基于随机减量法的分布式结构模态参数识别

摘要

Technology for partition processing and parallel analysis is essential to realize a dense array of wireless smart sensors network for measuring on large-scale civil structures.Here,a distributed data collecting approach for a system's modal identification was proposed based on the random decrement technique (RDT).The performance of the RDT-based distributed data collecting was tested using a two-side simply supported plate model.Using ISM400 wireless smart sensors,the random decrement function was calculated with the measured vibration acceleration time histories of the plate by adopting the natural excitation technique (NEXT).A time domain algorithm integrating NExT and the eigen system realization algorithm (ERA) was applied to identify modal parameters of the plate and combined with the method of stability diagram (SD) to eliminate false modes.The identification results were compared with those based on the centralized method.It was shown that the plate's modal shapes obtained with the proposed method are close to those obtained with the centralized method using the modal assurance criterion (MAC).%分区进行平行分析处理的技术已成为大型结构密集布置无线智能传感器网络检测结构系统的重要任务.提出基于随机减量法的分布式数据采集和模态识别方法.以两边简支板模型试验为例,采用ISM400无线智能传感器,通过自然激励法获得测试结构的响应信号,计算随机减量函数,然后运用特征系统实现算法提取系统的状态空间参数,并结合稳定图的方法剔除虚假模态,识别出结构的模态性能参数.以模态置信度为判据对比分析分布式算法与集中式算法的识别效果,结果表明两种算法吻合良好.