针对大型结构实际模态识别中存在难以准确对系统定阶和产生虚假模态等问题,提出了采用奇异熵导数结合稳定图技术进行子结构系统定阶并去除虚假模态的方法。以某简支梁结构为例,采用MATLAB软件仿真模拟振动测试和信息采集,进行数据处理分析。结果表明:奇异熵导数对信号信噪比的变化反应非常敏感,能够根据其零值点位置准确确定降噪的合理阶次;频率稳定图技术能够有效去除虚假模态,对子结构进行系统定阶,并有助于子结构之间进行模态阶次匹配,以获得准确的整体结构模态参数。
Aimed at the problems that was difficult to determine the substructure system order and false modal appeared in practice, differential quotient of singular entropy and stabilization diagram technology were used together to determine substructure system order and eliminate false modals. A MATLAB finite element model of a simply supported beam was taken for example to analyze the proposed method, by simulating vibration testing and data acquisition. The results show that differential quotient of singular entropy is sensitive to change of signal-noise ratio, and it can be used to determine reasonable noise reduction order according to the location of zero point. Frequency stabilization diagram technology is effective to eliminate false modals and get order for substructure system. It is helpful to match modal orders between substructures to obtain true modal parameters of whole structures.