摘要
井漏是钻井工程过程中经常遇到的复杂情况,极易引发井壁坍塌、溢流等事故。针对传统井漏研究机理复杂、边界条件考虑欠缺、随钻控制仅有监测的问题,应用大数据处理思路,采用数据关联性分析、归一化处理、离群点处理、不平衡处理等数据处理技术,优选深度神经网络模型,通过模型“创建-训练-优化”的关键技术,建立井漏预测模型方法。
Lost circulation is a complicated situation often encountered in the process of drilling engineering,and it is easy to cause accidents such as collapse of the well wall and overflow.Targetting at the complex mechanism of traditional lost circulation research,the lack of consideration of boundary conditions,the only monitoring while drilling control,the idea of big data processing is applied,and the data correlation analysis,normalization processing,outlier processing,unbalanced processing and other data processing technologies are adopted.The deep neural network model is optimized,and the method of lost circulation prediction model is established through the key technologies of model“creation-training-optimization”.
作者
和鹏飞
刘晓宾
陈真
史旻
陈玉山
HE Pengfei;LIU Xiaobin;CHEN Zhen;SHI Min;CHEN Yushan(Drilling and Production Company,CNOOC Energy Technology and Services Limited,Tianjin 300452,China;Tianjin Branch,China National Offshore Oil Corporation Limited,Tianjin 300450,China)
出处
《天津科技》
2019年第S01期21-23,共3页
Tianjin Science & Technology
关键词
数据处理
深度神经网络
井漏
钻井
data processing
deep neural network
lost circulation
drilling
