基于多井资料的天然气水合物储层随机模拟研究与应用

发布时间：2018-06-24 20:22

本文选题：随机模拟 + 孔隙度　；参考：《吉林大学》2015年硕士论文

【摘要】：天然气水合物的研究逐渐成为了世界上能源科学研究的热点，它广泛分布于极地冻土带及海底，其资源量相当巨大，有专家大胆估测天然气水合物的含碳量是所有煤、石油和天然气中碳含量的2倍，我国海域辽阔，天然气水合物富集区域多。本文的研究目的和意义为利用对单井资料的分析和处理明确天然气水合物矿体的层位并得到符合天然气水合物储集层的评价模型和方法；利用多井资料进行随机模拟研究以得到除井孔以外区域的孔隙度、含水合物饱和度分布情况，为研究区域水合物进一步勘探和研究提供必要的参数和依据。本文在分析天然气水合物的物理性质及研究区地层的物理性质基础上，明确了研究区域天然气水合物的层位，在前人工作研究的基础上建立了符合研究区域的孔隙度计算方法（校正后的密度孔隙度法）、饱和度计算方法（优选阿尔奇公式及双水模型）。同时本文提出了利用主成分分析法结合层内差异法及聚类分析法对研究区井进行自动分层处理，利用主成分分析法是为了综合各条高分辨率测井曲线的特征及减少用于分层的曲线条数方便计算，在利用层内差异法处理之后做聚类分析可以避免一些尖刺造成的分层过细。在单井分层的基础上，对分层数据做横向对比分析得到用于随机模拟研究的研究区的层面数据。此次在Petrel软件上利用随机模拟的方法对井孔以外的属性数据进行模拟研究。综合各随机模拟方法的优缺点及适用性选择序贯高斯模拟作为模拟方法。随机模拟一般分为数据准备、模型建立及模拟三步。以井数据为主的数据主要包括井的基本信息数据、测井数据、测井解释数据及分层数据等；利用处理得到层面数据建立模型，并在此基础上对模型进行网格划分，每个网格平面上代表100m*100m、厚度上根据层厚等比例划分精度一般在1m左右；在模拟之前，需对属性数据做统计分析及空间相关性分析，统计分析包括均值、最值、方差、协方差及相关性系数等；空间相关性分析是为了得到满足区域的变差函数模型。最终，在通过统计对比和抽样法的检验之后得到了反映全研究区的三维的属性数据（孔隙度、含天然气水合物饱和度）的分布。在现有的资料的基础上，本文处理得到的属性分布基本反映了研究区的分布特征，孔隙度的分布较为准确，，含天然气水合物饱和度的分布在趋势上基本符合，具体精度没有孔隙度高，下一步需搜集更多的地质、地震资料，完善储层数据库，这样模拟的精度会更高。
[Abstract]:The research of natural gas hydrate has gradually become the hot spot of energy science research in the world. It is widely distributed in the polar frozen soil zone and seabed, and its resources are very large. Some experts boldly estimate that the carbon content of natural gas hydrate is all coal. The carbon content in oil and natural gas is twice as large as that in China, and there are many gas hydrate enrichment areas in China. The purpose and significance of this paper is to use the single well data analysis and processing to determine the horizon of the gas hydrate ore body and to obtain the evaluation model and method that accord with the gas hydrate reservoir. In order to obtain the distribution of porosity and hydrate saturation in the area other than the well hole, the random simulation of multi-well data provides necessary parameters and basis for further exploration and study of hydrate in the studied area. Based on the analysis of the physical properties of natural gas hydrate and the physical properties of strata in the study area, the layer of natural gas hydrate in the study area is determined in this paper. On the basis of previous studies, the methods of porosity calculation (density porosity method after correction) and saturation calculation method (optimal selection of Archie formula and two-water model) were established in accordance with the study area. At the same time, this paper puts forward to use principal component analysis method combined with intra-layer difference method and cluster analysis to deal with the automatic stratification of wells in the study area. The principal component analysis (PCA) is used to synthesize the characteristics of each high resolution logging curve and to reduce the number of curves used for stratification. Cluster analysis can be used to avoid the delamination caused by some spikes. On the basis of single well stratification, the stratified data are analyzed horizontally to obtain the stratified data of the study area used in the random simulation study. The method of random simulation is used to simulate the attribute data outside the well hole in Petrel software. Considering the advantages and disadvantages and applicability of each stochastic simulation method, Sequential Gao Si simulation is chosen as the simulation method. Random simulation is generally divided into three steps: data preparation, modeling and simulation. The data based on well data mainly include well basic information data, log interpretation data and stratified data. Each mesh plane represents 100mb, and the accuracy of dividing the thickness according to the scale of layer thickness is generally about 1m. Before the simulation, statistical analysis and spatial correlation analysis of attribute data are needed. The statistical analysis includes mean value, maximum value, variance, and so on. Covariance and correlation coefficient, etc. The spatial correlation analysis is to obtain the variation function model of satisfying region. Finally, the distribution of 3D attribute data (porosity, gas hydrate saturation) reflecting the whole study area is obtained by statistical comparison and sampling. On the basis of the existing data, the attribute distribution obtained in this paper basically reflects the distribution characteristics of the study area, the distribution of porosity is more accurate, and the distribution of gas hydrate saturation is basically consistent with the trend. The accuracy is not as high as porosity, the next step is to collect more geological, seismic data and improve reservoir database, so the accuracy of simulation will be higher.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：P618.13

【参考文献】