Improved BWBN Model Of Reinforced Concrete Column And Its Parameter Identification

When subjected to intense earthquake action,most structure will enter the plastic stage.For RC column,its performance is very different from the elastic stage,and the components will often show obvious degradation after yielding.If the overturning moment caused by the horizontal earthquake and the vertical seismic component of the near-field earthquake are considered,the axial force on the column will also continue to change.According to the experimental research of previous scholars,it is found that the hysteresis characteristics of the variable axial force RC column and the performance under the fixed axial force are very different.Under the variable axial force,the stiffness degradation,strength degradation,ductility,etc.of the component will often be different.At the same time,there are phenomena such as hysteresis asymmetry.In order to make the hybrid test or simulation results more accurate,a resilience model that can describe the above phenomenon is needed.As a smooth curve model,the BWBN model can be used to describe the performance of most components,but it is not accurate enough to describe the hysteretic performance of the specific component of the RC column.For this,the following research has been conducted in this thesis:（1）The thesis uses an improved PSO based on genetic algorithm and chaos optimization algorithm to identify model parameters.In order to verify the recognition accuracy of this improved algorithm,four algorithms of PSO,CPSO,SA and CPSO-GA are used to identify the BWBN model with given parameters.According to the analysis of the recognition results,in terms of recognition accuracy,CPSO-GA has the best recognition effect for the 13-parameter BWBN model.In terms of recognition efficiency,although the number of convergence steps of CPSO-GA is long,its early accuracy surpasses other algorithms.At the same time,a large number of iterations occur with high accuracy,indicating that this improved algorithm effectively overcomes the premature convergence problem of PSO.（2）In order to make the BWBN model better describe the various degradation phenomena of the RC column,this paper makes corresponding improvements to the BWBN model based on the improved Park-Ang damage model.According to the test data of columns with constant axial force,it is found that the stiffness degradation,strength degradation,and post-yield stiffness have a clear functional relationship with the damage index,and the related items of the original model are modified according to this relationship.In order to further apply the BWBN model to the RC columns with variable axial force,this thesis first classifies the axial compression ratio at each level of 0.05 to replace the constantly changing axial force,and calculates the yield force,displacement and other parameters under monotonic loads separately for each level.The axial pressure ratio in the middle of each stage calculates the above parameters by linear interpolation.Then,the damage of the variable axis force column is calculated separately in positive and negative directions,which can describe the hysteresis asymmetry caused by the asymmetry of the damage.Finally,the correlation curve was used for comparison,the results are good,indicating that the improved model can be used in related research such as hybrid tests.At the same time,the research method of this thesis also has certain reference significance for the future research of the restoring force model of columns with variable axial force.（3）The improved BWBN model has many parameters and each parameter has a different degree of influence on the shape of the hysteresis.This thesis analyzed the influence and sensitivity of model parameters.According to the analysis result,it is found that the influence degree of parameter p、q、ψ、δ_ψ、λare small,and they can be considered to be fixed as constants during parameter identification.Finally,the test data is used to verify the identification method of fixed part of the parameters,and it is found that there is little difference between this method and the full parameter identification,which indicates that this method has good applicability.