Study On Working State Identification And Beat Vibration Of Ertan Arch Dam Based On Hilbert-Huang Transform

Flood discharge excitation of high arch dams can easily appear fatigue orvibration damage under long-term discharge load, seriously affecting the normaloperation of projects and benefits of play. We must seek for effective means ofdetection to quickly and accurately find the hidden dangers and diseases existing inhydraulic structures, predict of the security situation effectively, and assess itsoperation reliability. Hydraulic structural damage will lead to the mass or stiffnessloss in some specific part, and then change its dynamic characteristics, which will bereflected in the modal information. So how to obtain accurate modal information is animportant prerequisite for running condition monitoring of hydraulic structure. In thispaper, a series of studies are researched on this issue, including the following fouraspects:Firstly, study on vibration characteristic of Ertan arch dam under flood dischargeexcitation. The autoregressive AR model is the first time applied to analyzingspectrum characteristics of hydraulic structure vibration signals under flood discharge.The11measuring points arranging along the dam body are in the analysis ofamplitude domain、frequency domain and characteristic index along the top arch ringand crown cantilever, so that we can obtain the vibration characteristic of Ertan archdam under flood discharge excitation.Secondly, study on filtering method based on empirical mode decomposition(EMD).It is the first time to introduce the HHT method into the hydraulic structurevibration signal analysis, and attach to the EMD method whose essence is filtrationfrom high frequency to low frequency. Based on time characteristic scale of signalitself, we can regroup some IMF components, and constitute a high-pass, band-pass,low-pass filter to filter out noise. The EMD filtering method based on signalcorrelation and based on the combination of EMD and wavelet threshold filtering areput forward. The above two methods are applied in the measured vibration signaldenoise.Thirdly, study on “beat vibration” of Ertan arch dam. This text introduces theconcept of natural “beat”, analyzes of the effects caused by initial difference,frequency ratio and amplitude ratio. With the application of statistical analysis aboutfrequency domain and SSI results, combined with special features of “beat vibrations”itself, we can research “beat vibrations” of Ertan arch dam under discharge flow.Fourthly, study on modal parameter identification of high arch dam under flooddischarge excitation. Aimed at the difficult problem of exciting the prototype to test,combined with the characteristic of ambient excitation, the modal parameter of structure is directly identified according to the dynamic response of structure underambient excitation in the paper. First a method named NExT+HHT is proposed toidentify modal parameters of Ertan arch dam under the working state, and get the firstfour-order modal parameters. Aiming at the mode mixing phenomenon appearing inthe EMD process of response signal of closely spaced structure, it is the first time toput forward a band-pass filter and the spurious signal processing method ofcombining, which can be used to control the EMD process of vibration signal.andobtain modal component of each mode accurately. Then extract the free decayinformation from every modal component with the application of the random decrementtechnique, and finally obtain the structure modal parameters through the hilberttransform and least square fitting.