富县地区碳酸盐岩测井岩性识别方法

Logging-based method to identify the lithology in carbonate rocks,Fuxian area

为了解决鄂尔多斯盆地下古生界奥陶系马家沟组碳酸盐岩储层岩性类型复杂、变化快、难识别以及部分老井无测井解释剖面,地质研究人员无法快速有效判别岩性等问题,以富县地区马家沟组马五段为研究对象,分析了目前常用的岩性识别方法,通过取心资料确定了主要的岩石类型,结合碳酸盐岩中主要矿物的测井响应数值,得出了可采用光电吸收截面指数曲线与补偿密度曲线、补偿密度与补偿中子曲线两两包络的包络法与聚类分析—最小临近算法实现岩性解读的认识。研究结果表明:(1)包络法操作简便,能快速识别白云岩、石灰岩、石膏,尤其在含膏地层中优势明显,适用于生产中对岩性的预判,其操作关键点在于曲线左右刻度值的调整,其判别准确与否的关键在于曲线质量是否可靠;(2)聚类分析—最小临近算法结果更精确,其预测符合率高达92.31%,更适用于后期科研所需,但是该方法需要一定量的取心数据作为支撑;(3)目前上述两种方法在坍塌角砾岩的识别中都还存在着局限性,对于坍塌角砾岩的识别还需要借助成像测井以及地质认识来实现。

Most carbonate reservoirs of Lower Paleozoic Ordovician Majiagou Formation,Ordos Basin,can be hardly identified due to complex lithology and fast lithologic change.Furthermore,owing to some old wells without logging interpretation,geologists cannot conduct rapid and effective lithologic identification.So,Majiagou 5 Member in Fuxian area was taken as an objective to analyze the commonly-used identification methods.Then,the main rock types were determined based on coring data.And combined with the logging response values of the main minerals in carbonate rocks,the lithology was interpreted by means of the envelope method of photoelectric absorption cross section index curve-compensated density curve and compensated density curve-compensated neutron curve,and the clustering analysis-K-Nearest Neighbor.Results show that(1)the envelope method can easily determine dolomite,limestone,and gypsum,and especially does well in certain strata with gypsum.It is applicable to the lithological pre-judgment during production.Its key point lies in adjusting left and right scale values in the curves and its discrimination accuracy is dependent on the reliability of curve quality;(2)the clustering analysis-K-Nearest Neighbor provides more accurate results,and its prediction coincidence rate in this study is up to 92.31%.This method is more suitable for the later research,but it also needs more coring data;and(3)at present,both methods have limitations in the identification of collapsed breccia.Therefore,this identification cannot be realized without imaging logging and geological cognition.