泄流激励下水工结构应变模态参数时域辨识研究

为准确辨识泄流激励下水工结构的应变模态参数,利用应变与位移之间的关系,建立以应变及应变率为状态变量表示的系统随机状态空间模型,构建基于应变数据驱动的随机子空间算法模型。针对时域法所面临的噪声干扰以及由它引起的虚假模态和模型定阶问题,通过小波技术对时域信号进行消噪处理,利用奇异熵增量和改进的稳定图对模态参数进行精确辨识。模型试验表明此算法可以准确辨识泄流结构的应变模态,为泄流结构的损伤诊断与动态检测提供基础。

Study on strain modal parameters identification of hydraulic structure in time domain under discharge excitations

A stochastic system state space model that uses strain-displacement relationship and variables of strain and strain rate,is developed to identify strain modal parameters of hydraulic structure under discharge excitations,and a corresponding stochastic subspace algorithm is constructed on the basis of strain data.This model adopts wavelet technique for denoise of time-domain signals to avoid the shortcomings of time domain method,i.e.,noise pollution,false modal parameters and unknown order of model accuracy,and to improve its accuracy.A new parameter identification method of using increment of singular entropy and improved stabilization diagram is put forward.Model tests show that this method can accurately identify the strain modal of outlet structure under discharge excitations and that it provides a basis for diagnosis and dynamic detection of hydraulic structure damage.