基于贝叶斯网络的结构健康评估信息融合方法

On Information Fusion for Structural Soundness Appraisal in Bayesian Network

为了充分利用大型结构健康监测系统中来自不同时间与空间的多传感器信息资源,获得被测对象的一致性决策和估计任务,进而提高确诊率,介绍了从多传感器数据融合的概念、基本原理出发,提出的一种基于贝叶斯网络数据融合技术的结构健康监测方法。重点叙述了用于结构健康检测的朴素贝叶斯网络和扩展的朴素贝叶斯网络结构构建,以及网络节点概率的确定方法,并在项目中进行了试验。基于贝叶斯网络的结构健康评估方法有效地利用了各信息源之间的互补性,提高了健康评估的准确率、可靠性和稳健性。

For full use of the information source from multiple sensors at different times and spaces in the large structure soundness monitoring system and to obtain the consistent decision and expected task of the object to be detected for higher accuracy of diagnosis, the concept of data fusion from multiple sensors and the general mechanism are introduced, including a proposed method for structural soundness monitoring based on the data fusion technique in Bayesian Network, with detailed description of the naive Bayesian Network and its extended structural formation and the way to determine the rate of network nodes, which are tested in projects. This way of structural soundness appraisal through the network is of higher accuracy, reliability and stability with effective mutual compensation between data sources.