本文采用结构应变测试信号,引入了伪比能概念,直接在时域内对梁结构进行损伤识别。为了减少损伤识别的分析计算时间、提高识别精度,采用了“两步法”识别法:第一步,建立了伪比能(Pseudo Strain Energy Density)损伤识别指标,利用其来进行损伤定位;第二步,在损伤定位的基础上,利用遗传算法对损伤部位的具体损伤程度进行判别。并通过编程连接了Matlab与ANSYS,形成了完整的基于遗传算法的结构损伤识别系统。最后,通过一钢筋混凝土三跨连续梁有限元模拟数据,验证了本文提出的方法的有效性。
A strategy using strain test data for detecting the damage in beam structures in time domain is presented in this paper. The "Pseudo Strain Energy Density (PSED)" is defined and a two-step approach is set up to reduce the analyzing time and improve the precision of structural damage identification. The damage index PSED is proposed to implement damage localization. Also, the Genetic Algorithm (GA) is utilized to identify structural damage degree of the confirmed damaged locations. Furthermore, a systematical damage identification program is developed based on GA is developed by means of the software MATLAB, and the finite element (FE) software ANSYS that is employed to conduct the FE analysis of corrrplicated civil engineering structures, is embedded in this program with the interface programming technique. Finally, a three-span reinforced concrete continuous beam is utilized to verify the effectiveness and performance of this new damage identification strategy.