运用Fisher判别法实现储层中CO_(2)的识别

Identification of CO_(2)in Reservoirs Using Fisher Discriminant Method

我国某地勘探发现大量油气伴生的CO_(2),给后续的构造评价带来很大挑战。为提高CO_(2)识别有效性,提出将Fisher判别法用于储层中CO_(2)识别的思想。考虑流体性质差异,优选中子、密度两个测井敏感参数,C1/TGAS、C12/C345、C1/C2345三个录井敏感参数,并利用CO_(2)和烃类气的物理性质差异,创造性地引入热中子截面(TNXS)参数,作为能反映储层特点的原始判别因素。采用Fisher判别法对样本点进行反复训练,确定原始判别因素的线性组合形式,建立判别函数。通过判别图版确定流体类型,表明Fisher判别法对CO_(2)气层、烃类气层以及油水层有较好的区分度。Fisher判别法从传统技术、模型、算法的复杂性中超脱了出来,通过数据特征实现了数据类别的自适应识别,给复杂储层构造的油气勘探开发提供了新思路。

A large amount of CO_(2)associated with oil and gas was discovered in an exploration in China,which brought great challenges to the subsequent structural evaluation.In order to improve the effectiveness of CO_(2)identification,idea of using Fisher discriminant method was proposed to identify CO_(2)in reservoirs.Considering the differences of fluid properties,two measurement logging sensitive parameters,neutron and density,and three record logging sensitive parameters,C1/TGAS,C12/C345 and C1/C2345,were selected out.By taking advantage of the differences of physical properties between CO_(2)and hydrocarbon gas,the thermal neutron cross section(TNXS)parameter was creatively introduced.All these parameters were taken as the original discriminating factors that can reflect the characteristics of the reservoir.Fisher discriminant method was used to train the sample points repeatedly to determine the linear combination form of the original discriminant factors and the discriminant function was established.The fluid types were determined by the discriminant charts,showing that the Fisher discriminant method had a good discrimination for the CO_(2)gas layer,hydrocarbon gas layer and oil-water layer.Fisher discriminant method breaks away from the complexity of traditional technologies,models and algorithms,and realizes the adaptive identification of data categories through data characteristics,which provides a new idea for the oil and gas exploration and development in complex reservoir structures.