【作者】 毕雪亮；

【导师】 闫铁；

【作者基本信息】 大庆石油学院 ， 油气井工程， 2007， 博士

【摘要】 “最优化”是每一个工程或产品设计者所追求的目标。钻井是一个复杂的系统工程,在其过程中产生大量的不确定的、无序的信息。钻井工程的优化对提高生产率和经济效益具有重要的意义,是实现钻井工程科学化的必要环节。由于油气井工程的复杂性和不确定性,优化问题本质上呈现多目标性、影响参数复杂多变,迫切需要引入新的科学的优化方法及理论来解决。为解决工程实际中的优化问题,克服传统优化方法的局限性,本文提出基于知识发现的优化方法。其基本思想是:基于知识发现的数据、信息到知识的挖掘过程,是一种特定的模拟人类思维推理过程的概念到思维的知识学习模型。它可以在没有形成清晰、全面的数学模型,没有确切的量度标准的情况下,根据已有的经验、事实、观测数据进行分析和简化,从而发现内在规律、建立所研究问题的知识模型,进行知识推理和判断,最终获得合理而可行的优化方法。通过全面系统的研究及应用,论文的研究内容及取得的创造性研究成果主要包括:(1)对钻井过程中存在的不确定性问题研究。对钻井过程中产生的原始数据,应用不确定性理论中的方法进行分析总结,从钻井工程本身的基本特点出发,建立油气井工程数据信息的分析方法。(2)钻井工程信息数据库理论研究。结合钻井数据信息的不确定性特点,分析数据的处理模式和数据库系统的结构构建基础数据库,并对数据信息进行有效处理,为知识发现的各种算法、知识挖掘方法奠定基础。(3)人工神经网络优选钻头研究。在新型钻头使用效果及以往大庆深井所用钻头数据库的基础上,应用人工神经网络学习模型,利用自适应共振理论对大庆油田深井使用的钻头进行了优选。建立了钻头数据的知识发现分析模型。(4)基于支持向量机的钻具失效分析。在钻具失效统计数据库的基础上,结合支持向量机的统计学习模式的特点,分析大庆、海拉尔油田深井钻具失效基础数据,确定钻具失效的特点及原因,由此制定了预防钻具失效的措施。(5)井眼轨道的可视化理论及应用研究。在测斜信息数据库的基础上,应用Delphi与OpenGL软件工具实现井眼轨道显示和井眼轨道设计的可视化。由此建立了多障碍绕障设计模型。该方法为密井网井眼轨迹的三维描述、绕障井的井眼轨道设计及待钻井眼轨迹控制和轨道设计提供依据。本论文将知识发现体系的数据、信息、知识、智慧等部分紧密结合起来。从原始钻井工程数据到不确定性信息分析,及在信息基础上应用有效的挖掘方法提炼知识到最终决策用于生产实际中。本文的特色就是充分发挥了相应知识发现方法的特色,将其与信息属性紧密结合起来。这充分体现了知识发现理论在钻井工程中的重要性和巨大应用潜力。更多还原

【Abstract】 “Optimization”is the aim that every product designer pursues. Drilling is a complex system engineering, which produces plenty of uncertained random information. Optimizatioin of Dilling engineering is important to improve the efficiency and economic benefit. Due to the complexity and uncertainty of oil and gas drilling engineering, the optimization turns to be a multi-functional problem which has complex influence factors. New scientific optimization method and theory need to be designed to solve the problem.To solve the optimization problem in engineering field and get over the limitatioin of conventional method, an optimization method based on knowledge discovery is provided in this paper. It is a mining process from data and information to knowledge discovery. It is also a knowledge learning model specified in simulating human brain work process. Even without the mathematic model and standard measurement, knowledge discovery theory could help us to find the inherent law and build up the knowledge model with the analysis and simplification of the achieved experience, facts, and measuring data.Through systemic research and application, the content of the thesis and the creative serearch solution include:(1) Research on the uncertainty problem in drilling process. Analyse the orginal data obtained from drilling process using uncertainty theory. Build up the drilling data information analysis method.(2) Theoretical research on the drilling engineering information database. Based on the database information uncertainty, data processing model and structure of database system to build basic database is set up, which could process data information and establish a foundation of various arithmetic and knowledge discovery method.(3) Select drill bit on artificial neural network. Adaptive resonance theory is used in neural network learning model to optimize drill bit selection in deep well of Daqing Oilfield based on bit database. The knowledge discovery model of bit data processing is set up.(4) Drilling tool failure analysis based on support vector machine. Based on the drilling tool failure database and support vector machine study model, analyse Daqing, Hailaer oil field deep well drilling tool failure data and give a conclusion to its characteristic and reason, therefore a preventation measure could be established to direct the field work.(5) Visualization theory and application research on wellpath trajectory. Based on path survey database, a program was designed to implement wellpath trajectory and design visualizatioin using Delphi and OpenGL. The multi-obstacle path design model is set up.It provides the evidence to wellpath trajectory design, wellpath control and three dimensional wellpath trajectory descriptions.This paper unites data, information, knowledge, wisdom together in Knowledge Diskovery system. From original drilling data to uncertain information analysis, discovery knowledge based on information is used in field application. The feature of this paper is combining knowledge discovery with information property. It could contribute to wellpath trajectory desing and visualization theory, which shows the great importance and potential of knowledge discovery theory in drilling engineering.更多还原