多阈值BIRCH聚类在相控孔隙度预测中的应用

岩相及孔隙度预测在油气勘探中非常重要，为此，提出一种基于多阈值BIRCH（Balanced Iterative Reducing and Clustering Using Hierarchies）聚类的岩相预测方法，并以此为基础利用岭回归算法预测孔隙度。首先，根据地震波阻抗数据分布规律启发式设定初始阈值，根据簇之间体积的不一致性，动态增加阈值，使用Agglomerative算法进行全局聚类以划分岩相；然后，以井点处孔隙度和地震波阻抗数据为输入，在同一岩相内采用改进的岭回归方法预测孔隙度。模型实验表明，多阈值BIRCH聚类方法具有良好的稳定性和较高的计算效率，岩相划分准确。实际数据结果表明，该方法能够准确预测孔隙度。

Application of multi-threshold BIRCH clustering to facies-controlled porosity estimation

In view of the importance of lithofacies and porosity study in hydrocarbon exploration,we present an approach of multi-threshold BIRCH clustering for lithofacies classification,based on which we estimate porosity using ridge regression.The heuristic initial threshold is set in terms of wave impedance distribution,and the number of thresholds is increased dynamically according to inter-clustering volume inconsistency.Global Agglomerative clustering is then employed for lithofacies classification.For each lithofacies,a modified ridge regression algorithm is used to predict porosity based on well porosity.Model tests show that multi-threshold BIRCH clustering exhibits good robustness and computational efficiency for accurate lithofacies classification.A field data test shows that porosity could be estimated accurately using this method.