基于研究区块的高分辨率阵列感应测井(HDIL)资料,首先分析了利用交会图法进行流体识别的效果,指出了该常规解释方法的局限性。在此基础上,将基于遗传算法优化的支持向量机(GA-SVM)引入到测井解释当中,结合阵列感应测井和常规测井资料,建立了流体识别的非线性模型。利用该模型对45个训练样本进行回判,准确率为100%,对21个预测样本进行预测的准确率为90.48%,相对于常规解释方法,该流体识别方法具有更高的准确率。 Based on the high resolution array induction logging (HDIL) data of the study area, the effect of fluid identification by the cross plot method is analyzed first, and the limitations of the conventional interpretation method are pointed out. Based on this, GA-SVM is introduced into logging interpretation. Combined with array induction logging and conventional logging data, a nonlinear model of fluid identification is established. 45 training samples were judged by this model with 100% accuracy rate and 90.48% accuracy rate of 21 prediction samples. Compared with the conventional interpretation method, this method has a higher accuracy.