基于机器学习和测井数据的碳酸盐岩孔隙度与渗透率预测

准确预测碳酸盐岩储层孔隙度和渗透率对于碳酸盐岩油气藏储层评价具有重要意义。碳酸盐岩储层裂缝与溶孔广泛发育,基于经验公式从测井曲线预测储层孔隙度和渗透率具有较大误差。以中东某碳酸盐岩油藏为研究对象,选取914块取心井岩心,测定孔隙度与渗透率,利用随机森林（RF）、K-近邻（KNN）、支持向量机（SVM）和长短期记忆网络（LSTM）4种不同机器学习方法,通过测井数据进行孔隙度与渗透率预测,优化机器学习参数,筛选出适用于碳酸盐岩油藏的测井孔隙度与渗透率预测方法。研究结果表明：4种机器学习方法预测储层孔隙度结果差异不大,通过调整输入参数种类,可进一步提高孔隙度与渗透率预测效果,当以补偿中子（NPHI）、岩性密度（RHOB）和声波时差（DT）3种测井参数数据作为输入时,基于LSTM的储层孔隙度预测精度最高,孔隙度预测结果均方根误差（RMSE）为4.536 2;由于碳酸盐岩储层的强非均质性,基于机器学习的测井储层渗透率预测效果较差,相对而言,仅以NPHI作为机器学习输入参数时,基于RF的储层渗透率预测精度最高,渗透率预测结果的RMSE为45.882 3。     

Petrophysical evaluation of subsurface reservoir sandstones of Bengal Basin,Bangladesh

The present study has been carried out for core and log based petrophysical evaluation of subsurface reservoir sandstones considering at least one well from each of the Kailas Tila, Titas, Bakhrabad and Shahbazpur gas fields of Bangladesh. Core analysis results show that average core porosity, permeability and pore throat values are 20%, 209 mD and 44020 Å, respectively. These measured values support that the sandstone reservoirs are categorized as good quality reservoirs. Core porosity values usually exceed thin section porosity. Different cross-plots indicate that permeability is largely dependent on porosity. Permeability and porosity are also dependent on textural parameters that include size, shape, sorting and matrix of the reservoir sandstones. It reveals from both thin section study and different cross plots that diagenetic cements are the main controlling factors of the reservoirs. Quantitative determination of the volume of cements shows that cements have inverse correlation with porosity and permeability. Log based petrophysical parameters include shale volume, porosity, water and hydrocarbon saturation, permeability, moveability index and bulk volume of water. The average values of the mentioned parameters are 20%, 22%, 26%, 74%, 110 mD, 0.28 and 0.05, respectively. The most important parameters are porosity and permeability, which indicate that log porosity exceeds core porosity and log permeability significantly lags behind core permeability. The study also reveals that 23 gas zones covering total thickness of 385 m sandstones mostly posses good quality reservoirs except few moderate quality reservoirs.     

Logging Evaluation of the Ordovician Carbonate Reservoir Beds in the Lungudong Region,Tarim Basin

<正>In recent years,great progress has been made constantly in oil and gas exploration in the Lungudong region of the Tarim Basin.However,progress has been slow in the evaluation of its main oil-producing horizons-the Ordovician carbonate reservoir beds.Based on previous researches and on the various data such as drilling,geology and oil test,in combination with the interpretation of each single-well imaging and conventional logging data,and through analysis and comparison,the identification methods in imaging and conventional logging for four types of carbonate reservoir beds in this region are summarized in this paper.Calculation formulas for four reservoir bed parameters,i. e.shale content,porosity,permeability and oil saturation in this region are proposed;and reservoir beds in this region are divided into three levels(Ⅰ,ⅡandⅢ) by combining oil test data and logging data,The lower limits of the effective porosity of reservoir beds and the fracture porosity of effective reservoir beds are determined as 1.8%and 0.04%,respectively.The physical property parameters are calculated by conventional logging curves,and the most advantageous areas for reservoir development are predicted comprehensively.On the plane,the high-value zones of reservoir bed parameters are mainly concentrated in the N-S-trending strike-slip fault,the Sangtamu fault horst zone and near the LG38 well area;vertically,the reservoir bed parameters of the Yijianfang Formation are better than those of the Yingshan and Lianglitage formations.     

