论文部分内容阅读
杂卤石是四川盆地主要的固态钾矿物,川中地区大多数杂卤石层不纯,通常伴随石膏层、硬石膏层、盐岩层发育,甚至同层沉积,常规测井解释方法只能粗略地识别杂卤石层。以支持向量机理论和测井解释为基础,测井数据作为输入,构建预测模型,对川中地区下中三叠统杂卤石样本做精细识别,将识别结果与录井资料验证对比,正确率达到90%以上。再以预测模型为基础,结合含杂卤石岩性在测井曲线上的响应情况,构建杂卤石层分类识别模型,识别杂卤石层、石膏质杂卤石层和杂卤石膏岩层,识别正确率达到91.78%,与常规测井解释方法相比具有明显优势。结果表明,将支持向量机运用到找钾矿中具有广阔的前景。
Polyalstonite is the main solid potassium mineral in the Sichuan Basin. Most of the polyhalite layers in the central Sichuan are impure, usually accompanied by gypsum layer, anhydrite layer, salt rock formation and even the same layer deposition. The conventional log interpretation method can only roughly Identification of polyhalite layer. Based on support vector machine theory and well log interpretation, the logging data is used as input to construct the prediction model and to make a careful identification of the Lower Triassic polyhalite samples in central Sichuan Province. The verification results are compared with logging data verification, and the correctness rate Reached more than 90%. Based on the predictive model, combined with the response of the logarithm of the lithology of the polyhalite, the classification and identification model of the polyhalite layer is constructed to identify the polyhalite layer, the gypsum-type polyhalite layer and the halophilic gypsum rock layer, The recognition accuracy rate reaches 91.78%, which has obvious advantages compared with the conventional logging interpretation method. The results show that the application of SVM to find potash has a bright future.