摘要
模态参数的准确识别是对在线结构进行健康监测和损伤诊断的难点和核心之一。以拉西瓦拱坝水弹性模型为工程背景,对拱坝的模态参数进行时域识别。针对时域法所面临的噪声干扰以及由它引起的虚假模态识别与剔除和模型定阶问题,用小波技术对时域信号进行了消噪处理,综合自然激励技术和特征系统实现算法对环境激励下的拱坝结构进行模态参数识别,并用模态置信因子对虚假模态进行剔除。引入奇异熵的概念,建立了用奇异熵增量来实现系统定阶的方法和过程,研究表明,该方法得到的系统阶次有效可靠,使得定阶的界线更加清晰和稳定,避免了其它时域算法对系统定阶的盲目性,提高了数据处理和模态参数识别的速度。
The modal parameters of arc dam are identified in the time domain based on the hydroelastic model of Laxiwa arc dam. Aiming at reduction of noise disturbance, consequent identification and elimination of false modes, and solving the order- determination problem in time domain, wavelet technique is used to denoise the time domain signal. The natural excitation technique(NExT) and the eigensystem realization algorithm( ERA) are combined to identify the modal parameters of ambient-excited arc dam, and the false modes are eliminated by modal confidence factors. Singular entropy is introduced, and the method and the process of how to determine the order of system are realized through the increment of singular entropy. Study shows that the order estimated is available and credible through the method, and the borderline.. for oder determination is clearer and more stable. The method avoids the blindness in oder determination compared with other time domain algorithms, and improves the speed of data processing and modal parameter identification.
出处
《振动与冲击》
EI
CSCD
北大核心
2007年第12期101-105,共5页
Journal of Vibration and Shock
基金
国家自然科学重点基金(编号50538060
50679052)
新世纪优秀人才支持计划资助
关键词
模态参数识别
特征系统实现算法
奇异熵增量
定阶
小波分析
modal parameter identification
eigensystem realization algorithm
increment of singular entropy
determination of order
wavelet analysis
