基于多证据融合决策的间歇过程测量数据异常检测方法

发布时间：2018-04-01 17:31

本文选题：间歇过程　切入点：D-S证据理论　出处：《化工学报》2017年08期

【摘要】：间歇过程测量数据的高维、非线性、非高斯分布特征直接影响过程测量数据异常检测的准确性,为了融合多源数据异常检测信息,提升间歇过程测量数据异常检测精度,提出了一种基于多证据融合决策的间歇过程测量数据异常检测方法,该方法通过引入证据理论(Dempster-Shafer,D-S),采用主焦元判别伪证据和重新计算证据权重改进冲突证据处理方法,减小了冲突证据对多证据融合决策结果的影响,提高了间歇过程测量数据异常检测的准确率。构建了基于多证据融合的测量数据异常检测模型并将其应用到间歇过程测量数据异常检测决策判决中。实验结果表明,该方法能够融合多证据信息,有效地处理冲突证据,实现了间歇过程测量数据异常检测,降低了误检和漏检率。
[Abstract]:The high dimensional, nonlinear and non-#china_person0# distribution characteristics of batch process measurement data directly affect the accuracy of process measurement data anomaly detection. In order to fuse multi-source data anomaly detection information, improve the accuracy of intermittent process measurement data anomaly detection. An anomaly detection method for intermittent process measurement data based on multi-evidence fusion decision is proposed in this paper. By introducing evidence theory into Dempster-Shafern D-Sine, the main focus element is used to distinguish false evidence and the weight of evidence is recalculated to improve the method of dealing with conflicting evidence. It reduces the impact of conflict evidence on the decision results of multi-evidence fusion. The anomaly detection model based on multi-evidence fusion is constructed and applied to the decision decision of interval process measurement data anomaly detection. The experimental results show that, This method can fuse multi-evidence information, deal with conflict evidence effectively, realize abnormal detection of intermittent process measurement data, and reduce the rate of false detection and missed detection.
【作者单位】：北京化工大学信息科学与技术学院;
【基金】：国家自然科学基金项目(61240047) 北京市自然科学基金项目(4152041)~~
【分类号】：TQ050.7

【参考文献】