Vibration-based methods for damage localizations often rely on a damage feature defined in terms of changes of modal shapes and localize damage by detecting shape irregularities that are ascribed to local loss of stiffness. In this paper, the performance of several algorithms for damage localization is investigated and the results are compared in terms of the gain of information they provide, which is measured by the relative information entropy also known as Kullback–Leibner (KL) divergence. This parameter is a measure of the difference between two probability distributions and can quantify the information gain achieved using different statistical models. In this paper the relative entropy is used to compare the gain of information obtained using different damage-sensitive indicators retrieved from simulated structural health monitoring data. The investigation is carried out using structural responses simulated using the finite element model of a real bridge permanently monitored by the Italian Seismic Observatory of Structures.

中文翻译：

---

基于振动的破坏指标：基于信息熵的比较

基于振动的损伤定位方法通常依赖于根据模态形状变化定义的损伤特征，并通过检测归因于局部刚度损失的形状不规则性来定位损伤。在本文中，研究了几种用于损伤定位的算法的性能，并根据它们提供的信息增益对结果进行了比较，这些信息是通过相对信息熵（也称为Kullback-Leibner（KL）散度）来衡量的。该参数是两个概率分布之间差异的度量，可以量化使用不同统计模型获得的信息增益。在本文中，相对熵用于比较使用从模拟的结构健康监测数据中获取的不同损伤敏感指标获得的信息的增益。