The Analysis Of High Earth-rock Fill Dam Seismic Input And Seismic Deformation

Earth-rock fill dam construction is the most common type of dam in china, which is accounting for more than 90% of the total. High earth-rock fill dams are constructing in the southeast and northeast under the needs of western development and South-to-North Water Diversion.Most of these dams are located above high intensity of seismic region, which means the seismic resistance research is of high practical significance.The research on seismic issues of earth-rock fill dam is far fall behind the engineering practice. Theseismic input is the premise of earthquake research,which seriously affect the accuracy of earthquake prediction reliability. In actual, few of the high earth-rock fill damsuffered the disaster,which means the shortage in the seismic response and tactual damage records.Dam nonlinear dynamic theory of constitutive model and solving method is immatureyet, so, predictions by theoretical analysis and numerical simulation of the earth-rock fill damto the current seismic response and earthquake damage model is very difficult. Because of this, it’s also hard to sort out the high embankment dam failure process and mechanism of the earthquake, which affects the reliability of earth-rock fill dam seismic safety evaluation. The construction of this type dams has seriously hampered.This paper focuses on the impact of seismicinput on earth-rock fill dam seismic response and deformation and the permanent deformation parameters inversion of Zipingpu dam etc. for the above key issues to be solved ofearth-rock fill dam.The main contents and conclusions are as follows:(1)Generally, the design of hydraulic structures usually using the standard design spectrum which has been standardized. Periodic segment response spectrum has more and more influence on dam dynamic response by the growth of construction height of the dam. Attenuation coefficient γ determines the changes of long-period segment of the response spectrum in the value. Quantitative Analysis the impact of change the standard response spectrum decreased segment attenuation coefficient on seismic resistance of high earth-rock fill dam, and explore for high embankment dam attenuation coefficient values under the research of NGA(Abrahamson-Silva 2007 attenuation coefficient). The results shows that for the dams more than 200 m, its standard response spectrum normalized segment decreased attenuation coefficient γ take 0.7 is more appropriate than 0.9.The most commonly used prediction methodof near field strong seismic are ground motion prediction method works Green’s function method and stochastic finite tomographic method. The stochastic finite tomographic methodovercomes the limitations of the Green’s function method which may constrained by small earthquakes recorded and it also widens the estimated range of near field seismic.(2)Key of the stochastic finite tomographic method is to identify and determine the parameters of the ion source and the epicenter spectrum model. In this paper, corner frequency factor was added in stochastic finite tomographic method to solve the problem that when the corner frequency is different, the fitted results is better in high frequency and bad in low frequency. Using the improved finite fault method to consider the surface characteristics of big earthquake. The seismic motion time history of Wenchuan was predicted then by comparing the actual seismic motion time history. The improved method and the parameters of the ion source was compared too.(3) Threshold of BP neural network permanently deformed inversion model was optimized by artificial bee colony algorithm to solve the BP network model shortage of slow convergence and easy to fall into local minimum value. The convergence speed and accuracy of prediction of the network was increased. The improved model was verified by the data of Zipingpu CFRD got after the Wenchuan earthquake and compared with the model training method and prediction result of genetic algorithm optimization model. The result shows that: using the artificial bee colony algorithm to optimize BP neural network inversion model is applicable to permanent deformation permanent deformation parameters inversion and better than the genetic algorithm optimization model.(4) Seismic damage of Zipingpu dam after Wenchuan earthquake was considered in this paper by using improved stochastic finite fault method for synthesizing the seismic input acceleration process of the Zipingpu Dam site after Wenchuan earthquake. Then established theseismic response analysis model of dam –dam foundation–water to analysis the reason of the dam body deformation and deformation panel dislocation after earthquake. Comprehensive seismic method to improve the panel stress was recommended in this paper which may has a theoretical and practical significance to protect the damage CRFD suffered after extreme earthquake.