基于贝叶斯估计的叠前反演技术在平湖油气田的应用

The application of pre-stack inversion technique based on Bayesian estimation in Pinghu oilfield

利用地震振幅信息进行油气检测一直是地球物理工作者研究的热点,其中AVO技术由于其较高的成功率一直为勘探工作者所推崇。但AVO属性也包含了许多不确定性,即多种岩石可以引起任何类型的AVO响应。贝叶斯估计的叠前反演技术可以定量评价AVO属性的可靠性。它以已知井的测井曲线和AVO属性为输入条件,最后可得到目标区油、气、水的平面概率分布图,以此来减少AVO属性的不确定性。该技术在平湖油气田中山亭构造得到了成功的应用,检测结果与平湖9井吻合,根据叠前反演结果优选了平湖10井,该井主要目的层P8层钻遇油气层9.8m/2层,与预测效果相符。

Geophysicists use seismic amplitude information to detect oil and gas,and AVO technique is the most important method because of its high success rate.However,AVO attributes contain many uncertainties.Many kinds of rocks can give rise to any AVO response.The pre-stack inversion technique based on Bayesian estimation can evaluate quantitatively the reliability of the AVO attributes.It takes fixed wells logs and AVO attributes as input parameters,and gains the probability distribution maps of oil,gas and water in target area.This way decreased the prediction uncertainty of AVO attributes.This technique was used in Pinghu oilfield successfully that is matched by the Pinghu 9 well,and validated by the prediction of Pinghu 10 well with oil layers of 9.8m/3 layers in the target interval P8.