石油地球物理勘探2009，Vol.44Issue(5)：609-612,4.
井、震多尺度信息融合预测老油田浅层岩性气藏
Using well and seismic multi-scale information integration to predict shallow lithology gas reservoir in mature oilfield
摘要
Abstract
The lithology distribution for shallow gas reservoir in mature oilfield is quite complex, using seismic attribute analysis, gas-bearing sandstone forward modeling and sonic parameter prediction method which is based on Probability Neural Networks (PNN) to predict the distribution of the shallow gas reservoir are the effective tools for exploration of this kind of reservoir. Building geological model which is consistent with regional geology trends and applying the model in seismic forward modeling studies, summarizing the seismic reflection characteristics of the gas reservoirs, are the precondition for reservoir prediction in surface seismic exploration. Based on optimized attributes, PNN is used to integrate multi seismic attributes with well logging sonic data, then the distribution characteristics of the sonic parameters in 3d space can be predicted. On the one hand the workflow above avoid the information sidedness caused when only using single seismic attribute, on the other hand it realizes the reasonable extension of sonic parameters in the zone of interests over the 3d space, therefore the workflow provides a fast and effective tool for gas reservoir prediction. The workflow was applied in Honggang oilfield, Jilin, the seismic forward modeling results show that strong trough reflection from the top of the gas reservoir and strong peak reflection from the bottom of the reservoir can be seen, the anomalies of seismic attributes on seismic sections and the reflection characteristics from forward modeling gas-bearing sandstone are similar, thus major control factor for HI3 gas reservoir was identified, and the highly productive gas wells were the a-wards for the integrated studies.关键词
老油田/浅层气藏/多属性分析/正演/概率神经网络/多尺度信息融合Key words
mature oilfield/ shallow gas reservoir/ multi-attribute analysis/ forward modeling/ probability neural networks/ multi-scale information integration分类
天文与地球科学引用本文复制引用
张宪国,林承焰,张涛,王永刚..井、震多尺度信息融合预测老油田浅层岩性气藏[J].石油地球物理勘探,2009,44(5):609-612,4.基金项目
本研究由国家863重点项目(编号2007AA060501)及国家自然科学基金(编号40872094)资助. （编号2007AA060501）
