数据驱动的页岩油水平井压裂施工参数智能优化研究

Research on Data-Driven Intelligent Optimization of Fracturing Treatment Parameters for Shale Oil Horizontal Wells

曾凡辉 胡大淦 张宇 郭建春 田福春 郑彬涛

ZENG Fanhui HU Dagan ZHANG Yu GUO Jianchun TIAN Fuchun ZHENG Bintao

油气藏地质及开发工程全国重点实验室（西南石油大学）， 四川成都 610500 中国石油大港油田分公司石油工程研究院，天津 300280 中国石化胜利油田分公司石油工程技术研究院，山东东营 257000

National Key Laboratory of Oil and Gas Reservoir Geology and Exploitation(Southwest Petroleum University), Chengdu, Sichuan, 610500, China Petroleum Engineering Research Institute, PetroChina Dagang Oilfield Company, Tianjin, 300280, China Petroleum Engineering Technology Research Institute, Sinopec Shengli Oilfield Company, Dongying, Shandong, 257000, China

针对目前数智化压裂施工参数设计针对性不足、流程不畅通等问题，建立了基于数据驱动的压裂施工参数智能优化方法。以CD区块32口页岩油井为研究对象，采用主成分分析法处理代表储层地质特征、工程品质和施工参数的15项产量影响因素，使之降低维度，引入高斯隶属函数和熵权法进行储层压裂非均质性模糊综合评价，结合支持向量回归和粒子群优化算法，以产量最高为目标，推荐射孔位置、段长、簇间距、单位长度液量、单位长度砂量和排量。研究结果表明，渗透率、孔隙度、热解游离烃含量、单位长度液量和单位长度砂量为研究区块的产量主控因素。应用实例井采用优化的参数施工后，第一压裂段8簇均成功起裂，裂缝半长59.50～154.80 m，产量预测符合率为94.86%。研究表明，该方法可实现有效储层质量评价、产量预测和匹配储层地质条件施工参数的快速优化，推动页岩油等非常规储层高效开发。

A data-driven intelligent optimization method for fracturing treatment parameters was proposed to address the issues of insufficient pertinence and incomplete process design in digital fracturing treatment parameters. With 32 shale oil wells in the CD block as the research object, principal component analysis was used to reduce the 15 production-influencing factor dimensions representing geological attributes, engineering quality, and construction parameters of the reservoir. A Gaussian membership function and entropy weight method were introduced for a fuzzy comprehensive evaluation of reservoir fracturing heterogeneity. Combined with support vector regression and particle swarm optimization algorithms, the perforation location, segment length, cluster spacing, fracturing fluid intensity, sanding intensity, and discharge capacity were recommended with the highest production as the goal. The research results indicated that permeability, porosity, free hydrocarbon content by pyrolysis, fracturing fluid intensity, and sanding intensity were the main control factors for the production of the target block. All eight clusters of the first fracturing section of the application well have successfully initiated fractures during treatment with optimized parameters, with a half-length of 59.50–154.80 m and a production prediction accuracy of 94.86%. The method proposed can achieve effective reservoir quality evaluation, production prediction, and rapid optimization of treatment parameters that match reservoir geological conditions, promoting efficient shale oil development in unconventional reservoirs.

国家自然科学基金面上项目“大数据驱动的深层页岩压裂参数协同优化与实时调控研究” （编号：52374045）、四川省自然科学基金面上项目“深层页岩储层多簇射孔压裂竞争扩展多目标协同智能优化与调控”（编号：2023NSFSC0424）联合资助

https://doi.org/10.11911/syztjs.2023087