Temperature Change in Steel-Concrete Composite Bridge: Experimental and Numerical Study

In this article, a prototype empirical section of the composite girder was performed to verify the temperature distributions and the changes in concrete bridges under environmental thermal stresses. T-Beam composite section was placed directly subjected to environmental conditions to endure the change in convection, the convection includes solar radiation, surrounding air temperature and wind speed. In addition, this paper includes the 3D finite-element thermal studies represented by the COMSOL MULTIPHYSIC program, with particular reference to the influence of several parametric studies that represented by changing in concrete thickness, adding asphalt layer and finally change the value of wind speed. The section is prepared with fifteen thermocouples in different places in steel and concrete in the actual and theoretical model to evaluate the temperature distribution inside the composite bridges and their effects in the selected points. COMSOL MULTIPHYSIC showed a good ability to simulate convection, heat conduction and radiation within the surrounding environment. Such as thermocouple (TC3) the minimum temperature was 33.18 o C, 33.29 o C, the maximum temperature was 63.67 o C, 62.77 o C for experimental and theoretical results respectively while the maximum and minimum difference between experimental and FE temperatures is 2.96 and 0.099 for TC3 respectively. Therefore, it can be said that all the thermocouples results showed a good agreement between practical and theoretical results.


Introduction
Continuous bridges girders undergo frequent heating and cooling cycles from the environment that represented by sun radiation, surrounding air temperature and these conditions produce thermal stresses. Thermal stresses effects have often been related to damage of steel-concrete bridge girders. The nonlinear temperature distributions that stand up in bridges are not effortlessly anticipated and are often taken into consideration in a simplified manner, which can also cause structural behavior troubles [1,2]. Exposed structures, for example, bridges, are constantly heat loss and gain from sun radiation, convection and returning radiation to and from the surrounding air, which causes the difference in average bridge temperature and non-linear temperature distributions alongside the depth and the width of the structure [3,4]. Depending on the end situations of the girder, these temperature variations and gradients can also cause longitudinal enlargement and Shrinkage and the thermal loads caused stresses in directions perpendicular to the longitudinal axis. So that, the temperature variant can purpose critical cracking and deflections [4][5][6][7].To calculate temperature gradients in composite bridge girders, the quantity of solar radiation absorbed from the surface, the value of wind speed and the air temperature has to be recognized. The quantity of absorbed solar radiation is a function of the surface sun energy absorption. Wind speed, bridge route and geometry are key elements. Many studies on the impact of wind speed [2,8], bridge route [2,9,10] and geometric formation [2,9,11] had been found in the present-day literature, even as exact parametric studies the effect on temperature differences and distributions were found in concrete -steel girder. The purpose of this study is to show the result of the experimental study, results of the theoretical study by using COMSOL MULTIPHYSIC program and a comparison between practical and theoretical results and finally using several parametric to study the combined effect of changing concrete slab, adding asphalt layer and change the value of wind speed on the temperature gradients in concrete bridges.

Experimental work
Experimental works included casting a segment of composite concretesteel. Fig.1 shows the dimensions and locations of thermocouples in the composite bridge section. The steel beam depth is 500 mm, the width of the flange is 250 mm and the thickness of the web and flange is 8 mm.
Experimental work includes welding I-steel beam in distinct dimensions and with 8 bolts that had been used as shear connectors. After that, the plywood had been used to make the framework of the concrete slab.
The thermocouples sensor is used in this section to degree the temperature distribution and its effect on composite bridge at certain points and the entire number of thermocouples that have been used was (15), (7) in concrete and it symbolizes by (TC), (7) in steel and it symbolizes by (TS) and one between them is symbolized by (TCS).
Finally, the concrete was poured as shown in Fig.2, aspects of concrete slab and steel completely secluded through the use of foam to make the sample as an internal bridge and prevent switching heat from those faces as shown in Fig.3.

Environmental Results:
The environmental data were including measure three factors during July 24 that represent the summer season (sun radiation and daily air temperature are high in this season so; the temperature variance is high and clear). The first factor was surrounding air temperature, its essential in analysing the thermal behaviour of the model because it impacts the radiance of the exterior surface of the surrounding and the convective refrigeration over daylight hours. Fig.4 shows the behaviour of the surrounding air temperature. The second factor was solar radiation and which indicate the external heat source. Fig.5 indicates the sun radiation throughout 24-July. The most crucial factor affecting surface refrigeration by convection was wind speed, so the value must be known to take a look at the temperature gradients all through the day. Fig.6 shows the wind speed in the course of 24-July.

