| Damage evolution of civil engineering structure is a process that spans multiply temporal and spatial scales.In the civil engineering structure,there are significant differences in the spatial scale between the micro/meso defects in the material and the macro structure.What's more,the damage evolution rates in various space scales caused by various loads are also quite different,which makes the traditional single-scale damage analysis difficult to describe accurately the connection between meso damage evolution rules and overall structural degradation,and difficult to simulate the damage evolution process under different damage evolution rates reasonably and effectively,which brings hidden trouble to structural damage analysis and life cycle safety assessment.To meet the needs of engineering structural safety analysis,and in order to study the space-time multi-scale damage mechanism of a large-scale civil engineering structures,the paper aims to develop a simulation algorithm to analyze multi-scale damage evolution of the structural failure process.The main research work and achievements of this paper are as follows:In this paper,based on the continuum damage mechanics,a constitutive equation which can describe material softening phenomenon caused by the damage of metal material was prepared through the user material subroutine.Furthermore,damage parameters can be inverted according to the experimental data.The numerical solution obtained was compared with the analytical solution to verify the correctness and effectiveness of the user material subroutine.The seismic performance of two types of steel beam-column joints and a two-story steel frame with the beam-column joints under seismic repeated loading is analyzed.The evolution process of meso damage of beam-column joints and steel frame under quasi-seismic loading is studied.The influence of damage-induced material softening on seismic performance indicators is quantitatively analyzed.The relationship between damage-induced material softening and the seismic performance of the steel frame is then analyzed based on the spatial multi-scale model of the two layer steel frame.The results show that the user-defined material subroutine is correct and effective to describe damageinduced material softening behavior caused by material damage evolution.The results obtained by considering the damage-induced material softening are more consistent with the actual engineering situation,so the damage-induced material softening behavior should be taken into consideration in the process of damage failure of the engineering structure.The damage-induced material softening behavior affects the seismic performance of steel members,in which the stiffness degradation coefficient and strength degradation coefficient are greatly affected,while the viscous damping coefficient is less affected.In the early stage of the seismic load,the damage-induced material softening behavior has no effect on the seismic performance of the steel frame and can be ignored.But the damage-induced material softening region will expand rapidly,making the overall seismic performance of the steel frame decrease obviously just before the failure of the steel frame.The performance deterioration of steel frames is more abrupt and rapid compared with the steel members.The nonlinear stage of the skeleton curve of the steel members and the steel frame can be divided into two stages,the strengthening stage and the softening stage.The increase of the damage-induced material softening area is related to the performance deterioration and failure of the steel members and steel frame.Based on the D-N non-overlapping domain decomposition theory,a serial substructure nonlinear computation framework suitable for multi-scale damage analysis of engineering structures is established.The algorithm decomposes an engineering problem into several subcomponent problems,which reduce the computing scale of the engineering structure model efficiency.Based on a variety of development interfaces of commercial software ABAQUS,and two feasible algorithm implementation methods are proposed.A macroscopic-mesoscopic multi-scale model of a steel truss beam with defects is established,based on the DN non overlapping domain decomposition method through the ABAQUS Python interface.The response of the truss beam with defects under loading is analyzed to verify the feasibility of the serial substructure nonlinear algorithm.Through a large number of numerical experiments,the convergence properties of the serial substructure nonlinear calculation method,like the relaxation factor,the calculation precision,the size ratio and the initial displacement,are studied.It is suggested that the relaxation factor should take a maximum value in the case of convergence to improve the efficiency of calculation.A simplified boundary equations are introduced at the interface between the macro and meso models,resulting in the discontinuity of stress on the boundary.In the implementation of the serial substructure nonlinear calculation method,the mesoscopic substructures are sequentially calculated in turn,the algorithm can be improved to have higher parallelism.Therefore,a parallel substructure nonlinear computation framework suitable for multi-scale damage analysis of engineering structures is then established,based on the overlapping domain decomposition method.Through the multi-scale model of “L” concrete member with multiple substructures,and the damage characteristics under loading is then investigated.The results show that the algorithm can be used for the analysis on the structural damage failure problems with strong nonlinearity.Numerical experiments are carried out to study the effects of the parameters in the algorithm on the convergence speed and accuracy of the method.It is suggested that the relaxation factor should take a maximum value within the allowable range of the algorithm to improve the efficiency of calculation.These result lays the foundation for further application of this method to engineering practice.The strategy for numerical analysis on multi-scale evolution process of fatigue damage is proposed,and a spatial multi-scale algorithm for the analysis on the fatigue damage evolution process is established by writing Python script and user subroutines.An adaptive algorithm with variable step-size for numerical analysis on nonlinear high-cycle fatigue damage accumulation was proposed.The damage status of the hot spots of the Tsing Ma Bridge caused by high-cycle fatigue loading is analyzed based on two types of fatigue criteria.The results show that the structural life obtained by linear Miner criterion is conservative.If the nonlinear damage evolution criterion is used to predict the structural life,the adaptive algorithm can effectively improve the computational efficiency.The proposed spatial multi-scale algorithm was applied to analyze the interactive accumulation process of high and low cyclic fatigue damage.The spatial multi-scale model of the existing industrial building structure was established to analyze damage state and residual life of the industrial buildings under traditional serving load and sudden earthquake load with two types of fatigue criteria.The results show that,the macro cracks are more likely to form at the variable section(B points)of the crane beam during the service process.Structural damage is concentrated somewhere under high cycle fatigue load.The damage value of structural vulnerable parts is increased and structural life is lower when subjected to seismic loading.The results of the computation verifies the feasibility of the spatial multi-scale strategy. |
PDF Full Text Request