基于响应倒谱域特征的水电站厂房结构损伤评估
Damage Assessment of Hydropower Plant Structures Based on Response Cepstrum Domain Characteristics
赵玉宏 1李乐晨 2杨志超1
作者信息
- 1. 河南省水利科技应用中心,河南 郑州 450003
- 2. 河海大学 水利水电学院,江苏 南京 210098
- 折叠
摘要
数据驱动的建模方法已被广泛用于水工结构安全监测领域,但仍面临诸多挑战,包括振动响应噪声的干扰以及复杂模型泛化能力不足和过拟合现象.针对大型水电站厂房结构振动损伤问题,文章提出了一种基于响应倒谱域特征的新方法,通过分析振动响应的倒谱系数和主成分分析,有效捕捉结构振动的动态特性并减少噪声和负载的干扰,实现了厂房损伤准确识别和量化.
Abstract
Data-driven modeling methods have been widely used in the field of safety monitoring for hydraulic structures;however,they still face several challenges,including interference from vibration response noise,insufficient model generalization capabilities,and overfitting issues.To address the problem of vibration damage in large hydropower station's powerhouse structures,this paper proposes a novel method based on response cepstral domain features.By analyzing the cepstral coefficients of vibration responses and applying principal component analysis,the method effectively captures the dynamic characteristics of structural vibrations while reducing interference from noise and loads,enabling accurate identification and quantification of structural damage.
关键词
结构振动损伤/响应倒谱特征/主成分分析/异常检测Key words
structural vibration damage/response features in quefrency domain/principal component analysis/anomaly detection引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
万方数据