The Reliability Research Of Structural Seismic Rdliability Based On AK-MCS

As a typical natural disaster with strong and destructive power,it is still impossible for earthquake to establish a perfect prediction mechanism for earthquake and defensive measures are commonly used in response to the threat.The earthquake can make great damage to building structure while the potency and duration has strong randomness.In order to meet the performance-based seismic design concept,uncertainties such as structural material properties and randomness of seismic action should be considered in structural seismic analysis.The probabilistic seismic performance evaluation based on reliability theory is a popular research direction of structural seismic design.The limit state is highly nonlinear and implicit when the building structure is subjected to earthquakes,meanwhile,structural seismic response solution requires a large amount of computation.Therefore,it is difficult to solve the seismic reliability of the structure by the traditional reliability calculation method.In this paper,the AK-MCS method that combines the active learning Kriging(AK)with the Monte Carlo Simulation(MCS)is used to analyze structural seismic reliability.The main research contents are as follows:(1)The principle and basic steps of Kriging model calculation reliability are introduced and the performance of different reliability calculation methods is compared by numerical examples.MATLAB is used to implement AK-MCS model.Through numerical and project example,the best selection function is selected by comparing the different site selection function and the reliability and efficiency of reliability calculation before and after efficiency optimization.And the optimal selection function is selected to verify the optimization of the iterative stop point optimization and clustering optimization.(2)The AK-MCS method is applied to structural seismic reliability analysis and three different layers of reinforced concrete frames is built by OPENSEES.Elastic dynamic analysis was carried out to obtain the frame characteristic period,mode displacement and other information and MATLAB is used to implement the material properties of the frame and the randomness of the seismic action.Pushover analysis of the frame finite element model is used to obtain the relationship between the base shear force and the apex displacement,and the capability spectrum method is used to find the performance point and the corresponding base shear force.The load capacity and deformation capacity according to the seismic design requirements of the frame are taken as the limit state.Using the optimized AK-MCS method to analyze the structural seismic reliability and evaluate the seismic performance of three reinforced concrete frames.Comparing the reliability analysis results with the Monte Carlo simulation method,the accuracy and efficiency of AK-MCS is analyzed.(3)The three frameworks of natural seismic wave excitation were selected to analyze the dynamic seismic reliability of three different layer frame structures using the optimized AK-MCS method and Monte Carlo simulation method.The strength-frequency non-stationary random ground motion model is used as the three frameworks of ground motion excitation by MATLAB,and the dynamic seismic reliability of different layer frame structures under random ground motion is analyzed.The importance of each random variable of the frame structure to the static and dynamic seismic reliability is calculated,and the influence degree of the seismic performance of different random variables is analyzed.The research shows that compared with Monte Carlo simulation method,the optimized AK-MCS method could effectively reduce the number of structural real response times,and the efficiency is significantly improved in structural seismic reliability analysis.The accuracy of AK-MCS method can meet the structural seismic reliability analysis requirements,and the importance measure of each variable in the analysis results is consistent with the engineering experience.