Effect of Wall Thickness on Thermal Behaviors of RC Walls Under Fire Conditions
- First Online:
- Cite this article as:
- Kang, J., Yoon, H., Kim, W. et al. Int J Concr Struct Mater (2016) 10(Suppl 3): 19. doi:10.1007/s40069-016-0164-5
- 1.2k Downloads
The objective of this paper is to investigate the effect of thickness and moisture on temperature distributions of reinforced concrete walls under fire conditions. Toward this goal, the first three wall specimens having different thicknesses are heated for 2 h according to ISO standard heating curve and the temperature distribution through the wall thickness is measured. Since the thermal behavior of the tested walls is influenced by thickness, as well as moisture content, three additional walls are prepared and preheated to reduce moisture content and then tested under fire exposure. The experimental results clearly show the temperatures measured close to the fire exposed surface of the thickest wall with 250 mm thickness is the highest in the temperatures measured at the same location of the thinner wall with 150 mm thickness because of the moisture clog that is formed inside the wall with 250 mm of thickness. This prevents heat being transferred to the opposite side of the heated surface. This is also confirmed by the thermal behavior of the preheated walls, showing that the temperature is well distributed in the preheated walls as compared to that in non-preheated walls. Finite element models including moisture clog zone are generated to simulate fire tests with consideration of moisture clog effect. The temperature distributions of the models predicted from the transient heat analyses are compared with experimental results and show good agreements. In addition, parametric studies are performed with various moisture contents in order to investigate effect of moisture contents on the thermal behaviors of the concrete walls.
Keywordsnormal strength concrete wall moisture content wall thickness fire test heat transfer analysis
Reinforced concrete wall structures are widely used in Korea, especially for residential buildings. It is important to study the fire resistance of reinforced concrete walls, because the number of fire accidents in residential buildings is increasing every year. In Korea, concrete walls must be designed according to Korean building code specifications to meet required fire resistance ratings, which is determined by occupancy of the building and number of floors (Regulation for refuge and prevention of fire in building 2015). The fire resistance rating determines required length of time period that a structural member or system is able to function properly under fire. To design the structural members to resist fire, engineers choose a section for structural member from tables that can satisfy the required length of time period. The US and Europe provide guidelines similar to Korean building code in ACI 318-11 (2011) and Eurocode 2 (2006), respectively. Specifically, ACI 318-11 (2011) provides the chart for determining the fire resistance rating of concrete walls based on thickness and concrete type. Eurocode 2 provides a chart of minimum thickness of walls and cover based on the required fire resistance rating, which is determined by the number of exposure surface and loading condition.
However, quantitative evidence supporting the relationships between the thickness of a concrete wall and the resistance to fire has not been reported. In addition, a limited number of studies have been performed on structural walls, especially normal strength concrete walls, compared to other structural members such as columns and beams. Recent studies on reinforced concrete wall under fire conditions have focused on spalling in high strength concrete walls (O’Meagher and Bennetts 1991; Crozier and Sanjayan 2000; Ngo et al. 2013; Lee and Lee 2013). O’Meagher and Bennetts (1991) developed a theoretical computer-based model for analyzing the behavior of concrete walls under fire exposure. From their theoretical study, they concluded that the behavior of a reinforced concrete wall in one-sided fire exposure could be significantly influenced by the height to thickness ratio of the wall, the location of steel reinforcement, associated cover thickness, and end restraints. Crozier and Sanjayan (2000) focused on the behavior of slender high strength concrete walls subjected to standard fire conditions under combined eccentric axial and lateral loads to investigate the in-plane load capacity. According to their test results, slender walls subjected to low in-plane loads show that the in-plane loads made little difference on the behavior of the walls exposed to fire. Their study showed that concrete strength seems to have little influence on the in-plane load capacity of walls subjected to fire exposure on one surface. Ngo et al. (2013) found the differences in thermal behavior between normal and high strength concrete walls for ISO 834 standard fire testing. Specifically, it was observed that a hydrocarbon fire creates a significant increase in temperature in the initial period of the fire as compared to that specified by the ISO 834 standard. In addition, high strength concrete walls suffered severe damage due to spalling than normal strength concrete walls. Therefore, the hydrocarbon heating caused more explosive and earlier spalling in the high strength concrete walls than heating according to ISO 834 standard fire exposure. (Lee and Lee 2013) conducted experimental and numerical studies of the fire resistance of concrete walls exposed to fire on both sides. They tested eight walls with various height-to-thickness ratios, concrete strengths, and axial load levels. Test results showed that wall thickness and load level have a major effect on the performance of concrete walls in a fire. They proposed numerical methods for predicting the fire resistance of concrete walls. In their studies, the fire resistance of concrete walls was predicted based on the experimental observations of columns exposed to elevated temperatures.
