RESPONSE OF STEEL BUILDING FRAMES UNDER FIRE ENVIRONMENT

Abstract

Fires in multistoreyed buildings are occurring at a rapidly increasing rate in the world. The fire safety design of building frames, is a very complex problem involving structural, mechanical, environmental and economical consi derations. There is no established procedure for the Structural Engineer to deal rationally with fire hazard in the design of such structures. The most significant consequences of a rise in temperature due to fire sre, the reduction, of rigidity and strength. As such the estimation of the load carrying capacity depends on. the material-proper ties at temperature met within fire. The response of a structure, when both material and geometric effects are considered, becomes quite complex. Prom the structural engineers point of view, the most important aspect of the problem is that the structure of the building must preserve its load carrying capacity and avoid the possibility of total structural collapse under fire exposure conditions. The temperature reached in a structure during exposure to fire depends on the thermal properties of the material of the structure. As the estimation of the load carrying - 11 - capacity depends on the mechanical properties such as elastic modulus, yield strength, the knowledge of the above properties at elevated temperatures is very essential. Creep of metals at high temperatures is of greatest practical interest. Creep deformation is an important design consideration in any structure required to operate at high temperatures over a long period. The creep effect causes the stress distribution more uniform as the creep is non-linear. Numerical techniques for the analysis of deformation process of steel beams are found in some papers. A few countries like Sweden, Japan, Canada, France permit the use of numerical methods in their building regulations. In the Swedish method the temperature across the section of the beam is assumed to be constant, the beams are not restrained or continuous but assumed to be freely supported, the deflected shape of the beam is assumed to be a sine curve, in the Building research papers, Japan, the temperature is assumed to be constant in all joints and members, creep effects are neglected. British standard specification does not give any guidance on the numerical methods. However, the paper by Harmathy shows the approximate method of determining the deformation for a freely supported beam assuming that the deflected shape of the curve is sinusoidal. But the same - Ill - technique can not be applied for indeterminate-members. No method is available at present to evaluate the deformation process taking into consideration all the above factors. The fire performance of structures at present is given in terms of hours fire endurance or the maximum temperature that the member can withstand. The main interest should be the length of time that the structure can perform its function without either local failure or overall failure. Material non-linearity has to be taken into account due to the non-linear stress-strain relationship and creep laws. During the analysis, the deformed configuration of the structure causes change in geometry, thereby geometric non-LLnearity produces large deflections. Thus when both material and geometric non-linearities are considered the analysis of deformation process can become complex depending om the interaction of the effects as the failure load is approached. As the temperature distribution is not uniform, consideration of variation of temperature across and along the member plays an important role in the prediction of deformation process. The phenomenon of instability due to creep is another factor which may cause local or overall failure. - IV - This investigation is aimed at the study of deforma tion history of steel building frames under fire environment. An algorithm has been developed for the above analysis considering deterioration in the material properties due to rise of temperature, effect of creep and large deformations, sensitivity of strain rate as stress changes and an implicit time marching scheme. Acomputer programme has been developed for the analysis of steel frames, based on the above algorithm using incremental and iterative procedure. Based on such a study, the critical time of failure and the deformation process under fire exposure conditions for the following examples are predicted. Acolumn 1 mx 1 m(plane strain problem) is analysed for the deformation process, for creep and large displacement effects where 'E1 is constant and for fire exposure conditions where »E« changes with temperature /time. The results are compared with unified finite element approach and the comparison can be considered satisfactory. Owing to the deterioration of mechanical properties the column fails due to instability much earlier in the later case. To further check the accuracy of the developed programme, in the absence of sufficient literature, a portal _ T - frame is analysed for two cases (a) 3 elements in each member (b) 4- elements in each member. The results are nearly the same. The same portal frame is analysed for two load cases (i) Failure due to loss of strength (ii) Failure due to instability, in the former case the frame continues to preserve load till the frame leses its strength due to the deterioration of modulus of elasticity fE' to a very less value, thereby the critical time of failure is much more than that of the later case. It is assumed that the variation of temperature across the section, in all the members is constant. A two storeyed building frame is analysed for the fire condition. The progress of stress relief is plotted as a function of time which reveals that the frame fails due to the yield stress at the time of failure reaching a value much below the applied stress. Finally a three storeyed frame from the portion of a building is considered for the analysis. An overall failure takes place at the ground floor column at a time equal to 204 minutes due to the instability. An equivalent frame is selected from the main building frame assuming that the loads acting on the top floors are transferred to the ground floor columns. The frame is analysed for two cases (i) Two - VI elements in each member and (ii) Three elements in each member. It reveals from the diagrams that more number of elements in each member give accurate results. To study the behaviour of frames under ideal condi tions of actual fire (though the relationship between conduct ive and radiative components of an actual fire are unlikely to be produced in the laboratory tests) , each member of the frame is subjected to different temperature variation across the section so that the temperature can be maintained constant, at any instance of time, inside a room. The inside walls, roof and the floor of the rooms are maintained at constant temperature and the outside temperature is maintained at 200° o to 400 p less than the room temperature, at any instance of time. The column sizes are same in all the floors. The analysis is carried out for two and three storeyed frames and the critical time of failure is predicted. Thus, the critical time of failure, deformation process, stress distribution under fire exposure conditions can be predicted for any frame problem depending upon the capacity of the computer.