Temperature variance
Four thermocouples are used in this article to show the experimental results two thermocouples are used in concrete (TC3and TC12) and two in steel (TS9 and TS11). In general, Fig.7 & Fig.8 show the temperature variance in concrete during the Sunny day (24 July). The figures show the clear variance is during the Sunny day (24 July). Table 1 indicates the maximum and minimum recorded temperature distribution for TC3&TC12.   While Fig.9 & Fig.10 show that, the thermocouple in steel have jumps and these jumps came when the solar radiation becomes horizontal and falling directly on the steel section. Table 2 showed the maximum and minimum temperature distribution for TS9& TS11.

Finite Element Approach
COMSOL MULTIPHSICYCS Mussa et al. [4] finite element software version 5.5 was used to study thermal transfer analysis of temperature on the girder of a steel-concrete-composite bridge (T-beam) for a sunny day Fig.11 shows the mesh of composite bridge segment and Fig.12 shows the definition of solar location with respect to composite segment . It was assumed that the material properties that were used in the finite element analysis are isotropic and time-independent. From previous studies Emanuel et al. [12], the material properties were set, so that the specified values could efficiently simulate the real heat transfer technique. The Concrete material properties are density ρ = 2400 (kg/m3), thermal conductivity k = 1.5 (W/m.K), and heat capacity at constant pressure (specific heat) c = 900 (J/kg.K). Steel material properties are density ρ = 7850 (kg/m3), thermal conductivity k = 44.5 (W/m.K), and heat capacity at constant pressure (specific heat) c = 475 (J/kg.K). Apply air temperature, wind speed, and vertical and horizontal solar radiation at a distinctive time to represent the initial temperature of the entire bridge body at the start of the FE analysis procedure.
The analysis was started 24 hours before the scheduled time to decrease the effect of the initial temperature on the solution and to avoid error in the COMSOL program.

Comparison between Practical and Numerical Results
The chosen day, July 24, 2020, is used to conduct a comparative study of the temperatures recorded from the experimental composite steel bridge girder section and the temperatures predicted from FE thermal analysis. Four thermocouples are used to show the experimental and theoretical results, two thermocouples are used in concrete (TC3and TC12) and two in steel (TS9 and TS11). In general, Fig.13 to Fig.16 shows the result of the TC3, TC12, TS9 and TS11 thermocouples that showed a good match between the experimental temperature result and the finite element temperature result. For this reason, it can be concluded that the FE model is effective.
The TC3 and TC12 thermocouples located in the middle concrete slab and the top of the concrete slab respectively, Fig.13 and Fig.14 illustrate the comparison between the experimental results and the finite element results. From this figure, the general behaviour of the time-varying temperatures was nearly identical between the experimental results and the FE results. Table 3 shows a compare of the minimum and maximum temperatures for sunny days (July 24) for these thermocouples. The maximum difference between experimental and FE temperatures is 2.96 and 2.34 for TC3 and TC12.   The thermocouples TS9 and TS11 located in the steel web also in different places, Fig.15 & Fig.16 indicates the comparison between the experimental results and the results of FE. The general behavior of the figures showed a good match between practical and theoretical results. Table 4 shows a comparison of the minimum and maximum temperatures for sunny days (July 24) for these thermocouples. The maximum difference between experimental and FE temperatures is 1.78 and 2.54 for TS9 and TS11.

Parametric Study
Several parametric studies were performed to verify the effects of concrete slab thickness, change the value wind and using an asphalt layer on the temperature distribution of the composite section. Presumably, the studied bridges are located in Diwaniyah, Iraq (32 00N, 4455E). Two days were chosen to study the various parametric studies, and these days represent the extremes of weather conditions (24 July 2020 and 15 January 2020), and the same weather data obtained from the experimental work for these days were adopted. For all parametric studies, the average wind speed value was used which is equal to 1.232 (m/s) for the winter and it is equal to 1.712 (m/s) for the summer. Areal part of one of the spans inside the Abu Al Fadl Bridge, as shown in Fig. 17, was approved in the city of Diwaniyah, to study the effect of parametric studies on temperature change.

