克里金
计算机科学
钥匙(锁)
随机性
回归
回归分析
数据挖掘
机器学习
统计
数学
计算机安全
出处
期刊:Journal of Oil and Gas Technology
日期:2016-01-01
卷期号:38 (04): 80-89
被引量:9
标识
DOI:10.12677/jogt.2016.384040
摘要
科学高效的造斜率预测方法是优选钻井参数、提高井眼轨迹控制精度和效率的关键技术,对复杂定向井的高效、低成本开发具有重要意义。造斜率预测受多种因素的耦合影响,具有模糊性、随机性、非线性等特点,用数学力学模型难以描述。提出将Kriging代理模型应用到导向工具的造斜率预测中。从造斜率影响因素入手,重点分析了导向工具结构及造斜原理的差异对造斜率的影响;将现有的造斜率预测方法归纳为几何预测法、力学预测法和回归分析预测法3大类进行了综述和对比分析,指出了现有方法的局限性;从回归分析预测法的角度,提出了一种基于Kriging代理模型的造斜率预测新方法,分析了该方法的科学性和高效性,提出了使用该方法的具体预测步骤和应注意的关键问题,为造斜率预测提供了新途径。 The scientific and efficient prediction method of build-up rates was the key technology to optimize drilling parameters and to improve wellbore trajectory control accuracy and efficiency. It was of great significance for the efficient and low-cost development of complex directional wells. The prediction of build-up rate, which had the characteristics of fuzziness, randomness and non- linearity, was affected by various coupling factors. As a result, it was difficult to describe with ma-thematical and mechanical models, and it was proposed that Kriging surrogate model was applied to predict the build-up rate of steering tools. Starting from the analysis of the factors affecting the build-up rate, the influence of difference between the structure and build-up principles of the steering tools was emphatically analyzed. The current methods were summarized into three types of geometric prediction, mechanic prediction and regression analysis and prediction. By compar-ison between them, their limitations were pointed out. From the aspect of regression analysis and prediction, a novel method based on the build-up rate of Kriging surrogate model is proposed; the scientificalness and high efficiency of the method are analyzed; the concrete procedures for pre-diction and key issues for attention are proposed; it provides new ways for predicting the build-up rate.
科研通智能强力驱动
Strongly Powered by AbleSci AI