基于多智能体系统的钻井异常情况诊断技术  

DIAGNOSTIC TECHNOLOGY FOR DRILLING ACCIDENTS BASED ON MULTI-AGENT SYSTEM

作　　者：郭建明[1] 

机构地区：[1]西安石油大学石油工程学院

出　　处：《天然气工业》2008年第2期79-82,共4页Natural Gas Industry

基　　金：国家自然科学基金重大项目(编号:50234030);国家自然科学基金项目(编号:50474041);陕西省教育厅专项科研计划项目(编号:06JK205)的资助

摘　　要：在钻井作业过程中,现有的钻井事故与复杂情况的诊断与处理技术不能满足现场实际的需要,往往难于避免无法预料的钻井异常情况;研究领域中提出了数学模型方法,人工智能(AI)中的专家系统(ES),以及基于模糊的诊断方法、基于神经网络的诊断方法、概率逻辑推理诊断方法、基于遗传算法的诊断方法、基于灰色理论的诊断方法等以及各种方法相结合的综合诊断方法。这些方法各有优缺点和使用范围,也没有得到很好的应用。为此,提出基于多智能体系统的钻井异常情况诊断方案,用以充分应用钻井现场的实时动态钻井参数,使钻井异常情况的诊断与处理具体化、简单化、自动化,并提高钻井事故与复杂情况诊断与处理的实用性。The currently-used diagnostic methods and countermeasures can not solve so many problems that various unpredictable accidents easily happen in the course of drilling. Even though in this study area many diagnostic technologies have been put forward such as numerical model method, expert system in artificial intelligence, diagnostic method based on obscurity, diagnostic method based on logic and reasoning of probability, diagnostic method based on genetic algorithm, diagnostic method based on grey theory and comprehensive diagnostic method, the results showed that all those above-mentioned methods were not practical in actual use because of their own agreements and disagreements in various cases. Therefore, this paper put forward a diagnostic method for drilling accidents based on multi-agent system to estimate the real time dynamic drilling parameters in the field, which enables the diagnose and treatment of drilling accidents to be concrete, simple, automatic, and practical in use.

关 键 词：钻井工程 异常情况诊断 多智能体系统 知识库系统 实时预测 

分 类 号：TE28[石油与天然气工程—油气井工程] TP18[自动化与计算机技术—控制理论与控制工程]