油田采出水处理方案分布式专家系统研究  

Distributed expert system of oilfield produced water process program

作　　者：陈毅红[1] 周永宏[2] 李高平[1] 

机构地区：[1]西南民族大学计算机科学与技术学院,四川成都610041 [2]西华师范大学物理与电子信息学院,四川南充637000

出　　处：《计算机工程与设计》2012年第5期1978-1982,共5页Computer Engineering and Design

基　　金：国家"十五"科技攻关基金项目(2004BA616A-11-01);中央高校专项科研基金项目(11NPT02);四川省科技厅应用基础研究计划基金项目(2009JY0164)

摘　　要：针对现有油田采出水处理专家系统中存在的水处理流程、设备、水处理剂选择三者合一,导致不利于产生性价比高的方案以及系统工作自动化低的问题,提出将3种选择推理任务分开的协同式多专家系统。设计了基于叠代思想协同算法来协调3个子推理任务,提高了设计方案性能价格比及专家系统运行自动化水平。流程选择方面,提出基于不确定性的启发式全局优先搜索策略算法,并设计了该算法的启发函数来提高算法的效率。为设备选择和水处理剂选择设计了框架知识库和实用匹配算法,满足所选择设备和水处理剂的低成本要求。结果分析表明,该系统较之现有水处理专家系统在算法运行效率、结果有效性方面得到了明显提高。An synergistic multi-experts system with three divided inference tasks including process selection,device selection and water treatment material selection is proposed to overcome the shortcoming of exist experts system based on triple-in-one technology of above three selections,which lead to a low performance price ratio scheme and low automation level of expert system.The synergistic algorithm which manage three inference tasks is designed to improve the ratio and the automation based on iterative idea.As regards of process selection,the global enlighten priority search algorithm with uncertainty is proposed,and enlighten function is designed to improve the algorithm efficiency.Forthmore,the frame knowledge bases and match algorithm are designed for device selection and material selection and they ensure low cost.Results show that efficiency and effectiveness of the algorithm is improved obviously.

关 键 词：知识库 水处理不确定 推理 评价 协同 

分 类 号：TP182[自动化与计算机技术—控制理论与控制工程]