Abstract:

By combining strain mode with trust region optimization algorithm, a least-squares objective function was defined based on strain mode, then the detection problem was transformed into an optimization problem. Because the traditional modal strain index included both of damage location and damage degree information, there likely existed miscarriage of justice when the objective function was minimized, so the objective function based on strain mode was improved here. The improved optimization was carried out in two steps, the damage degree could be detected after it was located in the first step.Two numerical simulation experimental results show that the improved damage detection method is more desirable, and it provides a new idea for the applications of strain mode in practical engineering.

Key words: strain mode, trust region, objective function, damage detection

摘要：

将应变模态与信赖域优化算法相结合，提出了基于应变模态的最小二乘目标函数，将损伤识别问题转化为二次优化问题。由于传统的应变模态差既包含损伤位置信息，又包含损伤程度信息，对其进行目标函数最小化时容易产生误判,因此对应变模态目标函数进行了改进，改进后的优化分两步进行，实现了损伤识别的先定位、后定量。最后，通过两个数值算例对该方法进行了验证，结果表明改进方法的识别效果更为理想，为应变模态法在实际工程中的应用提供了新思路。

关键词: 应变模态, 信赖域, 目标函数, 损伤识别