史肖燕,季勇,崔猛,李忠明,赵飞. 基于符号聚合近似法的钻井液漏失类型自动识别[J]. 石油钻采工艺,2023,45(6):696-703. DOI: 10.13639/j.odpt.202302038
引用本文: 史肖燕,季勇,崔猛,李忠明,赵飞. 基于符号聚合近似法的钻井液漏失类型自动识别[J]. 石油钻采工艺,2023,45(6):696-703. DOI: 10.13639/j.odpt.202302038
SHI Xiaoyan, JI Yong, CUI Meng, LI Zhongming, ZHAO Fei. Automatic identification of drilling fluid loss types based on symbolic aggregate approximation[J]. Oil Drilling & Production Technology, 2023, 45(6): 696-703. DOI: 10.13639/j.odpt.202302038
Citation: SHI Xiaoyan, JI Yong, CUI Meng, LI Zhongming, ZHAO Fei. Automatic identification of drilling fluid loss types based on symbolic aggregate approximation[J]. Oil Drilling & Production Technology, 2023, 45(6): 696-703. DOI: 10.13639/j.odpt.202302038

基于符号聚合近似法的钻井液漏失类型自动识别

Automatic identification of drilling fluid loss types based on symbolic aggregate approximation

  • 摘要: 目前钻井液漏失类型判断需基于详细的地质工程信息,并辅以人工分析,漏失类型识别存在着主观性和滞后性。基于钻井液漏失原因和参数表征规律,建立了裂缝性、孔隙性、溶洞性和诱导裂缝性4种漏失类型特征曲线模版,并采用符号聚合近似(SAX)方法对特征曲线进行字符化转换;通过获取待识别井漏失特征曲线的SAX字符串,并计算与模版字符串的相似性,再根据量化的相似度自动识别漏失类型。实例井验证结果显示,该方法能直接利用录井数据自动识别漏失类型,识别效率相比传统的人工分析方法提升了90%以上。该方法既可用于海量历史数据挖掘分析,为后续钻井提供指导,也可用于实时漏失类型判断,为堵漏措施的选择提供科学依据。

     

    Abstract: Currently, the determination of drilling fluid loss types relies on detailed geological engineering information, supplemented by manual analysis, leading to subjectivity and delay in identification. Based on the drilling fluid loss causes and parameter characterization patterns, characteristic curve templates for four types of fluid loss types, that is, fracture, pore, dissolution, and induced fracture, were established, and the characteristic curves were transformed into symbolic sequences using symbolic aggregate approximation (SAX) method. By comparing the SAX string representation of the characteristic curve from the well under investigation with template strings and calculating the similarity, the fluid loss type can be automatically identified based on quantified similarity. Validation results from sample wells show that this method, utilizing logging data directly, can automatically identify the drilling fluid loss types, and achieves an identification efficiency improvement of over 90% compared to traditional manual analysis methods. This approach can be applied to large-scale historical data mining for analysis to guide future drilling operations, or can be applied to real-time fluid loss type judgement to provide scientific basis for selecting plugging measures.

     

/

返回文章
返回