摘要
基于油井基础数据、生产动态数据以及现场实时采集功图、电参、温压等实时数据,通过XGBoost大数据方法建立动液面计算模型。在大数据模型建立过程中,首先对影响动液面的因素进行相关性敏感性分析,最终选取24个相关参数;其次应用机器学习及深度学习算法,建立动液面计算模型,通过实测数据对模型计算结果进行验证,并对模型不断迭代优化,最终形成高精度动液面计算模型。
Based on the basic data of oil wells,production dynamic data and real-time data such as on-site realtime data acquisition,electrical parameters,temperature and pressure,this paper establishes a dynamic liquid level calculation model through XGBoost big data method.In the process of building the big data model,the correlation sensitivity analysis of the factors affecting the dynamic liquid level is carried out first,and 24 relevant parameters are finally selected;Secondly,the machine learning and deep learning algorithms are applied to establish the dynamic liquid level calculation model.The model calculation results are verified by the measured data,and the model is iteratively optimized to finally form a high-precision dynamic liquid level calculation model.
作者
汪金如
Wang Jinru(Taizhou Oil Production Plant of East China Oil and Gas Branch of China Petroleum and Chemical Corporation,Jiangsu 225300)
出处
《石化技术》
CAS
2023年第6期1-3,共3页
Petrochemical Industry Technology