The above review indicates that the behavior of concrete walls under fire exposure is affected by the wall thickness, magnitude of in-plane and lateral loads, concrete strength, and reinforcement ratio. However, no studies are reported on the quantitative relationship between the wall thickness and fire performance of concrete walls.
Therefore, the aim of this paper is to investigate the thermal behaviors of load bearing normal strength concrete walls with various thicknesses when one surface exposed to fire. In addition, the effect of preheating on the thermal behavior of concrete walls is investigated to evaluate the effect of moisture content on the performance of the walls.
Toward this goal, fire tests are performed on concrete walls having different thicknesses and amounts of moisture. In the experiments, temperature distributions through the thickness of the wall specimens are measured. In addition, finite element models are developed to simulate fire tests and the results from the models are compared with experimental results.
2 Experimental Program
2.1 Test Specimens and Variables
Summary of test conditions of walls.
Wall thickness (mm)
Amount of moisture
Mix proportions of concrete.
Specified concrete strength (MPa)
2.3 Test Set-up and Data Measurement
This study uses the deflection failure criterion suggested in ISO 834-2012 (2012) for the structures exposed to high temperatures. According to ISO-834, failure is determined when the axial deformation of the compression member reaches h/100 mm or the rate of deformation reaches 3 h/1000 mm/min within 2 h of fire, where h denotes the initial height of the specimen before the fire. For the walls in this study, the deflection failure limit is 6 mm.
3 Results and Discussion
3.1 Experimental Results
3.1.1 Wall Specimens Having Various Thicknesses
After about 20 min of heating, moisture in the concrete of W15 specimen begins to evaporate and steam starts to seep out through the unexposed surface. A similar phenomenon is observed with the thicker specimens, W20 and W25, but the moisture evaporation starts much later than after 20 min of heating. In addition, the largest amount of moisture flow is observed in W15 because it is easier for the moisture to migrate through lower thickness of W15 than through thicker walls W20 and W25.
Temperatures at C1 after heating for 2 h.
Temperature at points (°C)
Times taken for thermocouple C1 to reach various temperature levels.
Increasing time to reach at various temperature level (min)
3.1.2 Preheated and Non-Preheated Wall Specimens
However, temperatures at locations C2 and C3 of the preheated specimens are higher than those of non-preheated specimens. This is because heat is prevented from being transferred from the fire toward the opposite surface due to moisture clog inside the walls when the specimens are not preheated. In other words, preheated walls have little or no moisture clog, which leads to a better heat transfer through the thickness of the walls.
As noted above, the thickest wall having 250 mm thickness experiences the highest temperature measured at the surface exposed to fire and the shortest time to reach any temperature level, especially when the concrete has greater moisture content. The reason could be due to the moisture movement in concrete wall when exposed to one side fire.
3.2.1 Analytical Approach
The moisture content in concrete has a significant influence on thermal properties such as thermal conductivity and specific heat of normal strength concrete (Kodur 2014; Kodur et al. 2008; Szoke 2006). Therefore, it is important to consider the moisture effect in predicting thermal behavior of concrete structures under fire.
Only a few researches using numerical analysis or experimental studies on moisture migration and pore pressure development have been carried out, and they are mostly related to the spalling of high strength concrete (Beyea et al. 1998; Consolazio et al. 1998; Bazant and Thonguthai 1979; Khoylou 1997; Selih et al. 1994; Dwaikat and Kodur 2009; Kodur and Phan 2007). A practical modeling technique for predicting temperature distributions of concrete walls under fire considering movement of moisture has not been fully developed.
