【摘要】 常规井壁稳定预测方法应用于预探井时,由于资料缺乏及使用条件限制、计算参数误差、运算过程复杂等因素会严重影响其预测效果,针对这些问题提出了一种适用于预探井的随钻井壁稳定预测方法。通过考察波阻抗与孔隙压力、构造地应力、岩石强度之间的定量关系,建立包含波阻抗与井壁稳定力学参数的非线性模型,在此基础上利用神经网络算法识别波阻抗与地层三压力(孔隙压力、坍塌压力与破裂压力)之间的映射关系。通过神经网络分层建模及录井资料实时分析,实际钻进中利用地震反演波阻抗数据随钻预测钻头前方的井壁稳定性,及时优化钻井液密度以控制井壁稳定性。现场预探井的应用结果表明,新方法较之常规技术有效提高了适应性与运算速度,操作流程更为便捷,预测精度能满足工程需求。更多还原

【Abstract】 When the conventional method of predicting borehole stability is applied to the preliminary prospecting wells, the precision of prediction is often affected severely by the lack of data, limit of applying conditions, error of calculated parameters and complexity of operating courses. This paper presents a method of predicting borehole stability while drilling preliminary prospecting wells and establishes a nonlinear model including wave impedance and borehole stability mechanical parameters by investigating the quantitative relationships between wave impedance and pore pressure, in-situ stress as well as rock strength. Based on the nonlinear model, neural network algorithm is used to identify the relationship between the wave impedance and the three formation pressures(pore pressure, fracture pressure and collapse pressure). Through the establishment of the model by layered neural network and the timely analysis of logging data, the wellbore stability before the bit can be predicted while drilling by using the data of seismic wave impedance. Field application in preliminary prospecting wells shows that the new method has higher adaptability, faster calculation speed, and simpler operational procedure than conventional methods, and its prediction accuracy can meet the requirement of engineering.更多还原