Identification of"weakly responsive thick sand bodies"in braided river sand bodies using limited learning machine thickness prediction technology based on waveform constraints
Bohai W Oilfield is in the stage of high water cut and high recovery,and it is urgent to conduct in-depth study to maintain stable production.The main oil layer is in braided river sedimentation,with fast ver-tical and horizontal changes in sand bodies.In areas with strong seismic response,the well network is bet-ter.In areas with weak seismic response,the reservoir is not developed and the well network is incomplete.According to drilling statistics,not all areas with weak seismic response are thin sand bodies.The conven-tional meandering river sand body exhibits a"mud wrapped sand"feature,and the strength of its seismic response is correlated with the sand body thickness,while the braided river sand body exhibits a"sand wrapped mud".Through forward simulation analysis,it is shown that the sand body is affected by the sur-rounding rock and interlayer,and the strength of seismic response is irregular with the sand body thickness.By studying the limited learning machine thickness prediction technology based on waveform constraints,waveform classification of multiple reservoir structures in braided river sand bodies is carried out,and the limited learning of multi-attribute clustering is carried out to obtain the optimal estimation of multi-attrib-ute and sand group thickness.The research results have effectively identified the potential area of"weakly responsive thick sand bodies",assisting in the estimation of remaining oil,providing assistance for well po-sition adjustment,and also providing reference for thickness prediction in similar oil fields.
国家科技期刊平台
NSTL
万方数据