致密滩坝砂储集层孔隙分形特征、预测及应用

储层孔隙分形可以有效表征储层孔渗性能,综合反映储层孔隙结构特征及评价储层开发效果。为确定开发区块动用次序,对于东营凹陷西部区块沙四上纯下次亚段的致密滩坝砂储层,首先,利用薄片、物性及压汞等相关测试数据,计算致密滩坝砂孔喉分形维数（D）;其次,探讨分形维数与储层物性、孔隙结构参数相关性;然后,优选测井数据,建立孔喉分形维数的测井预测模型,并对东营凹陷西部区块滩坝砂分形维数平面分布进行了预测;最后,根据分形维数和油井产能相关性分析,建立了基于分形维数的储层评价标准,对研究区的致密滩坝砂储层进行了分类与评价。结果表明:当2相似文献   

多元回归分析方法在火山碎屑岩储层孔隙度计算中应用

火山碎屑岩储层具有岩性复杂、测井数据种类繁多以及区域大、层位多等特点,导致火山碎屑岩储层测井参数难于确定。以海拉尔盆地乌尔逊--贝尔凹陷的火山碎屑岩储层为研究主体,在准确识别岩性的基础上,对测井资料进行预处理,根据岩心分析资料计算其理论骨架参数值。依据孔隙度与中子、密度、声波三者之间的关系,将多元回归方法应用在孔隙度计算中,实现利用多元回归方法计算火山碎屑岩储层孔隙度。其结果与岩心分析孔隙度比较接近,绝对误差仅为1. 6%,能满足储量计算要求。     

用俘获截面解释碳酸盐岩储层孔隙度的方法研究

塔河油田主要为碳酸盐岩储层,裂隙、缝洞发育,影响常规测井数据的获取和准确性。当不能获得常规测井数据时,可由PND测井的俘获截面SIGMA解释储层孔隙度。根据CATO与饱和度和孔隙度的关系,对SIGMA进行了饱和度校正。计算的孔隙度与CNL有更好的一致性,相关系数高达0.91,证明了SIGMA与CATO联合计算孔隙度方法的有效性,为碳酸盐岩储层裂隙、缝洞发育时的孔隙度解释提供了一种行之有效的方法。     

元素俘获能谱在深层致密储层评价中的应用:以松科2井为例

松辽盆地沙河子组存在多段气测异常,显示了该地层具有致密气勘探潜力,但研究区深层致密储层非均质性强、厚度薄、互层发育且类型多样。在松科2井常规测井曲线分析基础上,利用元素俘获能谱精细反演矿物组分含量,对地层岩性进行准确识别,并建立了元素交汇图识别出5种类型储层。结合岩石物理实验资料,得到地层的变骨架孔隙度、脆性指数、有机碳含量等储层评价参数,依据气测资料及综合评价指数将储层分为三大类。研究结果表明:沙河子组地层(3 730~4 500 m)孔隙度主要分布在5%~9%,脆性指数主要分布在45%~60%,临近烃源岩TOC含量普遍较高且累计厚度较大,识别出3 798 ~3 831 m、3 950~3 990 m、4 150~4 200 m是具有勘探开发潜力的有利致密气层段。利用元素俘获能谱评价致密气储层结果准确,能为深层致密气储层优选及压裂改造提供参考依据。     

