基于BP神经网络和D-S证据理论的损伤识别方法

Damage Identification Method Based on Back-PropagationNeural Network and D-S Evidential Theory

目的有效利用结构健康监测系统中的多源传感器数据信息,进而提高复杂结构健康状况的正确诊断率.方法将BP神经网络与数据融合理论中的证据理论有机融合,提出一种决策级数据融合损伤识别新方法.为了验证所提方法的有效性,用1栋7自由度剪切型建筑模型的6种损伤进行了检验.结果研究发现,将BP网络和D-S证据理论相结合的综合诊断模型,可以有效地提高一些损伤模式的诊断率,具有良好的适应性.结论笔者所提方法优于单一信息建立模型的识别能力,表明它具有较好的容错性和识别精度,用于健康监测和损伤检测是可行的、有效的.

In order to make full use of multi - sensor data or information from multi - resources and improve the diagnosis accuracy in a structural health monitoring system, computational intelligence technology, e. g. neural network, and multi- sensor data fusion theory are employed to detect structural damage in this paper. A decision- level data fusion damage identification method, which combines BP neural network with evidence theory, is proposed. To validate the efficiency of the proposed method, six simulation damage patterns from a 7 - DOF shear- type building model are identified. The results show that the proposed method can not only improve the identification accuracy but also has good adaptive capacity. Furthermore, this proposed method is feasible and effective in damage identification and structural health monitoring.