水平井地质导向综合解释方法研究

发布时间：2018-03-18 12:27

本文选题：水平井　切入点：地质导向　出处：《西南石油大学》2017年硕士论文　论文类型：学位论文

【摘要】：水平井地质导向是在地质研究的基础上首先建立钻前地质导向模型,根据地质导向模型进行钻进,但是由于地层复杂多变化,需要在钻进时综合利用随钻测井、随钻测量和录井资料实时验证构造和储集层信息,不断修正地质导向模型,对着陆点、靶点深度进行实时预测,及时调整井眼轨迹,提高储层钻遇率。因此本文综合利用随钻测井和录井对水平井地质导向整个过程的解释方法进行综合研究。首先是对钻前储层构造模型进行研究。根据邻井资料结合曲线拟合的数学方法建立二维储层边界预测模型,它是水平位移与储层边界垂深的二元函数。此模型适用与起伏范围较大的地层,并且也可以适用于水平井穿过断层的情况。利用储层界面预测模型可以用于判断储层上倾下倾、计算储层的视倾角并判断钻头与储层边界的相对位置。在钻达储层前根据实时的数据对模型进行不断的修正以提高模型的精度。在水平井地质导向过程中造斜段的安全准确着陆是水平井钻井的关键一步,因此预测着陆点的位置是本部分的关键。本文主要研究了钻前预测着陆点方法和钻中预测着陆点方法。其中钻中预测着陆点的关键是实时识别标志层与邻井进行对比,因此进行了基于CBR(案例推理)理论进行人工智能识别标志层的研究,即通过分析当前井与邻井的标志层特征参数的相似度大小来判断是否钻达标志层。同时使用了同样的方法智能分析判断是否着陆。并且在着陆前对地质模型进行不断更新。在水平段钻进过程中为了实时的判断钻头在储层中的位置,考虑了随钻测量工具到钻头存在一定的距离,建立了基于BP神经网络的随钻预测模型,即利用实时的录井数据预测钻头处的随钻测井数据,并对预测的数据进行了校正,提高了准确性,保证最大程度的在储层中钻进。利用已钻井段的实时数据对储层地质模型进行更新,保证了后续的钻遇率。本文采用微软公司的Visual Studio 2013编程平台与SQL Server 2008数据库管理平台开发了水平井地质导向综合分析软件。该软件能够智能的识别标志层、智能判断是否着陆以及进行随钻预测,并结合钻井现场收集的数据验证了其可行性。
[Abstract]:The geological guidance of horizontal wells is based on the geological research. Firstly, the geological guidance model is established before drilling and drilling is carried out according to the geological guidance model. However, due to the complexity and variety of formation, it is necessary to make comprehensive use of logging while drilling. While drilling and logging data, the structure and reservoir information are verified in real time, the geological guidance model is constantly revised, the landing point and target depth are predicted in real time, and the borehole trajectory is adjusted in time. Therefore, this paper comprehensively uses logging while drilling and logging to comprehensively study the interpretation method of the whole process of geological guidance of horizontal wells. Firstly, the structural model of reservoir before drilling is studied. According to the data of adjacent wells, the structure of reservoir is studied. Combined with the mathematical method of curve fitting, a two-dimensional reservoir boundary prediction model is established. It is a binary function of horizontal displacement and vertical depth of reservoir boundary. This model is suitable for strata with large range of fluctuation, and can also be used for horizontal wells crossing faults. The prediction model of reservoir interface can be used to judge the updip and downdip of reservoir. Calculate the apparent dip angle of the reservoir and judge the relative position of the bit and the boundary of the reservoir. Before drilling to the reservoir, continuously revise the model according to the real-time data to improve the accuracy of the model. A safe and accurate landing is a key step in horizontal well drilling. Therefore, predicting the location of landing point is the key of this part. This paper mainly studies the method of predicting landing point before drilling and the method of predicting landing point in drilling. The key to predict landing point in drilling is to identify the marker layer in real time and compare it with the adjacent well. Therefore, based on CBR (case-based reasoning) theory, artificial intelligence recognition marker layer is studied. That is to say, by analyzing the similarity between the characteristic parameters of the current well and the adjacent well, we can judge whether the marker layer is drilled or not. At the same time, we use the same method to intelligently analyze and judge whether to land or not, and carry on the geological model before landing. In order to determine the position of the bit in the reservoir in real time during horizontal section drilling, Considering the distance between the measurement tool while drilling and the bit, a prediction model based on BP neural network is established, that is, using the real-time logging data to predict the log data while drilling at the bit, and the predicted data are corrected. Improved accuracy, guaranteed maximum drilling in the reservoir, and updated the reservoir geological model using real-time data from the drilled section, In this paper, a comprehensive analysis software of horizontal well geological guidance is developed by using Microsoft's Visual Studio 2013 programming platform and SQL Server 2008 database management platform. The software can identify the marker layer intelligently. Intelligent judgment of landing and prediction while drilling is carried out, and the feasibility is verified by combining the data collected from drilling site.
【学位授予单位】：西南石油大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TE243

【相似文献】