基于PCA-BP的钻头优选方法与应用

Selection and application of bit type based on PCA-BP

周启成 3 赵春艳¹ 周伟勤⁴ 冯小琴³

中国石油新疆油田分公司，新疆 克拉玛依 834000 长江大学石油工程学院，湖北 荆州 434023 中石化中原石油工程有限公司，河南 濮阳 457001 延长石油集团研究院，陕西 西安 710000

Xinjiang Oilfield Company, PetroChina, Karamay 834000, China Petroleum Engineering College, Yangtze University, Jingzhou 434023, China Zhongyuan Petroleum Engineering Co. Ltd., SINOPEC, Puyang 457001, China Research Institute, Shaanxi Yanchang Petroleum (Group) Co. Ltd., Xi′an 710000, China

涪陵页岩气勘探开发过程中，由于岩石可钻性差、硬度高等导致的机械钻速低问题越发突出。针对这一问题，文中建立了基于主成分分析与神经网络优选钻头类型的方法。首先，根据测井解释与地统计学原理，获得岩石抗钻参数纵向与横向的分布规律；然后，对现场钻头使用情况进行统计，确定钻头使用状况的表征参数，并根据主成分分析法的降阶原理对多维表征参数进行处理，获得综合表征参数；最后，利用BP神经网络建立岩石抗钻参数、地层层位和钻头类型与综合表征参数之间的映射关系，从而实现钻头类型优选。以焦石坝地区的现场资料进行实例分析，结果表明，该方法实用性强，能够满足工程实际需要。

With exploration and development of shale gas in Fuling Area, the problems of poor drillability, high hardness and low drilling rate became more and more significant. Considering this condition, the method of bit type selection, based on principal component analysis and BP neural network, was established in this article. Firstly, the distribution of rock resistance parameters on vertical and horizontal were built using log interpretation and geostatistics. Then, the characteristic parameters of bit conditions were defined from counting the data of bit conditions. The synthetical characteristic parameter was calculated using principal component analysis. Finally, according to the BP neural network, mapping relation between rock resistance parameters, formations, bit types and synthetical characteristic parameter was established. With this mapping relation, the optimal selection of bit type is achieved. Taking the data of Jiaoshiba Area as the examples, the result of analysis shows that this method is practical and could satisfy the needs of drilling engineering.

10.6056/dkyqt201602025