全文下载：2013040427
陈莹莹1,2，徐余法1,2，陈国初2
(1.华东理工大学 信息科学与工程学院，上海 200237；
2.上海电机学院 电气学院，上海 200240)
摘要： 针对D_S证据理论直接处理高冲突证据合成时会出现诊断结果不合理的问题，提出了一种相对加权证据合成法。在多个证据融合时，该方法综合利用证据的均值距离和冲突强度表示证据冲突，并根据各证据与其它证据的冲突程度确定其权重系数，在此基础之上采用相对加权的方法对证据的基本概率指派函数修正，然后用D_S规则合成。将所提出的方法应用到风力发电机组的故障诊断中，结果表明了所提方法对高冲突证据融合的有效性。
关键词: D_S证据理论；冲突强度；故障诊断；证据均值距离
中图分类号： O 359文献标志码： A
Application of an Improved D_S Evidence Theory Set to the Fault Diagnosis of Wind Turbine
CHEN Ying-ying1,2, XU Yu-fa1,2, CHEN Guo-chu2
（1.SchoolofInformation Scienceand Engineering,EastChinaUniversityof Science and Technology,Shanghai200237,China;
2.School of Electrical Engineering, Shanghai DianJi University, Shanghai 200240, China）
Abstract: D_S evidence theory involves unreasonable diagnosis when evidence highly conflicts. A new approach of combination of relatively weighted belief functions is proposed to solve the problem . If many pieces of evidences are to be combined, the amount of conflict between evidences is firstly evaluated by both evidence distance and conflict intensity, and every piece of evidence is given a weight coeffficient according to its amount of conflict with the others, Then by using the relative weighting of evidence，the mass function could be obtained. The basic probability assignment function is synthesized by using D_S rules. Then the proposed method on the diagnosis of wind turbine is applied. The experimental results show the effectiveness of the proposed method for high-conflict evidence.
Key words: D_S evidence theory; conflict intensity; fault diagnosis; evidence average distance
收稿日期：2012-08-24
基金项目： 国家自然科学基金项目（60772006，70773041）、上海市教委重点学科资助项目（J51901）、上海市教委科研创新重点项目（09ZZ211）.
作者简介： 陈莹莹（1988-）,女, 硕士研究生.
