论文部分内容阅读
在碳酸盐岩地层中,裂缝是非常重要的储集空间类型,因此,识别裂缝具有十分重要的意义。普光气田碳酸盐岩储层具有很强的非均质性和各项异性,储层类型复杂,为裂缝的识别增加了较大的难度。本文主要以常规测井资料为主,结合岩心资料、薄片分析资料以及成像测井资料,总结出不同类型的裂缝的测井响应特征。根据对裂缝测井响应特征的分析,引入支持向量机(SVM)方法,通过对所选样本进行归一化训练,并对普光气田4口井进行预测,准确率高达80%。
In carbonate formations, fractures are very important reservoir space types, so identifying fractures is of great importance. Carbonate reservoirs in Puguang Gas Field have strong heterogeneity and anisotropy, and the types of reservoirs are complex, which increases the difficulty of identifying fractures. In this paper, we mainly use conventional well logging data. Combining with core data, slice analysis data and imaging logging data, the logging response characteristics of different types of fractures are summarized. Based on the analysis of the response characteristics of fracture logging, the support vector machine (SVM) method is introduced. Through the normalized training on the selected samples, the prediction of 4 wells in the Puguang Gasfield is as accurate as 80%.