随钻方位伽马测井成像畸变机理与地层产状计算方法研究

Using the azimuthal gamma logging data while drilling, the borehole imaging can be get for formation evaluation and geo-steering, the key problem is to determine the actual formation interface and attitude in order to track the drilling target position and adjust the well path in real time, which is still monopolized by foreign company..In high angle or horizontal well the deviation between the axis of logging tools and the well axis makes the visual detection and the actual shape of the borehole do not coincide, the amplitude of logging signal in different circumferential directions and the shape of formation interface will be distorted. In addition, the roundness of borehole is worse than that of vertical well, and the logging data are sparse, which makes it difficult to recognize the formation boundary with borehole imaging..In view of the above core problems, closely combined with the needs of geo-steering of azimuthal nature gamma-ray LWD tool，this research will reveal the mechanism and law of eccentric influence on the amplitude and investigation depth of gamma logging nearby formation surface by means of physical experiments and numerical forward modeling，to analyze the distortion characteristics of borehole imaging amplitude and formation interface shape, to obtain the influence relationship of instrument eccentricity on the calculation of formation dip angle. .Based on this, to study the high-fidelity interpolation recovery method of the sparse logging data, obtain the global optimization fitting and optimal level set segmentation method to detect the fuzzy interface of borehole image precisely, and gives the corresponding formation attitude calculation method..This study will solve the problems such as the quality of azimuthal gamma LWD borehole imaging is poor, the automatic formation interface identification rate is lower, and the attitude quantitative calculation error is larger, and it will lay the theoretical foundation for forming the azimuthal LWD Geo-steering technology with independent intellectual property rights.

利用随钻方位伽马测井资料井壁成像可进行地层评价、实施地质导向，其关键是准确判定地层界面与产状以跟踪靶点位置，目前为国外垄断。大斜度、水平井中随钻测井仪器不居中使得视探测形状与实际井眼形状不吻合、井壁测井信号幅度与地层界面形态产生畸变，此外井眼较直井圆形度差、随钻测井数据稀疏，导致了井壁成像地层边界难以识别。本项目针对该问题、结合国内随钻方位伽马测井地质导向需要，以物理实验和数值正演模拟为手段、揭示近地层界面随钻方位伽马测井信号幅度与探测深度受偏心影响的机理和规律、分析井壁成像幅度与地层界面形态畸变特征，得出仪器偏心对地层倾角计算的影响关系；以此为基础，研究井壁稀疏测井数据的高保真插值复原成像方法，基于井壁成像全局优化模糊区域水平集最优分割地层界面追踪思路、得出高精度地层界面自动识别和地层产状计算方法；为形成自主知识产权随钻方位测井地质导向技术奠定方法理论基础。

随着我国对非常规、深层油气勘探开发的需求，随钻测井地质导向技术成为提高油气层钻遇和采收率的关键技术之一。利用随钻方位伽马测井资料通过井壁成像可进行地层评价、实施地质导向，其核心是准确判定地层界面与产状、以跟踪靶点位置，国外技术总体处于引领地位。.项目结合我国随钻方位测井仪器研发现状和地质导向需求，针对近地层界面随钻方位伽马畸变机理及规律不清问题，发明了快速正演算法、设计了物理模拟实验装置并进行了研制，通过正演揭示了井斜、仪器偏心对随钻方位伽马井壁成像的影响特征及规律，给出了相应矫正方法；针对实时采样稀疏、成像质量差问题，综合地质统计和机器学习给出了符合井壁成像空间结构特征的实时插值重构方法，提高了井壁成像质量；针对地层界面井壁成像自动识别问题，给出了最优动态阈值与最小水平集图像分割方法，并通过图像边界多重链码向量追踪、地层界面遴选与分类等算法实现了地层边界的自动划分判识；针对地层产状计算问题，通过对霍夫变换、正弦拟合、鲁棒最小二乘最优拟合算法的优化比较，实现了不同井壁成像模式地层界面的法矢量求解和地层产状定量计算。为利用方位伽马成像自动判识地层界面和定量计算地层产状奠定了理论方法基础。.为形成自主产权技术，发明了二维图元窗口裁剪绘图方法，并采用双缓存与GPU加速技术开发了基于随钻方位伽马测井成像的井筒信息一体化地质导向工业软件，不同油田多次应用测试、目的层钻遇率均在90%以上；此外成像处理与产状计算模块被集成到中石化工程院相关软件平台并推广应用。.通过总结部分研究成果，发表期刊论文8篇、SCI检索5篇，会议论文2篇；参加境外国际学术交流会1次，国内会议分组报告1次、特邀报告1次；申请国内发明专利4项，授权中国发明专利2项、WIPO 1项，登记软件著作权5件。培养毕业硕士12名、在读硕士6人，晋升副教授1名。组织召开国际学术研讨会4次，邀请国外高校与公司专家开展系列学术讲座4人次。