阿尔金山前带东段基岩储层主要特征及评价

为了解阿尔金山前带东段基岩储层主要特征,对储层进行评价.在野外露头踏勘、岩心观察描述、薄片鉴定及相关分析测试的基础上,结合普通测井资料、成像测井资料及孔渗密实验数据等进行分析.结果表明,阿尔金山前带东段基岩储层岩性主要是岩浆岩和变质岩,岩浆岩以花岗岩为主,变质岩以片麻岩为主.储集空间类型主要包括未完全充填裂缝、溶蚀孔洞和基质微孔3类.基岩储层孔隙度介于0.004%~9.760%,平均孔隙度在1.663%~3.844%;渗透率介于最大值0.002~33.239mD,平均渗透率在0.020~3.836 mD,片麻岩类物性整体好于花岗岩类,孔隙度和渗透率之间没有明显的相关性.基岩储层主控因素包括岩性及矿物组合、古地貌、先存断裂及破碎带、后期岩浆侵入及热液作用4个方面.结合基岩储层的物性特征及形成的主控因素,建立了多因素控制的基岩储层评价标准,将储层分为好、中等、较差、差4类.评价结果表明,研究区多以Ⅱ类储层为主,主要见于花岗片麻岩,其次是花岗岩、黑云母斜长片麻岩等.东坪鼻隆、牛中-牛北斜坡可见Ⅰ类储层;尖北斜坡、牛东鼻隆、冷北斜坡多为Ⅱ类储层.      

FMI测井技术在川东北地区碳酸盐岩溶孔溶洞型储层评价中的应用

碳酸盐岩储层储集空间复杂多样,溶蚀孔洞发育且具有较强非均质性。根据实际岩心表面溶孔发育空间位置、均匀程度和溶孔直径,结合FMI特征,建立了FMI储层类型精细解释图版,并应用于非取心段储层类型的解释。川东北地区下三叠统飞仙关组、上二叠统长兴组的孔洞型储层可细分为三类:较连通型储层、局部连通型储层和孤立型储层。在储层类型解释基础上,结合FMI特征半定量地计算溶孔面孔率。通过取心井的验证,FMI测井技术可有效地应用于评价溶孔溶洞型碳酸盐岩储层的非均质性。     

Quantitative Analysis of Porosity Heterogeneity: Application of Geostatistics to Borehole Images

High levels of heterogeneity in many carbonate reservoirs have raised concerns about the validity and relevance of small-scale measurements from core plugs and high-resolution logs. While the measurements themselves may be accurate, they may not be representative of the average formation properties. A related question is one of reconciling the measurements made in small volume of investigation data (e.g., core plugs), with the measurements from relatively large volume of investigation data (e.g., wireline logs). This paper presents a technique to quantitatively describe the porosity heterogeneity in a borehole at the scale of several tenths of an inch. The method involves treating high-resolution borehole imagery as a 2D sample from a 3D data volume, and applying geostatistical analysis to these data. We compute the experimental semi-variogram and upscale its range and sill to larger (several inches) scales of measurement to predict the impact of heterogeneity on conventional core plug and logging tool porosity measurements. The resulting dispersion variance between the different measurement scales support the interpretation, application and comparison of these porosity measurements. This technique was applied to an Early Cretaceous carbonate reservoir in Abu Dhabi. We found that the scale of the heterogeneity is typically less than 1– 2in., so that while significant heterogeneity is observed at the core plug and smaller scales of measurement, the larger-volume logging tool measurements smooth out the heterogeneity and show considerably less variability. The differences between porosity measured in core plugs can be completely accounted for by this upscaling effect.     

Alam El Bueib reservoir characterization,Tut oil field,North Western Desert,Egypt

