Impact of time after fire on post-fire seismic behavior of RC columns

Abstract

Autogenous recovery after fire exposure is known to be effective on the behavior of reinforced concrete structural members. This is substantially attributed to the variations on the residual mechanical properties of concrete as a function of time after fire. For the first time, in this study, an experimental study was carried out to examine the impact of time after fire on (i) post-fire behavior of small-scale specimens (cubes and cylinders), (ii) seismic behavior of full-scale flexure-controlled reinforced concrete columns. Outcomes of the post-fire seismic test results were investigated on a total of four columns one of which was an unheated reference specimen, whereas the other three were heated for 90 minutes following the ISO-834 fire curve. Seismic tests were carried out on fire exposed columns 30, 60 or 360 days after fire exposure. The test results indicated that 90 minutes of fire exposure caused a reduction in lateral load capacity, ductility and stiffness of the columns with respect to the unheated column. On the other hand, impact of time after fire exposure on residual lateral load capacity and ductility of the columns was found to be limited while the column subjected to seismic test 30 days after fire exposure exhibited less stiff behavior with respect to the columns tested later. Furthermore, an analytical study is conducted for the prediction of seismic behavior of reinforced concrete columns after fire exposure considering the variations in residual properties of concrete by time, and the proposed model is found to be in good agreement with the test results.

Introduction

Reinforced concrete (RC) columns are known to lose a fair amount of compressive strength and stiffness after being exposed to fire [1]. However, the loss of residual mechanical properties of concrete due to fire exposure is largely recoverable in the long-term [2]. Therefore, autogenous recovery of concrete as a function of time after cooling is of paramount interest. Post-cooling behavior of concrete can be separated into two stages [3]. At the first stage, after cooling to ambient temperatures, it is observed that the compressive strength and elastic modulus of concrete may be further reduced from its original strength under high temperatures until a period of time because of continuing disintegration of the microstructure in concrete. This first stage of post-fire cooling period is reported to be approximately 30 days from the beginning of the cooling [4], [5]. Because of the reaction between the concrete and moisture in the air (needed for re-hydration of cement), after around 30 days, recovery in compressive strength begins referring to the second stage of cooling period and concrete can recover up to 90% of its original strength in one year after being subjected to temperatures of 500 °C [4], [6]. This is attributed to the fact that capillaries initially filled by C-S-H gels are opened during fire exposure and refilled by rehydration products due to the moisture in air owing to the carbonation process [2]. Hence, this rehydration process reduces the porosity of concrete and results in strength recovery. However, there is not a consensus between the reduction ratios in compressive strength depending on the time after fire due to the fact that residual compressive strength after fire may be affected by various parameters such as concrete composition, heating and cooling scenario, specimen size, axial restraint. For instance, Harada et al. [4] indicated that residual compressive strength decreases up to 50% of the initial value 30 days after fire, and this reduction ratio does not further vary until 60 days after fire exposure. On the other hand, Papayianni and Valiasis [5] mentioned that the compressive strength would reduce to about 41% of the original value 30 days after being subjected to fire and then around 15% enhancement on compressive strength is expected for the next 30 days (at the age of 60 days after fire, residual compressive strength ratio is about 48% of the original value). Residual mechanical properties of reinforcement are also effective on the behavior after fire. Cooling would often restore the material to its original state for temperatures up to 450 °C for cold work and 600 °C for hot rolled steel reinforcement [6], [7] irrespective of time after fire. As can be drawn from the findings of the literature review, the load-bearing capacity of a structure tends to decrease after the maximum temperature is achieved, then reaches to a minimum value, and finally recovers partially or completely after cooling [8]. These variations in load-bearing capacity due to the thermal inertia and additional loss of material mechanical properties during cooling stage, may result in structural failure during the cooling phase of a fire rather than under elevated temperatures [9], [10], [11]. This phenomenon is one of the major reasons that formed a motivation for the present study.

In other respects, performance of any building subjected to fire should be inspected considering probable future earthquakes in earthquake prone areas. Weaknesses in RC structures arising from insufficient lateral load capacity and stiffness due to low concrete compressive strength and substandard detailing of reinforcing bars, result with vulnerable buildings during earthquakes [12]. Fire damage may also cause inadequate concrete compressive strength even for well-engineered structures, and can lead to vulnerable building classes as commonly exist in developing countries. Keeping this issue in mind, post-fire seismic behavior of RC structural elements designed according to current guidelines is simply unknown up to date. Majority of the existing studies considered the behavior after cooling under only service loads combined with uniaxial/biaxial bending [13], [14], [15], and the studies investigating fire-damaged RC columns under seismic loading have not been extensive. Studies conducted by Yaqub and Bailey [16] and Bailey and Yaqub [17] included seismic retrofitting of shear-critical RC columns after fire exposure while Mostafaei [18] pointed out hybrid fire testing principles on a single high strength concrete column under monotonic lateral loading. None of these studies considered the post-fire seismic behavior of conventional flexure-controlled RC columns. Furthermore, impact of autogenous recovery in concrete after fire on the post-fire seismic behavior of RC columns has not been investigated yet neither for shear nor flexure-controlled columns. As mentioned above, structural performance assessment of a fire-damaged structure may lead to misleading results due to the reduction in compressive strength of concrete particularly in the initial days after fire exposure. Therefore, this study presents the experimental findings on the effect of fire damage on seismic behavior of RC columns for the first time with a particular focus on the impact of autogenous recovery of concrete with time after cooling. In the scope of this study, uniaxial compression tests were carried out on cylinder and cube specimens in order to determine the variation of compressive strength depending on the time after fire. Also, four full-scale RC columns were designed and constructed complying with the major design codes (e.g. [19]). The columns except the reference one, were exposed to ISO-834 standard fire [20] for 90 minutes, and then kept in ambient conditions for 30, 60 or 360 days until the test day. The columns were then subjected to constant axial loads combined with reversed cyclic lateral displacements simulating seismic loading. The responses of the columns were analyzed in terms of lateral load-displacement relationships, ductility and stiffness. However, the column tested 360 days after fire experienced premature buckling of longitudinal bars due to concrete settling problems during casting process and therefore, only the hysteretic response of this column was given in the study, and it was excluded from further discussion. As aforementioned, there is an obvious inconsistency in the literature in terms of the recovery of concrete with time after fire, which is particularly encountered during the two months after fire exposure. After this period of time, it is generally accepted that the recovery rate of concrete almost proportionally increases with time after fire, and concrete after fire is expected to reinstate its original strength after one year [4], [6]. Therefore, it should be noted that the column tested 360 days after fire can be deemed to be a less critical specimen in the test matrix. A combined thermal and structural analysis was also conducted to predict the post-fire seismic response of the columns considering the variations in mechanical properties of concrete by time after fire. The results of the analytical work demonstrate that the generated algorithm within the current study is in good agreement with the test results. Demonstrated applicability of the given algorithm within the manuscript in terms of prediction of post-fire seismic response considering the compressive strength variation with time, may serve to the prediction of the response of the column tested 360 days after fire as well.