论文部分内容阅读
为了有效地利用结构健康监测系统冗余、互补的信息进行结构健康状况评估,该文提出了一种将粗集、数据融合和概率神经网络(PNN)有机地结合在一起的损伤识别新方法。它先用粗集进行属性约简来降低数据的空间维数,然后运用PNN进行融合计算来处理冗余、不确定信息,最后进行融合决策和损伤识别。在粗集属性约简过程中,提出了运用K-均值聚类的方法进行数据离散的处理方法。为了验证所提方法的有效性,对2个数值算例的多种损伤模式进行了识别,并与没有经过粗集处理的PNN损伤识别方法进行了比较。研究发现,该文所提方法不仅可以降低数据的空间维数,而且具有很高的损伤识别精度。
In order to make effective use of the redundant and complementary information of structural health monitoring system to evaluate the structural health, a new method of damage identification is proposed, which combines rough set, data fusion and probabilistic neural network (PNN) organically. Firstly, it reduces the space dimension of the data by using attribute reduction of rough sets, and then uses PNN to calculate the redundant data and uncertain information. Finally, fusion decision and damage identification are performed. In the process of rough set attribute reduction, a method of data discretization using K-means clustering is proposed. In order to verify the effectiveness of the proposed method, a variety of damage modes of two numerical examples are identified and compared with the PNN damage identification method without rough set processing. The research shows that the proposed method can not only reduce the spatial dimension of data, but also have high damage identification accuracy.