In this study, 3D finite element models of all the specimens are generated to simulate fire experiments using ABAQUS 6.10-3 (Theory Manual 2010). The model uses 8-node linear brick elements for the concrete walls and the reinforcing bars. Even though the experimental results show one-dimensional heat propagation through the thickness, the 3D models are generated in order to use the temperature distributions predicted from the transient heat analysis for the further studies of mechanical analysis of the fire damaged walls. In the model, temperature-dependent thermal properties such as effective specific heats and conductivities of concrete are adopted from Eurocode 2 (EN 1992-1-2). Time dependent temperatures of ISO standard heating curve are prescribed to one surface of each wall model, while initial temperature is given as 20 °C for all the surfaces. Detailed modeling methods can also be found from the previous study of Choi et al. (2012).
For preheated walls having 200 and 250 mm under 300 °C, the thermal conductivity of moisture clog is determined as 50 % of the conductivity for plain concrete because moisture is actively evaporated under 300 °C and delays heat transfer when the concrete reaches over 300 °C. Conductivity is increased by 60 and 20 % for W20V and W25V, respectively. For the W15 model, the thermal conductivity of the moisture clog is 50 % of the conductivity for plain concrete regardless of temperatures in order to change the temperature from that of W15 V beyond 300 °C.
3.2.2 Analysis Results
In the Figs. 13a and 13b, analytical results of temperature distributions for the case of walls having 150 mm thickness are compared with experimental results. In the W15 V model, it is assumed that the moisture clog is not formulated due to moisture evaporation during preheating process, while the W15 model includes moisture clog zone. The assumption is supported by the experimental results showing that the heat passes relatively easily in the preheated wall than the non-preheated wall. Since the temperature distributions predicted from the non-preheated and preheated models having thickness of 150 mm are well matched with the experimental results, the assumption can be considered as acceptable. Figures 14 and 15 show the time–temperature curves of walls having thicknesses of 200 and 250 mm, respectively. In these models, it is assumed that the moisture clog zones are formulated for the preheated walls as well as the non-preheated walls. This assumption is made because the experimental results show that the moisture of the preheated walls is not fully evaporated during preheating when the specimens have thicknesses of 200 or 250 mm. Because moisture clog zone delays heat propagation, the predicted temperatures in the dry zone increase rapidly during heating, while temperatures behind the moisture clog zone are not increased effectively. Such delay in heat propagation becomes significant in the models of W25 and W25V, as moisture content increases with wall thickness. The analytical results of the W20, W20V, W25, and W25V models having the moisture clog zone are in good agreements with the experimental results and justify the assumption.
To investigate the effects of wall thickness and moisture content in concrete on the temperature distribution, six walls are heated on one side according to the ISO 834 standard fire curve. The wall thicknesses are varied as 150, 200, and 250 mm and half of the fabricated walls are preheated before the fire tests to reduce the amount of moisture. Cross sectional temperatures are measured at three different locations through the wall thickness during heating. The experimental results clearly show the temperature measured close to exposed surface is higher in the thicker walls than the temperature measured at the same location of the thinner wall. This is because the moisture clog inside thick walls with 200 and 250 mm thicknesses prevents heat being transferred and this acts as insulation. This is also confirmed by the experimental results of temperature distribution of the preheated walls, showing that better heat transfer occur from the preheated walls as compared to non-preheated walls. Finite element models are developed to simulate fire tests and the results from the model are compared with experimental results.
The proposed analytical modeling technique having moisture clog zone is able to predict thermal behavior of normal strength concrete walls under fire with consideration of moisture contents. However, accuracy from the model can be increased when the better information on the moisture clog zone and moisture dependent thermal properties of concrete are included the model. Therefore, further studies are needed to investigate the position of moisture clog zone in the real sized structural members and to see how the zone moves as heating time passes.
This research was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (No.2-2013-3322-001-1). The research was also supported by Korea Agency for Infrastructure Technology Advancement (KAIA)—16AUDP-B100356-02.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.