An integrated petrophysical well log and core data analysis of the Alam El Bueib-3D (AEB-3D) and Alam El Bueib-3E (AEB-3E) reservoirs, Tut oil field, North Western Desert, Egypt, allows the evaluation of those parameters critical to their reservoir characterization. The determined reservoir properties investigated are lithology, depositional paleoenvironments, shale volume, porosity, fluid saturation and net pay thickness. As the analysis of reservoir pressure is essential to specify the driving mechanisms that control the behavior of fluids within reservoirs, pressure analysis identified the Alam El Bueib-3D reservoir to be partial water drive and Alam El Bueib-3E to be an active water drive. In addition, the pressure analysis identified two additional hydrocarbon-bearing reservoirs in the Alam El Bueib-3D. Stacking patterns of the Alam El Bueib-3D reservoirs are generally fining upward sequences reflecting fluvial to braided fluvial paleoenvironments. The Alam El Bueib-3E channel shows blocky patterns with sharp bases and tops diagnostic of distributary channels. Statistical analysis of core permeability and porosity values of the Alam El Bueib reservoirs indicate very good reservoir quality. Scatter diagrams of porosity values derived from core analysis for the AEB-3D and AEB-3E reservoirs reveal correlation coefficient (r) values of 0.93 and 0.64, respectively, indicating strong linear relationships. Plots of porosity values against permeability values also show strong linear relationships in all the reservoirs indicating that the AEB-3D and the AEB-3E are permeable and with strong pore communication. The permeability measurements indicate restricted reservoir heterogeneity as plots of vertical permeability values against horizontal permeability values exhibit linear trends. Altogether, the reservoir characterization of the Alam El Bueib reservoirs demonstrates their excellent potential for hydrocarbon production.     

鄂尔多斯盆地西南部镇泾油田延长组致密砂岩储层裂缝测井识别

低孔低渗致密砂岩储层的测井评价和开发难度较大,裂缝发育情况对这类储层的产能影响十分重要。为了利用常规测井资料实现致密砂岩储层裂缝发育情况的评价,基于井径、电阻率、密度、声波等常规测井曲线计算出8个裂缝指示参数,然后利用这8个参数构建了一个裂缝综合识别参数。该裂缝综合识别参数对裂缝的识别正确率更高。将这一套方法应用于鄂尔多斯盆地西南部镇泾油田延长组致密砂岩储层裂缝识别,对于所有样本,裂缝综合指标法识别正确率为81.4%以上。本研究给出的方法对于利用常规测井资料评价致密砂岩储层的裂缝发育情况和有效性提供了一种途径,对致密砂岩储层的勘探开发有一定的帮助。     

塔河油田奥陶系生物扰动型碳酸盐岩储集层特征及其孔隙度计算样本检验模型

塔河油田奥陶系一间房组和鹰山组生物扰动型碳酸盐岩储集层非常发育。在岩芯观测的基础上,通过薄片显微镜观测与柱塞孔隙度、渗透率分析,对研究区生物扰动型碳酸盐岩储集层宏、微观特征和成因机制进行了研究,并基于岩芯资料和常规测井数据提出了一种新的孔隙度计算样本检验模型。结果表明:1）研究区奥陶系生物扰动储集层在岩芯上多呈不规则状或绸带状分布的含油斑块,潜穴充填物以白云石为主,储集空间主要是晶间孔。随着生物扰动指数的增加,平均孔隙度呈先增大后减少趋势,平均渗透率则一直增大;2）在早—中奥陶世的沉积过程中,（古）生物在碳酸盐沉积物表面或内部建造了大量形态各异的潜穴,生物扰动使得潜穴充填物和围岩基质在成分、结构上出现差异,这些差异伴随着其后的成岩过程逐步增大,这对储集层的储、渗能力具有重要影响;3）在总结前人测井孔隙度计算方法的基础上,将生物扰动指数引入密度测井孔隙度计算模型中,提出了一种基于岩芯资料和常规测井数据的孔隙度计算样本检测检验模型,并用岩芯柱塞测试结果验证了该孔隙度计算模型的有效性和准确性;该模型弥补了现有常规孔隙度模型的不足,其计算产生的拟岩芯数据可为下一步建立的通用生物扰动碳酸盐岩储集层孔隙度计算模型提供可靠的样本训练数据,也对这类储层的储量估算与油藏描述有一定的借鉴意义。     

塔里木盆地古城墟隆起西端奥陶系碳酸盐岩储层特征及预测

塔里木盆地古城墟隆起西端发现了多层奥陶系碳酸盐岩储层,深入分析其储层特征及形成机理,对油气勘探开发具有重要意义.为此利用该研究区内岩心和测井资料,对该区的奥陶系碳酸盐岩储层类型和特征进行了研究,发现该区存在多种储层类型.结合研究区的构造演化特征和岩溶系统的分析,并通过与塔河地区的储层进行对比认为储层的原生孔隙、裂缝是次生孔隙和形成的基础条件,而构造运动产生的裂缝是促进岩溶发育、控制储集体发育的关键因素,而古城墟隆起西端内幕岩溶型储层主要发育在6300 m以下.进一步应用振幅变化率、相干体分析等技术,有效预测了古城墟隆起西端的岩溶型碳酸盐岩储层发育区.     

