piànico-Sèllere Basin (northern Italy)

A Quaternary interglacial lake sediment record from the Piànico-Sèllere Basin (northern Italy) consists of biochemical calcite varves with intercalated detrital layers. At the end of the Piànico Interglacial, continuous varve formation was replaced by predominantly detrital sedimentation. However, 427 varve-years before this shift, an abrupt increase in the frequency and thickness of detrital layers occurred. Microfacies analyses reveal a total of 152 detrital layers, ranging from 0·2 to 20·15 mm in thickness, deposited during the last 896 years of the Piànico Interglacial. Three microfacies types are distinguished: (i) graded layers, (ii) non-graded silt layers, and (iii) matrix-supported layers. The position of detrital layers within an individual varve provides additional information on the season in which they have been deposited. Microfacies analyses in combination with varve counting further enabled precise varve-to-varve correlation of the detrital layers for two sediment sections cropping out ca 130 m apart. The detailed intra-basin correlation allows the source regions of detrital layers to be inferred. Moreover, micro-erosion at sub-millimetre scale has been established. Of the described facies types, only the accumulation of summer and spring graded and non-graded silt layers abruptly increased before the end of interglacial varve formation whereas non-graded winter silt and matrix-supported layers are randomly distributed over the entire study period. Heavy rainfalls are assumed to have triggered spring and summer graded layers, so that the occurrence of these layers is thought to be a proxy for extreme precipitation events in the past.     

青藏高原沱沱河盆地渐新世湖相叠层石:韵律纹层记录的古气候条件

沱沱河盆地通天河剖面渐新世雅西措群发现一套产于湖泛过程中的柱状叠层石。通过对叠层石纹层形态和显微组构的研究,发现这些叠层石具有典型的富藻纹层和富屑纹层交替的显微结构特征,发育三级生长节律,纹层对之间渐变接触,纹层带之间间断明显。采用激光微区取样的方法,针对4块样品的8个纹层对,系统采集了16对碳氧同位素数据,结果是富藻纹层同位素值整体上均高于富屑纹层,正相关性明显。这表明富藻纹层生长时期气候炎热湿润,富屑纹层生长时期相对寒冷干燥。认为湖相叠层石生长受周期性气候变化的控制,其韵律纹层记录了丰富的古气候信息。     

Modally-Graded Rhythmic Layering in the Skaergaard Intrusion

The origin of igneous layering in the Skaergaard and other intrusionsis the subject of much debate. Currently held theories for itsorigin include gravitational sorting, rhythmic nucleation, andchemical diffusion models. This paper presents data on small-scale,modally-graded, rhythmic layers, one of several distinct typesof layering found in the Skaergaard. These layers occur onlyin the Layered Series of the intrusion, in which they are awidespread, almost ubiquitous feature. They are characterizedby strong variations in modal composition, with magnetite, ilmenite,and olivine concentrated in the lower part of the layers, pyroxenein the middle of the layers, and plagioclase in the upper partof the layers. In addition, the layers are size graded, withthe largest grain sizes of the different minerals generallyoccurring where the mineral is most abundant. The chemical compositionof both plagioclase and Ca-rich pyroxene do not vary in anyregular manner within the layers studied. The forsterite compositionof olivine is consistently higher in the lower part of the layers.This trend, however, correlates very strongly with the modalvariations of olivine and can be explained by minor amountsof subsolidus reequilibration between cumulus and adcumulusolivine, suggesting that the original composition of cumulusolivine within individual layers was constant. These results are consistent with layers formed by some gravitationalsorting process and, in several cases, are at odds with othertheories that have been proposed for the origin of igneous layers.Based on these data and the widespread evidence for currentinteraction (crossbeds, trough-shaped layers, and internallygraded layers), we propose that these layers result from periodicdisruption of a static zone of crystal growth on the floor ofthe magma chamber by convection or density currents. This mechanism,however, is almost certainly not responsible for the formationof all of the types of layering found in the intrusion.     

塔里木盆地东河1油藏低渗层层序控制特征及其对注气开发的影响

通过测井和岩心分析,将塔里木东河1油藏东河砂岩段从下到上划分为前滨-上临滨-下临滨相带,并进一步划分出11个准层序。经岩相学分析,将砂岩段低渗层分为钙质胶结低渗层、杂基充填低渗层和混合低渗层。钙质胶结低渗层主要分布在准层序内部,以方解石胶结砂岩为主,受成岩作用控制;杂基充填低渗层主要分布在准层序边界处,多为海侵背景下的滞后沉积产物;混合低渗层在准层序内部及边界均有分布,但数量很少,无明显规律性,砂粒间主要由杂基和方解石胶结物混合充填,充填物含量低。结合测压资料认为,发育在准层序边界处的杂基低渗层保留了更多的初始沉积特征,易形成横向连续性好的低渗层,是注气开发的重点关注对象。     

油、水、干层的地球化学识别

本文介绍一种判断油、水、干层的地球化学方法，根据储层样品中有机抽提的的含量和组成特征，结合油层物性，地层测试和试油资料编制织识别油、水，干层的图版，提出判别油、水、干层的地质／地球化学指标，确定油层的地球化学参数的下限，并运用此方法对松辽盆地新站油田主要油层分布段的油、水、干层进行判别。根据试油资料验证表明，地球化学判断结果比传统的地质录井及地球物理测井资料判断结果更符合实际情况。     

Electron microscopic investigations of sand material in the core drillings in the Great Hungarian plain

The study may inform about the electron microscopic investigations of the sand layers in seven core drillings in the Danube-Tisza Interfluve region. The primary aim of the investigation was to obtain exact data about the blown sand layers under the surface. In five out of six drillings on the alluvial fan of the Danube blown sand layers were explored (the number of layers varied between 2 and 7). There weree apparently phases in the formation of the alluvial fan when on certain parts flooding was replaced by eolian processes. The blown sand layers under the surface, however, are only several metres thick at the most. The seventh drilling (at Jánoshalma) proves that this region constitutes the continuation of Transdanubia and does not belong to the alluvial fan of the Danube. At Jánoshalma the driller bored through 26 loess, sandy loess, and loessy sand layers, and more than 10 blown sand layers under the surface in the Quarternary sequence.     

A preliminary evaluation model for reservoir hydrocarbon generating potential established based on dissolved hydrocarbons in oilfield water

OILFIELDWATER,WHICHEXISTSCONCOMITANTLYWITH OILORGASINSUBSURFACERESERVOIRS,ISAKINDOFIMPOR TANTCOMPONENTOFRESERVOIRFLUIDANDALSOISADRIVING FORCEANDCARRIERFORHYDROCARBONMIGRATION,ACCUMULA TIONANDHYDROCARBONRESERVOIRFORMATION.OILFIELDWA TER,HYDROCARBONANDRESER…     

藏北羌塘盆地褶皱形变研究

羌塘盆地由于受多期构造活动的影响形成多个构造层,不同构造层的变形特征存在明显差异。其中三叠纪构造层多形成一些紧闭倒转的小型背、向斜褶皱,侏罗纪构造层内多发育大型宽缓的背、向斜及复背斜、复向斜,而白垩纪—新近纪构造层多位于向斜核部和背斜翼部,形成宽缓的褶曲。褶皱变形以近东西向为主,从老到新不同构造层形成的褶皱由紧闭到宽缓,存在一定的继承性和递进叠加的特点。褶皱轴迹的空间展布及变形特点表明不同时期板块的拼合挤压是导致盆地变形的主要驱动力,基底断裂及基底凸凹变化对褶皱的展布和变形有一定程度的控制和影响。