| A large number of experiments show that the joint shear deformation and the slide of longitudinal bar through beam-column joint which cause the deformation of beam ending even reaches half of the total deformation cannot be ignored. As a key part of the performance-based design, beam-column joint suffers complex forces in earthquake, and its performance is affected by many factors. Large nonlinear deformation of the beam-column joint will cause serious pinch effect and reduce the structural energy dissipation capability.Firstly, choose some beam-column joint species with complete test data as analysis object and establish corresponding finite element model in ABAQUS. Steel bar and reinforced concrete are fully controlled by using EMBEDED function, regardless of the bond-slip between them. The response of model is simulated under the cyclic loading and monotonic loading. The analysis show that the load-displacement curve under monotonic load is in good agreement with the bearing capacity despite the large load stiffness. The load-displacement curve under cyclic load is plump and no pinch effect., while,there is a big difference in contrast with the result of experiment.On the basis of the previous finite element model, SPRING element are used to consider the effect of bond-slip between steel and concrete. Selective analysis the bond-slip constitutive under monotonic loading and some corresponding adjustments are made according to the test. The analysis show that the load stiffness is more suitable and has no effect on the bearing capacity after using the SPRING element. Also analysis the effect of load-displacement curve under cyclic load after using the SPRING element. But considering the characteristics of the SPRING element which is only applied to monotonic loading in ABAQUS and the actual situation of cyclic loading, other methods are necessary.Simulate beam-column joints under monotonic loading and cyclic loading using ABAQUS.By comparing the calculated coordinate data obtained from the simulation results with the corresponding original experimental data, the results' reasonability is proved. On this basis, the proposed shear calculation model is evaluated by mathematical principles using the results under monotonic loading. It also proved that the parallelogram assumption is reasonable, while the assumption of diagonal is unreasonable and the diagonal rotation cannot be ignored. Get the same conclusion when further extended to the simulated results unde cyclic loading. calculate the measured shear angle calculated by the node's coordinates., and compare with the shear angle respectively using the proposed two calculated methods.The analysis show that there is a big difference among them, despite the similar total shear angle.Mo Linhui proposed a longitudinal bar stress-slip(?-s) model in 2013.But the processing of data and parameters is relatively rough, and the given results are nested and complicated. On this basis,after further collecting data and enlarging the sample size, the relationship between the longitudinal reinforcement stress and the corresponding longitudinal reinforcement slippage is got. New modified formulas for each feature point are acquired by SPSS to sregression and optimization Choose some beam-column joint tests, use the bond-slip constitutive model with new modified formulas and take the influence of bond slip into consideration. Contrasting the above with the experimental results and simulation results by Mo Linhui, the rationality and simplicity of the new modified formula are verified.This article's main innovation points are as follows:(1) Modify the bond-slip constitutive models in SPRING element, and establish a more appropriate bond-slip constitutive for the beam-column joints under monotonic loading.(2) Based on ABAQUS simulated data, it shows that the assumption of diagonal exists a difference in shear calculation model.The shear angles calculated by the proposed two methods are also existe a big difference.(3) By optimizing the ?-s model proposed by Mo Linhui, get the more reasonable and simpler formulas for each feature node. |
PDF Full Text Request