Effect of slab thickness
Concrete slab thickness in composite bridges is an important factor affecting the temperature gradient within the segment. To study the effect of this parameter, four values were chosen to represent the thickness of the concrete slab 150, 200, 250, 300 mm, with no change in steel dimension. Fig.18 show the temperature change during the day at the upper surfaces of the concrete on a hot day (July 24) and cold day (January 15) respectively. From Fig.18 it can be seen that the temperature of the upper surface is not affected by the change in the thickness of the concrete slab on summer and winter days, because of changing in concrete thickness does not effect on the amount of energy exchanged between the upper surface and the surrounding atmosphere. Fig.19 shows an increase in the effect on the temperature of the bottom surface when the thickness of the concrete slab decreases on a hot day, while this effect decreases on a cold day, and the reason is due to the closing of the distance between the two. So the lowest concrete surface is affected more than other surface. The difference in surface temperature and temperature in the concrete core is an important factor in determining the temperature distribution. Fig.20 indicates the effect of the change in thickness on the difference between upper surface temperature and temperature at 60 % and 80% from depth of concrete slab thickness on a hot and cold day (24 July, 15 January). From Fig.20 and 300 mm, respectively. From the above data for the two figures, a set of observations can be recorded, where the temperature difference rises with the rises in the thickness of the slab, and the maximum difference occurs on a hot day (July 24) at 12:00 noon, as well as the minimum difference that occurs on the same day.    Fig.21 shows the temperature gradients is linear and positive between the top surface and 80% of the concrete slab thickness, at the same time as its miles negative between the bottom surface and 80% of the concrete slab thickness. Fig.22 indicates that the reversed temperature distribution is concave, and the temperatures of the upper and lower surfaces are approximately identical. Table 5 shows the effect of slab thickness at the maximum and minimum average temperature of the composite bridge. The maximum temperature became recorded on July 24 between 14:30 and 15:30, whilst the minimum average temperature turned into recorded on January 15 at 15:15. The distinction between the maximum and the implied minimal temperature is 54.84 ° C, 51.80 ° C, 49.48 ° C, and 47.61 ° C for the thickness of the slab is 150, 200, 250, 300 mm respectively.

Effect of asphalt thickness
Several parametric studies are performed to discuss the effects of the varying convection. One of these parameters was the effect of the asphalt layer. The different values thickness of the asphalt were used and these values were 0, 6 and 10 cm. Zero value of asphalt layer thickness means no asphalt layer, concrete slab is the surface exposed to solar radiation.
The properties of asphalt material are solar absorption coefficient a = 0.92, density ρ = 2100 (kg / m 3 ), thermal conductivity k = 1.1 (W /m.K), specific heat c = 920 (J / kg. K), and emissivity coefficient (ɛ = 0.93). Fig.24 shows the temperature difference of the top surface of the concrete. In summer, it can be seen that the temperatures of the top surface of the sample, which is 6 cm thick, are approximately the same as that of the sample not covered by asphalt, with a time difference between the two curves. The temperatures of a sample covered with asphalt with a thickness of 10 cm are lower than other samples. The reason is due to the greater depth of asphalt compared to another sample which means that the amount of heat that reaches the surface of the concrete is less compared to the sample that has a thickness of 6 cm. During the night, samples covered with asphalt retain heat compared to sample without asphalt, as the upper surface is not subjected to direct cooling at night. In winter, the effect is almost negligible, and the reason is the small amount of solar radiation during this period of the year. Fig.25 and Fig.26 show the temperature gradients with the depth of the bridge installed at 15:00 and 1:00, respectively. Fig.25 indicates the temperature difference inside the concrete slab ranges from approximately 0.5 to 3 °C when comparing the sample covered with a 6 cm asphalt layer with the reference sample (without the asphalt layer), while the difference is greater when comparing the reference sample with the sample covered with an asphalt layer 10 cm long. Fig.25 shows that the temperature difference is concentrated inside the top of the concrete slab. This difference is limited when the sample covered with a 6 cm asphalt layer is compared with the reference sample (without the asphalt layer), while the difference is clear and tangible when comparing the reference sample with the sample covered with a layer Asphalt 10 cm long. Fig.26 indicates the effect of adding an asphalt layer on the maximum and minimum average temperature of the composite bridge. Table 6 shows the average maximum and minimum temperatures for the composite bridge for the different asphalt layers. The differences between the maximum and minimum average temperatures decrease slightly with the increase in the thickness of the asphalt layer.  These results are in accordance with the British standard BS 5400-2 [13], where the presence of an asphalt layer with a thickness of 100 mm reduced the differences between the maximum and minimum average temperatures to 3 °C, while the results of the current study showed the difference between the maximum The minimum average temperature is 4.18 °C for the same thickness (100 mm). AASHTO [14] considered that these differences due to the asphalt layer are small and can be neglected. Fig. 27 shows the eeffect of asphalt layer on the maximum and minimum average temperature of the composite bridge.

