基于聚类算法的城市桥梁安全运营状态评价

Clustering Algorithm Based Operation State Evaluation of Urban Bridge

为提高城市桥梁运营安全性,提出了一种通过聚类算法来判别结构损伤程度的解决方案。首先建立结构有限元模型,并通过有限元模型计算结构在不同程度下的频率响应,然后采用K-means算法对频率进行聚类。结果表明,采用K-means算法可以以较大辨识度区分不用损伤情况,且桥梁损伤程度与聚类辨识度存在相关关系,该关系可为进一步提高聚类效果提供依据。

To improve the safety of urban bridge during operation,this paper proposed a kind of clustering algorithm to distinguish the degree of structural damage.First,structure finite element model was established,then based on which,the frequency was calculated.Second,the K-means algorithm was applied to clustering the results under different degree of damage.The result shows that the K-means algorithm can distinguish with larger recognizable without damage,and the bridge damage degree was correlated with the degree of clustering recognition,the relationship can provide the basis for further improve the effect of clustering.