论文部分内容阅读
尽管通过测井数据预测烃源岩有机碳含量(TOC)具有经济快速的明显优势,但由于不同地区及地层的特殊性,其预测误差限制了该方法的广泛应用。利用前人提出的预测烃源岩TOC含量的自然伽马测井法、体积密度测井法、多元回归分析法和BP神经网络模型,结合鄂尔多斯盆地延长组的地质特征建立了研究区TOC含量预测的经验公式和模型。优选认为BP模型为延长组TOC含量预测精度最高的方法,并将其应用于盆地南部YK1井延长组,得到9 220个样点的TOC值,辨识出4段连续厚度大于10m的有效烃源岩,分别位于长_9~1段、长_7~3段、长_7~2段和长_7~1段,并且长_7~2—长_7~3段为TOC含量最高的优质层段。研究发现,这些有效烃源岩层段均不同程度地发育油页岩,并且同一层段TOC含量预测值在不同地区的钻井中可能出现明显的差异,这与延长组不同沉积时期古湖泊沉积相的分布和差异相一致。
Although log-based prediction of organic carbon content (TOC) in source rocks has obvious economic advantages, its prediction error limits the extensive application of this method because of the particularities of different regions and strata. The natural gamma ray logging method, bulk density logging method, multiple regression analysis method and BP neural network model which are used to forecast the TOC content of source rocks proposed in the past are used to establish the prediction of TOC content in Yanchang Formation of Ordos Basin. Empirical formula and model. BP model is considered to be the best method to predict the TOC content in the extended group and to apply it to the extended group YK1 in the southern part of the basin to obtain the TOC values of 9 220 samples and to identify 4 effective source rocks with a continuous thickness greater than 10 m , Respectively, located in the long _9 ~ 1, long _7 ~ 3, long _7 ~ 2 and long _7 ~ 1, and long _7 ~ 2 ~ _7 ~ 3 paragraphs for the TOC content of the highest quality Layers. It is found that these effective hydrocarbon source rocks develop oil shale to some extents, and the predictions of TOC content in the same layer may have obvious differences in drilling in different areas. This is in line with the fact that the sedimentary facies of ancient lakes in different depositional periods Distribution and differences are consistent.