海上油田薄储层预测技术研究

中海石油(中国)有限公司深圳分公司,深圳 518000

薄层预测;神经网络曲线模拟;敏感参数构建;地震波形相控模拟

Research on Prediction Technology of Thin Reservoirs in Offshore Oilfield
HOU Kai,XIONG Shuquan,WAN Nianhui,LIU Ping,LI Yongfeng,LIANG Quanquan,YANG Xiaojiang

Research Institute , CNOOC Ltd., Shenzhen 518000,China

Thin layer prediction;Neural network curve simulation;Sensitive parameters construction;Seismic waveform driving inversion

DOI: 10.13512/j.hndz.2022.01.13

备注

针对南海东部A油田钻井资料少,储层预测不确定性大的难题,探索了神经网络曲线模拟方法,模拟生产井的声波曲线,增加地震反演的井控程度,提升储层预测准确性。同时针对油田3~5m的薄层数量多、储量占比大、非均质强的问题,讨论了一种地震波形相控模拟方法,该方法在提高储层预测纵向分辨率的同时,也有较为可靠的横向分辨率,体现了地震波形特征,符合地层的地质沉积规律,有效的减少了反演的随机性,为海上油田薄层油藏的增产挖潜提供了有力支持。
Aiming at the problem of less drilling data and large uncertainty of reservoir prediction in A oilfield in the eastern South China Sea,this paper explores the use of neural network learning method to simulate the acoustic curve of production wells,and increase the well control of seismic inversion and improve the accuracy of reservoir prediction. At the same time, for the problems of large number of thin layers, large proportion of reserves and strong heterogeneity during 3-5 m of A oilfield in the east of the South China Sea,a seismic waveform controlled inversion method is discussed. This method not only improves the vertical resolution of reservoir prediction, but also has a more reliable horizontal resolution,which reflects the characteristics of seismic wave shape,conforms to the geological sedimentary law of the formation,effectively reduces the randomness of inversion,and provides strong support for increasing production and tapping potential of thin reservoir in offshore oilfield.
·