基于生长分层自组织映射网络的岩性识别模型

针对复杂类型油气储层的岩性识别难的问题，提出了一种新的基于生长分层自组织映射网络的岩性识别模型，并对金衢盆地金66测井数据进行仿真试验。结果表明，它是一种操作简便、易于实现的模型；既保留了自组织映射网络的优点，又具备其自身的优势；不但能对输入数据进行正确的聚类，而且能将输人数据中的层次继承关系直观地展现出来，从而可进一步提取出输入数据的共性与特性，并有助于对高维数据的深层次分析。在复杂的油气藏领域中可以应用生长分层自组织映射网络进行岩性识别，具有广泛的应用前景。

Lithology Identification Based on Growing Hierarchical Self-organizing Map Network

In order to solve the problem of hardly identifying the lithology on the complex reservoir, we bring forward a model on the identification of lithology based on growing hierarchical self-organizing map networks, and a test on JIN66 well to identify well lithology has been made. The result of simulation indicated that it is practicable and effective. The network retained the advantages of the self-organizing map, and had its own advantages. It not only correctly classifies the input data, but also directly demonstrates the hypotaxis of category inheritance on the input data. So it can distill the universality and characteristics of the input data, and contribute to a deeper level of analysis of the high-dimensional data. Therefore in the field of complex reservoir,growing hierarchical self-organizing map network has used for lithology identification, which has broad application prospects.