基于变系数ΔlogR技术的烃源岩TOC精细评价

TOC evaluation using variable-coefficient ΔlogR model

关键词:  测井  烃源岩  总有机碳  变系数ΔlogR技术  乌尔逊凹陷  海拉尔盆地

Key words:  well logging  source rocks  total organic carbon(TOC)  variable-coefficient ΔlogR model  Wuerxun sag  Hailaer Basin

摘要:

海拉尔盆地乌尔逊凹陷主力烃源岩层(南一段)烃源岩非均质性强,总有机碳(TOC)波动明显,利用有限的测试样品表征整套烃源岩层的生烃潜力具有较强的不确定性;烃源岩层的测井响应表现为高自然伽马、高声波时差、中等电阻率,易于识别,但是烃源岩测井评价参数的通用性差,难以建立起具有普遍适用性的TOC预测公式。针对上述问题,采用变系数ΔlogR技术评价该区烃源岩TOC,将ΔlogR技术中的经验参数视为待定系数,从模型参数地质意义出发,通过针对单井和“分区、分相”预测模型中的待定参数,完成全区120口井TOC测井评价,利用测井计算TOC实现烃源岩分级评价,通过“钻井定频率、层序定展布”的方法,得到南一段不同丰度级别烃源岩的厚度。结果表明,变系数ΔlogR技术预测TOC误差平均为16.6％,比传统方法平均降低19.9％;南一段差烃源岩厚度一般15～20 m,中等烃源岩厚度一般10～20 m,优质烃源岩的厚度一般20～45 m;南一段差烃源岩、中等烃源岩、优质烃源岩的体积分别为0.46×10¹¹、1.08×10¹¹和1.59×10¹¹ m³。

Abstract:

The source rocks in the First Member of Nantun Formation of Wuerxun sag, Hailaer Basin are characterized by strong organic heterogeneity with large variations in total organic carbon(TOC) values, resulting in the organic abundance of source rocks which cannot be represented by limited and discrete test samples. Source-rock sections can be easily recognized by high natural gamma ray, high interval transit time and medium resistivity logging response, but it is difficult to establish a universal TOC prediction formula, because the key parameters in ΔlogR model vary from well to well. Unlike in conventional approaches where ΔlogR is approximated empirically, an new approach that assumes ΔlogR as unknow parameters (k and a) is applied. In the variable-coefficient ΔlogR model, "k" is predicted according to single well, whereas "a" is predicted in the light of zones and facies. This variable-coefficient ΔlogR model is applied to 120 wells. Sources rocks in the First Member of Nantun Formation are divided into three grades according to the relationship between measured TOC and Hydrogen index (IH). TOC limit for poor source rocks, medium source rocks and excellent source rocks are 1.0％ and 2.0％, thus logging-derived TOC is used to classify source rocks. Within the sequence stratigraphic framework, the depth of source rocks of various grades is determined by calculating thickness ratio of various-grade source rocks in system tracts of single well and by using the Kriging interpolation between all the wells. The results show that logging calculated TOC by the variable-coefficient ΔlogR model agrees well with the measured values. The TOC predicting error by the variable-coefficient ΔlogR model is about 16.6％, with 19.9％ reduction than the traditional ΔlogR model. For the three rock types, poor source rocks, medium source rocks, and excellent source rocks, their depths are 15-20 m, 10-20 m, and 20-45 m; and their volumes are 0.46×10¹¹m³,0.46×10¹¹m³, and 1.08×10¹¹m³, respectively. These three types account for 14.7％, 34.5％, and 50.8％ of total source rocks in volume.