基于贝叶斯网络的钻井作业风险评价模型的研究与实现

发布时间：2018-04-29 13:40

本文选题：钻井作业 + 风险评价　；参考：《西南石油大学》2015年硕士论文

【摘要】：在石油行业,钻井作业是一项投资大,风险高的工程。它的特点是多工种、多工序、连续作业、环节多、规模大、技术复杂。本文主要从钻井作业现场人的不安全行为与物的不安全状态两方面来分析其中的不安全因素。钻井作业是石油勘探与开发中的一个十分重要的环节与手段。近些年来石油勘探开发规模的不断壮大,整个国家对石油的需求量增加,从而对钻井作业的安全性要求逐渐提高。为了保证钻井作业的安全,对其进行风险评价势在必行。相比传统的风险评价方法,贝叶斯网络在坚实的数学基础上克服了多态性、非独立性、时序性等问题,通过概率图形的方式直观的表示元素、节点之间的依赖关系,具有双向推理的性能,从而可以进行前向预测分析以及后向的诊断分析。因而,本文进行了贝叶斯网络在钻井作业风险评价中的研究,主要包括以下几个方面： (1)了解国内外有关风险评价的研究现状,对现有的风险评价方法进行比较。 (2)识别钻井作业的风险和确定风险评价标准。钻井作业的风险包括人的不安全行为和物的不安全状态等影响因素。并根据钻井安全等相关研究确定了风险评价标准。 (3)采用贝叶斯网络分析法对钻井作业的风险隐患进行分析；在钻井作业风险评价模型的结构学习基础上,构建出风险评价的贝叶斯网络图；在钻井作业风险评价模型的概率预测部分,使用Matlab计算出某个时间段内钻井作业现场的风险概率,根据风险评价标准对这个时间段的钻井作业进行风险级别的归类。 (4)综合运用贝叶斯网络的结构学习和概率预测设计了贝叶斯网络钻井作业风险评价模型。通过现有的安全评价系统中的有关数据对贝叶斯网络风险评价模型进行验证。最后本文采用C/S结构和.NET开发环境,在Windows环境下以Visual Studio2010为开发平台,使用C#语言和Sql Server数据库,对钻井作业风险评价模块进行设计与实现。
[Abstract]:In the petroleum industry, drilling operation is a project with high investment and high risk. Its characteristics are multi work, multi process, continuous operation, many links, large scale, and complex technology. This paper mainly analyzes the unsafe factors in the two aspects of the unsafe behavior of the people in the drilling operation and the unsafe state of the objects. In the development of a very important link and means. In recent years, the scale of oil exploration and development is growing, the demand for oil is increasing in the whole country, thus the requirements for the safety of drilling operations are increasing gradually. In order to ensure the safety of drilling operations, it is imperative to evaluate the risk of the drilling operation.
Compared with the traditional method of risk assessment, Bayesian networks overcome the problems of polymorphism, non independence, timing, etc. on the basis of solid mathematics, which are intuitively expressed by the pattern of probability, and the dependence between nodes has the performance of bidirectional reasoning, so that the forward prediction analysis and the backward diagnosis can be analyzed. In this paper, the research of Bayesian network in risk assessment of drilling operation is carried out, mainly including the following aspects:
(1) to understand the research status of risk assessment both at home and abroad, and compare the existing risk assessment methods.
(2) identifying the risk of drilling operations and determining the criteria for risk assessment. The risk of drilling operations includes factors such as unsafe behavior of human beings and the unsafe state of objects, and the risk assessment standards are determined according to the related research of drilling safety.
(3) the Bayes network analysis method is used to analyze the risk of drilling operation, and on the basis of the structure learning of the drilling operation risk assessment model, the Bayesian network diagram of risk evaluation is built. In the probability prediction part of the drilling operation risk assessment model, the wind in a certain time period is calculated by using Matlab. Risk probability is classified according to the risk assessment standard for drilling operations in this time period.
(4) the Bayesian network drilling operation risk evaluation model is designed by using Bayesian network structure learning and probability prediction, and the Bayesian network risk evaluation model is verified by the relevant data in the existing security evaluation system.
At last, this paper adopts C/S structure and.NET development environment, and uses Visual Studio2010 as the development platform under Windows environment, uses C# language and Sql Server database to design and implement the drilling operation risk evaluation module.

【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】：TE28;TP18

【参考文献】