测井资料计算源岩有机碳含量模型对比及分析

Models for Calculating Organic Carbon Content from Logging Information: Comparison and Analysis

研究首先从理论分析入手,根据Δlog R模型在基线、有机碳含量背景值、叠合系数K、成熟度参数方面存在不足或不便推导出改进的Δlog R模型,针对改进Δlog R模型在测井参数选取方面的不足提出逐步回归模型;然后用具体实例对比三种模型应用效果,结合实例分析说明了改进的Δlog R模型优于Δlog R模型、逐步回归模型优于改进Δlog R模型的原因;最后建议使用逐步回归模型计算源岩有机碳含量。对区域有机碳含量评价时,建议以沉积相为基础分区带评价,同时提出几点模型应用过程中应注意事项,为今后测井途径计算有机碳含量提供方法思路。

Δlog R model is widely used in calculating organic carbon content in the past with the stepwise regression model which seldom selected.Stepwise regression model is superior to other models,the paper has proved it from both theory and practice and provided guidance for the further use of the model.So the paper can be divided into three parts. Part of model analysis and comparison： In the paper,Δlog R model,improved Δlog R model and stepwise regression model are analyzed.The article is arranged as follows： firstly,introduce Δlog R model briefly;secondly,bring in improved Δlog R model on the basis of deficiencies of Δlog R model;Finally,present stepwise regression model make up the shortages which improvedΔlog R faced.The paper begins with Δlog R model,pointing out its deficiencies on convenience、 objectivity and applicability from the aspects of baseline value,background value of TOC,composite coefficient K,maturity parameter;Improved Δlog R model is achieved under the principle of Δlog R model,it is more concise and objective by establishing functional relation between organic carbon content and logging data.But improved Δlog R model use only two curves（acoustic transit time and resistivity） to calculate TOC,the calculating results sometimes are not satisfactory.This paper attempts to calculate organic carbon content from a variety of log curves and stepwise regression model is presented.Stepwise regression model is more flexible and objective because it filters logging curves which are closely related to organic carbon content based on actual situation. Part of examples and analyses： In this part,we compare three models with examples and demonstrate the superiority of the stepwise regression model from both theoretic and practice.It is easy to know from the model analysis part that improved Δlog R model is superior to Δlog R model on convenience,objectivity and applicability.From the form of equation we can see that Δlog R model is a special form of stepwise regression model.So,stepwise regression model is better than other two models theoretically;Examples in the paper also show that calculating results of stepwise regression model are more accurate because of its advantage in logging variables selection and combination. Part of suggestions and precautions： we suggest using stepwise regression model to calculate TOC from logging curves.The method which divided large blocks into several small blocks based on sedimentary data is recommended.Log response of TOC is susceptible to interfering factors underground,some interfering factors can made certain logging curves unable to reflect organic carbon content and are excluded from the function.It is difficult to establish a universal formula for the whole block because the interfering factors vary from place to place.It is reasonable to divide large blocks into several small blocks because the factors which impact the logging response of organic carbon are similar in the same small blocks and these factors cancel each other to some extent.At last,we point out some precautions including incompact sediments,poor borehole,low porosity（tight） intervals,depth and lithology and so on.