Research On Improved Methods For Modal Identification And Damage Detection Of High Dam

By now,China has accomplished the most dams in the world with advanced technology,rich experience and high speed of construction.There are a batch of high dam beyond 120 meters constructed in China.But,the flow pulsating pressure is very likely to result in structural vibration or even fatigue destruction of flood discharge structures.Therefore,it is important to study the dynamic response and damage identification of flood discharge structures under discharge excitation.The numerical simulation,hydro-elastic model experiment and new data fusion method were carried out to study modal frequency change regularity of RCC dam,data level and decision level fusion method of vibration signal,modal parameters and damage identification of flood discharge structures in this paper.The paper mainly studies the following:1)research on vibration response characteristics of flood discharge structures.The numerical simulation was carred out to study effects of strength,thickness and quantity of rolling layer on high RCC dam modal frequency.The effect of rolling layer properties on the vibration frequency of RCC dam is basically the same as that of conventional concrete dam.When the thickness of the rolling layer is certain,the thickness of the rolling layer vary from 5cm to 10 cm,the elastic modulus of the soft layer is more than 70% of the elastic modulus of the non-damaged layer,the calculated results show that the modal frequencies of RCC dam are close to those of conventional concrete dam.2)Study on the method for working state identification of flood discharge structures.According to characteristics of multi-sensor vibration response of high dam discharge,data level fusion method of multi-sensor vibration signals based on variance contribution is proposed.The validity and de-noising effect of this method is verified through applying it to data fusion of vibration displacement signals on arch ring measuring positions in Jinping I high arch dam and data fusion of signals measured in the hydro-elastic model of RCC gravity dam.The new data level fusion method could reflect holistic dynamic characteristics of high arc dam.Besides,it has stronger applicability in the aspect of digging for dam body structure with closely spaced frequencies and high-frequency weak signals.3)Research on damage detection Method of flood discharge structures.According to information fusion technology and D-S evidence theory,this paper put forward the improved D-S evidence fusion method on the basis of Pignistic probability metrics and conflict coefficient.Compared with other methods,better convergence and recognition accuracy of this method is verified through typical cases.When given two evidences,the correct result can be recognized with the method.And the final identification accuracy is up to 98.4%.This paper summarizes sensitive characteristic indexes of flood discharge structure damage and identifies the damage of these indexes in RCC dam.The final identification accuracy is up to 99.89%.When conflict between evidences is not violent,the method could efficiently avoid or reduce effect of erroneous evidence on recognition of damage position compared with other methods.