钻前准确预测井壁稳定性是防止钻进过程中井壁失稳的有效手段。声波时差和地层密度是井壁稳定分析的两个关键参数。根据地震记录与地层声波时差及密度之间存在的非线性关系,提出了利用地震属性钻前预测井壁稳定性的方法。从井旁地震记录中提取地震属性,通过RBF神经网络在已钻井段的地震属性与声波时差及密度之间建立起映射模型,以此为基础预测待钻地层的声波时差和密度。运用预测结果结合井壁稳定力学模型,确定岩石力学参数和地应力状态,计算井壁坍塌压力和破裂压力,确定安全钻井液密度窗口,实现钻前井壁稳定预测。该方法在塔里木油田的应用中取得了良好的预测效果。
Predicting borehole stability accurately before drilling is an effective way to control borehole instability in drilling. The interval transit time and density of formation are key parameters in borehole stability analysis. The method of predicting borehole stability before drilling by using seismic attributes is presented based on the nonlinear relationship between seismic data and interval transit time and density. The seismic at tributes are extracted from borehole-side seismic trace firstly. Through radial basis function neural network the mapping model is constructed between seismic attributes and interval transit time and density in drilled forma tion, based on which the interval transit time and density under borehole bottom is predicted. The prediction results and the borehole stability mechanics model are used to determine rock mechanical parameters and in-situ stress state, and to calculate the collapse pressure and fracture pressure of the borehole, thus obtaining the safe drilling fluid density window to achieve the prediction of borehole stability before drilling. This method has yielded excellent predicting results in practical application.