Liu Qiwei, Kong Fanmin, Chen Xiaolong, Wang Guangsheng, Li Kang
School of Information Science and Engineering, Shandong University, Qingdao, 266237, Shandong, China.
Sci Rep. 2024 Oct 23;14(1):25071. doi: 10.1038/s41598-024-76032-6.
With the diminishing availability of oil and gas resources, the Rotary Steerable System has become increasingly important. However, the vibrations and shocks during the drilling process pose challenges to the Measurement While Drilling. In recent years, the application of machine learning in the field of petroleum exploration has gradually expanded, especially in the estimation of geological parameters and lithology discrimination. However, there is still limited research on drilling tool attitude measurement. To address the above-mentioned issues, this study proposes a method that combines drilling tool attitude sensor data with artificial neural networks to improve the accuracy of dynamic inclination measurement. This method utilizes machine learning techniques, combining real-time z-axis acceleration signals and magnetic induction signals, and employs a deep learning model to invert the x and y-axis acceleration signals, thereby achieving high-precision measurement of dynamic inclination angles. Experimental results show that Long Short-Term Memory model, under simulated measurement conditions with different rotational speeds, yields dynamic inclination curve errors ranging from 0.4° to 0.7°, significantly reducing the errors compared to the original measurements. This method not only improves the accuracy of inclination angle measurement but also demonstrates strong adaptability to different rotational speeds, providing more accurate data support for drilling operations.
随着石油和天然气资源的日益减少，旋转导向系统变得越来越重要。然而，钻井过程中的振动和冲击对随钻测量构成了挑战。近年来，机器学习在石油勘探领域的应用逐渐扩大，特别是在地质参数估计和岩性判别方面。然而，在钻井工具姿态测量方面的研究仍然有限。为了解决上述问题，本研究提出了一种将钻井工具姿态传感器数据与人工神经网络相结合的方法，以提高动态倾角测量的精度。该方法利用机器学习技术，结合实时z轴加速度信号和磁感应信号，并采用深度学习模型对x和y轴加速度信号进行反演，从而实现动态倾角的高精度测量。实验结果表明，在不同转速的模拟测量条件下，长短期记忆模型产生的动态倾角曲线误差在0.4°至0.7°之间，与原始测量相比误差显著降低。该方法不仅提高了倾角测量的精度，而且对不同转速具有很强的适应性，为钻井作业提供了更准确的数据支持。
