基于融合推理模型的钻井液优化设计系统研究

发布时间：2018-05-17 05:31

本文选题：钻井液 + 融合推理　；参考：《西南石油大学》2015年硕士论文

【摘要】：钻井液设计作为钻井工程设计的重要内容,也是钻井液现场施工的重要理论依据。就目前调研看来,我国大多数油田的钻井液设计主要还是依靠专业设计人员通过对油井相关数据的分析,在进行大量实验的基础上,综合可参考的历史资料以及设计者积累的经验来完成的。这种传统的设计方式存在设计结果因人而异,设计书格式不够统一,难以与国际接轨等诸多缺点。因此,为了提高钻井液设计质量,利用计算机对设计进行辅助,将人工智能系统引入到设计中去是解决传统钻井液设计方式上这些不足的较为普遍方法。同时随着油气勘探开发技术的飞速发展和需求量的不断攀升,现代钻井技术对钻井液提出了更新更高的要求,各种新型钻井液技术也不断得到应用和发展,在追求高效低成本的今天,智能化的钻井液设计及管理技术也得到了更多的关注,因而开发更加实用于现代钻井液设计及钻井液数据管理的软件已十分必要。基于以上认识,本文研究了基于专家规则推理(Rule-Based Resoning, RBR)技术、基于范例推理(Case-Based Resoning, CBR)技术以及支持向量机(Support Vector Machine,SVM)技术在钻井液设计中的应用方法,并建立了这三种推理技术的融合推理模型,以此模型开发了钻井液优化设计系统,用于对钻井液的设计过程加以辅助,提高钻井液设计效率和设计质量。本系统将专家规则推理与范例推理以及支持向量机的回归机融合在一次钻井液设计推理过程中,避免了单一推理模型由于推理过程过于简单,不符合专家思维以及推理结果不准确而引起的系统实用性差等问题。本文完成的主要研究工作和取得的成果如下： (1)调研了传统钻井液设计的一般过程及工艺原理,总结了一套类似于专家思维的钻井液计算机设计逻辑思路。 (2)收集了多条钻井液专家经验,并按一定规则建立了规则库；收集了各大油田钻井液技术资料,并按一定原则建立了钻井液范例库；通过收集的钻井液配方,以室内实验数据为基础,结合支持向量机建立了预测钻井液配方的模型库。 (3)研究了基于规则推理技术、基于范例推理技术和支持向量机的基本原理和它们在钻井液设计中的应用方法,以及三种技术的融合应用方法。 (4)根据钻井液设计的工艺特点,以Visual Studio2010为开发平台,vb.net和c++为设计语言开发了钻井液优化设计系统。将系统的设计结果与现场5口井的应用实例进行对比,结果发现该系统较好的完成了钻井液的体系选择和性能参数设计。
[Abstract]:Drilling fluid design, as an important content of drilling engineering design, is also an important theoretical basis for drilling fluid field construction. According to the current investigation, the drilling fluid design of most oilfields in China mainly depends on the professional designers through the analysis of the relevant data of the wells, on the basis of a large number of experiments. Comprehensive reference of historical materials and designers accumulated experience to complete. The traditional design method has many shortcomings, such as different design results, not uniform design format, difficult to connect with the international standards, and so on. Therefore, in order to improve the design quality of drilling fluid, it is a common method to solve these problems in the traditional drilling fluid design mode by using computer to assist the design and to introduce artificial intelligence system into the design. At the same time, with the rapid development of oil and gas exploration and development technology and the increasing demand, modern drilling technology has put forward higher requirements for drilling fluid, and various new drilling fluid technologies have been continuously applied and developed. Nowadays, with the pursuit of high efficiency and low cost, more attention has been paid to the intelligent drilling fluid design and management technology, so it is necessary to develop more practical software for modern drilling fluid design and drilling fluid data management. Based on the above understanding, this paper studies the application methods of Rule-Based Resoning-based (RBR-based), Case-Based Resoning-based (CBR-based) and support Vector Machine (SVM) in drilling fluid design based on expert rule reasoning (RBR), Case-Based reasoning (CBR) and support Vector Machine (SVM) in drilling fluid design. The fusion reasoning model of these three reasoning technologies is established, and a drilling fluid optimization design system is developed based on this model, which is used to assist the drilling fluid design process and improve the design efficiency and design quality of drilling fluid. In this system, expert rule reasoning, case reasoning and regression machine of support vector machine are combined in the primary drilling fluid design reasoning process, and the single inference model is avoided because the reasoning process is too simple. Some problems such as poor system practicability caused by inaccuracy of reasoning results and so on are not in accordance with expert thinking. The main research work and results achieved in this paper are as follows: 1) the general process and process principle of traditional drilling fluid design are investigated, and a set of logical thinking of computer design of drilling fluid similar to expert thinking is summarized. (2) the expert experience of several drilling fluids has been collected, and the rule base has been established according to certain rules; the technical data of drilling fluid in major oilfields have been collected, and the drilling fluid sample bank has been established according to certain principles. Based on laboratory experimental data and support vector machine (SVM), a model base for predicting drilling fluid formulation was established. (3) the basic principles of rule-based reasoning, case-based reasoning and support vector machine, their application in drilling fluid design, and the fusion of the three techniques are studied. According to the technological characteristics of drilling fluid design, a drilling fluid optimization design system is developed with Visual Studio2010 as the development platform and c as the design language. The design results of the system are compared with the application examples of 5 wells in the field. The results show that the system selection and performance parameter design of the drilling fluid are well completed.
【学位授予单位】：西南石油大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TE254

【参考文献】