普光地区碳酸盐岩储层孔隙类型测井识别及孔渗关系

普光地区长兴组和飞仙关组碳酸盐岩储层孔隙度与渗透率之间没有严格通用的数学关系,导致储层渗透率计算具有很大困难.通过对该地区测井资料、常规薄片、铸体薄片和岩心物性等资料进行分析,表明碳酸盐岩孔隙类型是影响孔渗关系的主要因素.基于常规测井资料构造出对孔隙结构比较敏感的测井特征:声波时差与密度比值和深浅侧向电阻率比值,可用于对该地区碳酸盐岩孔隙类型进行识别,再针对不同的孔隙类型建立相应的孔渗关系模型,用于计算该地区储层渗透率.实例资料处理结果表明,模型计算渗透率与岩心分析渗透率符合较好,且井间规律具有一致性,基于孔隙结构建立的储层孔隙度与渗透率模型能较好地确定储层渗透率.      

Logging evaluation of deep multi-type unconventional gas reservoirs in the Songliao basin,northeast China: Implications from continental scientific drilling

Multi-type unconventional gas-bearing reservoirs with different lithologies and gas accumulation potential occur in the deep part of the Songliao basin. However, the reservoirs are non-homogeneous, the gas components differ substantially, and not all types of gas-bearing reservoirs have been identified or evaluated. The International Continental Scientific Drilling Program (ICDP) is used as an example to conduct qualitative and quantitative evaluations of deep multi-type unconventional gas-bearing reservoirs using conventional and specialized logging data. The core test data are used to determine the physical properties. The porosity and permeability are compared and analyzed using different methods and models. The results show that the reservoirs have low to ultra-low porosity and ultra-low permeability. Based on the comparison of the rock mechanical parameters and mineral composition, brittleness evaluation parameters are proposed for different types of deep reservoirs in the study region. The mineral brittleness index is highly consistent with the brittleness index based on rock mechanics. An identification method for deep multi-type gas-bearing reservoir and a classification approach for different gas properties are established based on the logging response and parameter interpretation. The methane gas reservoirs have low density (DEN) and low compensated neutron logging (CNL) values and high acoustic (AC) time difference and high resistivity (RT) values. The CO₂ gas reservoirs have lower RT values and higher CNL values than the hydrocarbon gas reservoirs. The comprehensive analysis of deep gas source rock conditions and of the source-reservoir relationship provides insights for the evaluation of deep multi-type unconventional gas reservoirs. The discovery of high hydrogen content is of significant importance for developing new areas for deep natural gas exploration.     

沁水盆地南部TS地区煤层气储层测井评价方法

煤层气是一种自生自储于煤岩地层的非常规天然气资源,其储层测井评价内容及方法不同于常规天然气,在煤层气勘探开发过程中更关注于有关煤岩工业分析组分、基质孔隙度、裂缝渗透率及煤层含气量等一系列关键的储层参数。针对沁水盆地南部TS地区煤层气勘探目标层,分析了各种测井响应特征,采用回归分析法计算煤岩工业分析组分;针对煤层气含量影响因素众多且较为复杂的特点,结合相关地区煤岩样品实验分析结果,利用基于等温吸附实验的兰氏煤阶方程估算煤层含气量参数;通过煤岩孔隙结构的分析,采用变骨架密度的密度孔隙度计算公式求取煤岩总孔隙度,利用迭代逼近算法计算裂缝孔隙度;根据煤岩裂缝中面割理发育而端割理不甚发育的特点,以简化的单组系板状裂缝模型计算煤岩裂缝渗透率。通过对TS-A井进行实际计算,结果表明,煤岩工业分析组分和煤层含气量计算结果精度高,总孔隙度一般在5.5%左右,而裂缝孔隙度则大多小于0.5%,裂缝渗透率主要分布在0.001×10-3～10×10-3μm2之间,孔渗参数计算结果与相邻井区现有资料相符。采用测井方法可以快速、系统地对煤层气储层多种参数进行准确评价。     

