基于FDD法的模态参数连续自动识别及频率变异性分析

Continuous Automatic Identification of Modal Parameters and Analysis on Frequency Variability Based on FDD Method

模态参数是分析桥梁运营状态变化的重要指标,可为桥梁状态评估提供数据基础。基于频域分解法(FDD)的模态参数连续自动识别方法,分析了模态参数与环境之间的关系。采用前、后向线性均值滤波方法对奇异值曲线进行滤波处理,剔除噪声干扰引起的虚假毛刺峰值;提出了Dvp指标,结合波峰波谷检测算法识别得到了真实模态参数对应的奇异值峰值;引入K-means聚类法实现了奇异值曲线真实峰值的自动选取。针对实际健康监测数据存储格式,设计了模态参数连续自动识别框架,并提出了融入模态参数的自动识别算法,实现了模态参数连续、自动识别。将所提出的方法运用于斜拉桥的实测数据中,通过连续自动识别得到一个月的模态参数,分析了环境对结构频率的影响程度。研究结果表明:所提出的基于FDD的模态参数自动识别方法能够实现模态参数的自动识别;所提出的融合自动识别算法的数据自动分析框架能够实现模态参数连续、自动识别,能解决桥梁健康监测数据的连续自动分析;桥梁结构频率受环境温度影响较大,随机环境因素所引起的结构频率随机性不可忽略。

Modal parameters are important indexes to analyze the change of bridge operation status,which can provide data basis for bridge status evaluation.A continuous automatic modal parameter identification method based on frequency domain decomposition(FDD)was proposed,and the relationship between modal parameters and environment was analyzed.The forward and backward linear mean filtering method was used to filter the singular value curve to eliminate the false burr peak caused by noise interference.The D vp index was proposed,and the peak value of singular value corresponding to the real modal parameters was obtained by combining with the peak and trough detection algorithm.The K-means clustering method was introduced to realize the automatic selection of the true peak value of the singular value curve.According to the actual health monitoring data storage format,a framework of continuous automatic identification of modal parameters was designed,and an automatic identification algorithm incorporating modal parameters was proposed to realize continuous and automatic identification of modal parameters.The proposed method was applied to the measured data of cable-stayed bridges,and the modal parameters for one month were obtained through continuous automatic identification.The impact of the environment on the frequency of the structure was analyzed.The research results show that the proposed automatic modal parameter identification method based on FDD can realize the automatic identification of modal parameters.The proposed data automatic analysis framework combined with automatic identification algorithm can realize continuous and automatic identification of modal parameters,and solve the continuous and automatic analysis of bridge health monitoring data.The bridge structure frequency is greatly affected by ambient temperature,and the randomness of structure frequency caused by random environmental factors cannot be ignored.