Damage Mechanism Of Steel Structures And Fragmentation Process Of Masonry Walls Under Blast Loading

Recently, large-scaled terrorist explosions occur frequently all over the world. On the other hand, accidental explosions induced by misoperation or improper storage of inflammable substances (e.g. natural gas, firework) also occur commonly in the daily life. In urban environment, the explosions would result in serious casualty and economical lose, and more severe disaster would be caused by serious damage or progressive collapse of buildings subjected to blast load. Therefore, the destruction level caused by blast would be mitigated through improving the explosion resistance of structures. In this thesis, three aspects of structural response under blast load are discussed, and they are 1) a method for evaluating the damage of steel square tubular column caused by explosion and post-explosion fire is proposed; 2) the damage patterns and collapse mechanism of planar lattice structures and steel frame structures are discussed; 3) the fragmentation process of masonry infill wall subjected to blast load is studied. More detailed introduction and some conclusions are presented as the following.(1) A method for evaluating the damage degree of steel column subjected to blast load and fire action is introduced. Johnson-Cook strength model and Bonora damage model are employed to describe the constitutive model of mild carbon steel, in which both the plastic stress hardening and strain rate effect are considered, and the mechanical damage caused by blast load is also taken into account. The reliability of the proposed material model is verified through a numerical simulation of a fieldtest. Pressure-Impulse (P-I) diagram is established for evaluating the damage level of steel square tubular column subjected to blast load. Moreover, according to the steel temperature in ISO834 standard fire environment and the strength reduction law suggested by Euro-code, the fire resistance of the explosion-survived steel column is studied. The effects of blast pressure, impulse and fire time on the column residual capacity are discussed. A Pressure-Impulse-Fire time (P-I-t) coherence function is presented, and it can be used to predict the damage degree of the tubular column under the integrated action of blast and fire. A parametric study on the effects of geometric sizes is conducted.(2) The dynamic response and collapse patterns of planar lattice structures with four common different layouts of steel column under blast loads are numerically studied, and the critical blast pressures of the lattice structure with different explosion-evolved columns are calculated. The results indicate the following four conclusions, which are listed as: 1) If the planar lattice structure is only supported by the side columns, catastrophic collapse is likely to happen when the explosion occurs near the corner column, and the collapse process would begin with the buckling of the compressive bars in the corner region; 2) If the columns are also set up at the location of frontispieces, the anti-explosion capacity of corner column is reinforced, and local collapse could happen, instead of global collapse; 3) If center columns are designed in the lattice structure, the critical blast pressure is not related with the location of outside explosion; 4) If there is a cantilevered section beyond the span, the anti-explosion capacity of corner columns is obviously smaller than others, and the collapse pattern is identified as a global fall inside the span, and the cantilevered section is turned over.(3) A numerical method for studying the structural response caused by blast loads is proposed, and the main idea of this method is to divide the analysis into two individual steps. In this thesis, this method is employed to study the dynamic response and collapse process of steel frame structure subjected to blast load. Firstly, The Remap tool in AUTODYN is used to simulate the propagation of blast wave, and the reflection pressure on the surfaces of frame members are measured by numerical gauges. Based on the numerical results, the distribution of pressure along member surface is discussed. Secondly, a refined finite element model of a steel frame structure is modeled, and the blast loads measured above are applied on the surface of frame members. The numerical simulation is carried out by explicit solver LS-DYNA, and the structural response and damage induced by blast load are numerically investigated. The results reflect that the blast pressure on the surface of column or beam is much smaller than wall surface involved in same blast scenario, and the pressure around frame joints is much different with other locations. The empirical method provided by TM5-1300 is not accurate enough to predict the pressure on the surface of beam or column. Steel frame structures can survive in a medium-scaled blast event, but would partly damage or progressively collapse if they are subjected to a large-scaled blast load.(4) A numerical method for studying the fragmentation of fragile materials subjected to impact loads is proposed, and the method is employed to analyze the fragmentation of masonry infill wall under blast load in this thesis. The proposed method is established with the principles of fracture mechanics, the theories of micro-crack development and aerodynamics. With the implement of explicit solver, the numerical method can be used to determine the fragment size and launch distance quantitatively. The modified Drucker-Prager criterion is employed to describe the constitutive model of brick and mortar, and the mechanical damage and strain rate effects are taken into account. On the other hand, a homogenized masonry material model is introduced in the thesis, and the efficiency and reliability of the homogenized model is verified through a comparative analysis. Based on several numerical cases with different blast loads, a statistical study of fragment size and ejection distance is carried out. The results reflect that the distribution density function of fragment size can be represented by the generalized extreme value distribution. The distribution of fragment ejection distance is different with different scaled distance. If the scaled distance is smaller, the ejection distance follows the generalized extreme value distribution as well, and if the scaled distance is comparatively bigger, the density function can be represented as the exponential function. Moreover, both the mean value and variance of fragment size are linearly related with scaled distance, and in the case of launch distance, the relationships between mean value, variance and scaled distance can be described as Boltzman equations.