Probabilistic Progressive Collapse Analytical Method Of Steel Frame Structures

Steel frame structure is a common type of building system which has high strength and ductility with rapid construction speed.The probability of structural collapse is small but the loss might be huge.Therefore,research on probabilistic progressive collapse of steel frame structure need to be carried out.The dissertation focuses on steel frame structures.First,a reduced numerical model of steel frame structure is proposed considering the cantenry action of steel connection and membrane action of composite floor systems.Based the numerical models,probabilistic steel connections are established and reliability analysis of steel frames structures is carried out to study the probability of collapse.Since probabilistic analysis of structural collapse often relies on time-consuming numerical simulations,a design-oriented progressive collapse resistance capacity model of steel frame structure is proposed to improve the computation efficiency.On the other hand,the direct Monte Carlo method requires large number of samples which is low efficiency in predicting small failure probability.Therefore,based on subset simulation technique,a rapid probalistic progressive collapse method is proposed to reduce the number of samples in reliability analysis.The primary work and achievements of this dissertation are as follows:(1)A reduced numerical model of steel frames structures is proposed considering the cantenry action of steel connection and membrane action of composite floor systems.Based on the proposed numerical model,numerical simulations of existing testes are carried out as a validation of the reduced model.The influences of cantenry action of steel connection and membrane action of composite floor on structural collapse resistant capacity are also studied.Results shows that : 1)The proposed reduced connection model can effectively and efficiently simulate the complex mechanical behaviors of steel connection,including the axial force-moment interaction effect resulting from catenary forces,the changes of rotation centers and the fracture of steel connections.The proposed model can also simulate the membrane action of composite floor under large deformation conditions.2)Neglecting the cantenry action of steel connection will over-estimate the collapse resistance capacity of steel structure while neglecting the membrane action of composite floor will under-estimate the collapse resistance capacity.3)The steel frame structure subjects to small vertical deformation and damage due to the loss of columns from moment bays.However,the steel frame structure subjects to progressive collapse due to the loss of columns from gravity bays.The gravity frames with beam-column jointed by simple connections are the weakest part of the whole building against progressive collapse.(2)A probabilistic analysis methodology of steel frame structures against progressive collapse considering uncertainties in strength and ductility of steel connections is proposed.Based on the reduced model of steel connection,an analytical framework for developing probabilistic connection models is proposed.The connection models developed are further introduced in probabilistic progressive collapse of structural building.Fragility curves of steel frame structures corresponding to different limit states are developed.The proposed method is appdied to steel gravity frames jointed by simple connections.Discussions on the influences of steel connection and slab continuity on collapse vulnerability are given.Results from the present probabilistic method are compared with those from deterministic approach.Results show that: 1)The proposed method can effectively simulate the randomness of mechanical behavior of steel connections due to the uncertainties factors.It also can be used to evaluate the collapse resistance capacity of structural bulidng from the perspective of relability theory.2)For the steel connection analyzed in this study,the coefficient of variations of ductility of steel connection ranges from 0.08 to 0.37 while the coefficient of variations of strength of steel connection rages from 0.08 to 0.16.The mechanical behaviors of steel connections composed of brittle components present a larger variation.3)The compressive strength of concrete,yield strength of steel deck,and ultimate strength and ductility of steel connections are the most influential random variables on collapse resistance capacities based on the sensitivity analysis results.3)Robustness evaluation results according to UFC guidelines are rather conservative with probability of non-exceedance ranging from 0.91 to 0.98,for the steel frame structures analyzed in this study.On contrary,using load combinations from GSA may lead to a non-conservative design with the probability of non-exceedance ranging from 0.25 to 0.38.(3)A design-oriented probabilistic capacity models for steel frame structures subjected to column loss is proposed.The probabilistic capacity models are formulated by adding explanatory terms to an existing deterministic model.The explanatory terms are selected to correct the bias in deterministic capacity model.The ranges of design variables are determined according to design guideline and common practice.The D-optimal design algorithm is used to generate the virtual design configuration.Based on the reduced numerical models,virtual experiment data are generated from finite element simulations to calibrate the model parameters.The calibration of model parameters is conducted using Bayesian approach.A progressive regression & deletion method is adopted to simplify the proposed capacity models.As an application and validation of the probabilistic capacity models,reliability analyses of steel gravity frames jointed by simple connections are carried out.The developed fragility curves are compared with those from finite element analysis cooperated with Monte Carlo simulation.The results show that: 1)The proposed capacity models are accurate in predicting the resistance capacity of the steel frame structures,compared with the existing analytical models and the results from numerical simulations.2)The proposed model does not require numerical simulations in progressive collapse analysis,which improves the computation efficiency in predicting failure probability of structural collapse.(4)An efficient probabilistic assessment framework is proposed to evaluate the failure probability of structures against progressive collapses.The subset simulation technique is introduced in the reliability analysis and further combined with the alternative load path method and two-step method in rapid predicting probability of failure of structural building.The proposed method is applied to a ten-story steel frame structure to study the failure probability of structure under column removal and blast loading.The results show that: 1)compared with the direct Monte Carlo simulation method,the proposed method requires less number of samples and is more efficient in predicting failure probability of structural collapse.It also be adopted to determine the most effeictive measures in protecting structures against collapse.2)For the steel frame structure analyzed in this study,the conditional failure probability of structure given the column failure exceeds 0.85 if the blast load is directly targeted as the column from gravity bays.Therefore,the lack of structural robustness is the main reason leading to the structural collapse against blast.If the blast load is directly targeted against the column from moment bays,the conditional failure probability is not senstive to the changes of standoff distance and TNT mass and ranges from 0.17 to 0.24 Therefore,the structure is robust after column loss so that the most effective measure in protecting structures is to design the structural members blast-resistant.