基于盲源分离的损伤识别方法  

Damage identification method based on blind source identification

作　　者：林友新[1] 周翠[2] 李宏男[2] 

机构地区：[1]广东电网公司电力科学研究院,广东广州510080 [2]大连理工大学建设工程学部,辽宁大连116024

出　　处：《地震工程与工程振动》2013年第6期158-163,共6页Earthquake Engineering and Engineering Dynamics

基　　金：国家重点基础研究发展计划(973计划)项目(2011CB013605);国家自然科学基金委创新研究群体基金项目(51121005)

摘　　要：考虑环境因素和结构运行状态的损伤识别方法,在结构健康监测中越来越受到人们的关注。本文利用监测中提取的结构振动特征,提出基于盲源分离的损伤识别算法。该算法通过盲源分离中的AMUSE法可以将信号分为环境源和损伤源,其优点是不需要测量环境因素即可去除环境源影响。当结构处于健康状况时,结构损伤源信号很小;但当损伤发生后,该源会急剧增大。以预应力混凝土梁作为数值算例对该方法进行验证,温度作为影响结构反应的环境条件,且温度与结构特征参数具有非线性关系,算例成功地剔除了环境因素的影响,增大了结构损伤识别预警的准确性。Researchers have paid more and more attentions to damage identification method under varying environ- mental and operational conditions for structural health monitoring. In this paper, the AMUSE method based on blind source separation （BSS）is applied to vibration features identification during the monitoring process. The A- MUSE method could separate signals into environmental and damage sources. The advantage of the method is that it does not require to measure environmental parameters. When a structure is under health condition, the damage source remains small,and it increases significantly when structual damage occurs. The proposed method is illustra- ted using computer-simulated data by a prestressed concrete beam, and temperature as an environmental condition has a non-linear effect on the vibration features. The simulation case works successly in eliminating environmental effect, which increases the precision of damage alarm.

关 键 词：损伤识别 盲源分离 环境因素 温度效应 AMUSE 

分 类 号：TU37[建筑科学—结构工程] TP3[自动化与计算机技术—计算机科学与技术]