低渗储层可动流体核磁共振研究

以低磁场核磁共振岩心分析技术为主,辅助以岩石薄片、环境电镜扫描、高分辨率CT、成像测井等常规测试手段,对国内3种不同岩性的典型低渗储层共计277块基质岩样进行可动流体评价研究,分析储层可动流体分布规律和影响因素,计算可动流体的T₂截止值和孔喉半径下限值。研究表明：低渗储层可动流体参数变化范围较宽;岩样渗透率越高,其与可动流体参数的相关性越强;低渗透砂岩储层可动流体百分数与孔隙度、渗透率之间相关性均较好,利用非线性最优化方法得到的计算值与实测值吻合很好;微裂缝的发育程度、孔径分布、粘土矿物含量及其充填程度等微观孔隙结构特征是低渗储层可动流体的主要影响因素。     

碳酸盐岩储层测井评价方法

从碳酸盐岩储层类型划分、储层参数表征、流体性质判别到数学方法的应用等4个方面,系统总结厘定了国内外碳酸盐岩储层测井评价的方法技术,通过对比分析各种评价方法可知:根据一定的数学算法,综合利用常规测井和成像测井不仅能很好地识别储层类型,而且能实现对储层参数的半定量、定量计算评价。利用常规测井和测井新技术资料(电测井和声波测井)计算出缝洞储层的各项物性参数,在此基础上,结合常规测井、核磁共振测井和偶极横波测井能进一步判别出流体性质。通过与岩心分析结果比较,该储层参数计算方法不仅有效,而且简单可行。数学算法和测井新技术资料在复杂碳酸盐岩储层评价中发挥不可替代的作用,能解决复杂碳酸盐岩储层的计算评价问题。以碳酸盐岩储层特征及流体性质为切入点,结合各种地质、地震和测井资料,借助数学算法,准确地识别储层类型与判别流体性质,并精确求解碳酸盐岩储层特别是缝洞储层的物性参数,将是碳酸盐岩储层测井评价的重点及发展方向。     

2D probabilistic prediction of sparsely measured earth properties constrained by geophysical imaging fully accounting for tomographic reconstruction ambiguity

Many hydrological, environmental, or engineering exploration tasks require predicting spatially continuous scenarios of sparsely measured borehole logging data. We present a methodology to probabilistically predict such scenarios constrained by ill-posed geophysical tomography. Our approach allows for transducing tomographic reconstruction ambiguity into the probabilistic prediction of spatially continuous target parameter scenarios. It is even applicable to data sets where petrophysical relations in the survey area are non-unique, i.e., different facies related petrophysical relations may be present. We employ static two-layer artificial neural networks (ANNs) for prediction and additionally evaluate, whether the training performance of the ANNs can be used to rank geophysical tomograms, which are mathematically equal reconstructions of physical parameter distributions in the ground. We illustrate our methodology using a realistic synthetic database for maximal control about the prediction performance and ranking potential of the approach. For doing so, we try to link geophysical radar and seismic tomography as input parameters to porosity of the ground as target parameter of ANN. However, the approach is flexible and can cope with any combination of geophysical tomograms and hydrologic, environmental or engineering target parameters. Ranking of equivalent geophysical tomograms based on additional borehole logging data is found to be generally possible, but risks remain that the ranking based on the ANN training performance does not fully coincide with the closeness of geophysical tomograms to ground truth. Since geophysical field data sets do usually not offer control options similar to those used in our synthetic database, we do not recommend the utilization of recurrent ANNs to learn weights for the individual geophysical tomograms used in the prediction procedure.