Abstract: Undrilled formation pressure prediction is very complex. There are some methods available for predicting pressure,such as extrapolation by adjacent well,regional pressure analysis and seismic velocity spectrum analysis. But in oceanic area, some difficulties arise because of low well spacing density and limited information under serface. This paper introduced a new way to pressure prediction in undrilled formation by means of extrapolating sonic logging using artificial neural network while combining halfway log data with seismic data. This new technic and the procedure were presented in this paper. The extrapolating prediction for some oceanic wells in a area brought good result. It shows this research work is feasible and valuable.

Key words: log data, seismic data, undrilled formation, extrapolation, pressure prediction

摘要： 井底以下地层压力预测是一项十分复杂的研究课题。目前采用的方法主要有邻井外推、区域压力分析和地震速度谱分析等。但在海洋钻井的条件下,由于井网密度较稀,地下已知信息有限,给压力预测造成了很大困难。本文提出一种利用中途测井资料,结合地震信息,通过地震特征参数提取并用神经网络技术建立地震与测井之间的关系和外推井底以下声波曲线,进而预测未钻达地层压力的新方法。文中论述了实现这一新方法的技术和过程。经在某地区几口海洋钻井的外推预测实践,取得了良好的效果。

关键词: 测井, 地震, 井底以下, 外推, 压力预测