ΔLogR技术与BP神经网络在复杂岩性致密层有机质评价中的应用

Application of ΔLogR technology and BP neural network in organic evaluation in the complex lithology tight stratum

本文系统分析ΔLogR技术应用于复杂岩性致密层有机质评价中存在两方面的局限性:参数选取方面,测井曲线选取过于单一,无法有效削弱致密层段复杂岩性和孔隙度等因素对计算有机碳含量的影响;构建模型方面,人为剔除异常点存在随机性与偶然性误差,影响建模准确性.针对上述问题,本文建立了BP神经网络模型,并将其应用于柳河盆地柳参1井下桦皮甸子组烃源岩有机质评价.研究结果表明,在不剔除异常点情况下,BP神经网络模型计算TOC值和实测116组TOC值相关性达到0.886,显示建模效果良好.分别应用BP神经网络和ΔLogR模型,计算研究区致密层纵向上连续的TOC曲线,BP神经网络模型的计算TOC曲线与实测TOC数据基本吻合,而ΔLogR模型的计算TOC曲线吻合度较差.因此在测井资料完善的情况下,本文建议使用BP神经网络评价复杂岩性的致密层有机质.

This paper points out the ΔLogR technology has a couple of limitations for organic evaluation in the complex lithology tight stratum through systematic analysis. On the choice of parameter,selection of well logging parameters is too single to weaken effectively the influence of complex lithology and porosity,and other factors to the calculation of organic content in tight layer; on the construction of the model,eliminating artificially abnormal points brings random errors to affect the accuracy of the model. Aimed at these problems,the BP neural network model was established to calculate vertical continuous organic carbon content in tight stratum of the Xiahuapidianzi Formation in Liuchan-1 well,Liuhe basin.Without eliminating artificially abnormal points, the correlation between the TOC values calculated by BP neural network model and the 116 sets of measured TOC values reaches to 0. 886,which shows the modeling effect of BP neural network is good. Then the BP neural network model and the ΔLogR model were respectively applied to work out the vertical continuous TOC curve of the tight layer. The TOC curve calculated form the BP neural network model agrees well with the measured TOC values,while the latter does not. Therefore,in the case of complete logging data,this article recommends building the BP neural network model for evaluating organic matter in complex lithology tight stratum.