Effect of wind speed
The third parameter was to study the effect of wind speed on the temperature gradient. Wind speed affected on the heat transfer coefficient (hc). Three different values of wind speed were used and these values are 0, 2 and 4 (m/s). Fig.28 and Fig.29 shows the effect of wind speed on the temperature change of the top and bottom surface of the concrete slab on (July 24). From the figure it can be seen that the temperature decreases with increasing wind speed where the maximum difference in temperature is 7.78 and 12.60 ° C for upper surface and the difference for lower surface was 2.76 C and 3.67  when comparing the temperature curves of wind speed 2 and 4 (m/s) with the wind speed curve temperatures equal to zero, respectively. The maximum differences between the temperatures curves were appear 15.45 for the top and appear at sunset for bottom surface. Fig.30 shows the impact of wind speed on the temperature gradient of the composite bridge at 14:00 (July 24). From this figure, it could be observed that the maximum difference between the temperature curves occurs at the top surface of the concrete slab, and the temperature gradient will increase with the decrease in the wind speed. Fig.31 shows the effect of wind speed on the average maximum and minimum temperature of the composite bridge, while Table 7 shows the average maximum and minimum temperature of wind speed 0, 2 and 4 m / s. From the above figure and table, it could be observed that the wind speed has no effect in the winter season, but in the hot season, the wind speed is influential in the maximum mean temperature as the average temperature decreases with increasing wind speed.

Conclusion
Based on the data estimated from the composite section analysis under the environmental conditions of Iraq weather using the COMSOL MULTIPHYSICS program and from the practical results, the following conclusions can be drawn:

Experimental Conclusions
1. The maximum air was 50.03 o C while the daily minimum value was 4.09 o C so, the difference between the two values was, 45.94 o C. The maximum solar energy of 24July was 980 (W/m 2 ), while the minimum sun radiation was appeared in cold day and its value was 498.7 (W/m 2 ) 2. The temperature variance is clear during the hot day (24 July) 3. The thermocouple within the steel segment has jumped within the measured temperature, and this temperature increase is throughout the duration of sunrise and sunset due to the exposure of the steel section to direct sunlight.

The Numerical Conclusions
1. The FE model using COMSOL MULTIPHYSIS of the composite section showed good ability to simulate convection, thermal conductivity and radiation within the surrounding environment. Therefore, the thermocouples showed good agreement between FE and the experimental results. 2. The numerical study showed that the thickness of the concrete slab has no effect on the temperature of the top surface, while the effect of the thickness of the concrete slab is noticeable on the temperatures of the bottom surface of the concrete slab and the temperatures inside the concrete slab. 3. Existing asphalt layer leads to an increase in the solar energy gained during the day due to the black color of the layer, but on the other hand, the presence of the asphalt layer with a certain thickness will lead to lower temperatures at the bottom surface of the asphalt layer, so the presence of a medium thickness layer (50 or 60 mm) has a slight effect on the temperatures of the top surface of the concrete slab during the day while the increase in thickness of this layer leads to a decrease in the temperature of the top surface of the concrete. 4. During the day, when comparing the top surface temperature of a composite segment without asphalt layer with the compound section of asphalt layer thickness of 60 mm, the temperature difference is less than 2 °C, while when the layer thickness is 100 mm, the difference is more than 9 °C. While at night, noticed the opposite, as the temperatures are hotter when there is an asphalt layer and the temperature increase during this period with the increase in the thickness of the layer as a result of not cooling the concrete surface by wind when covering the concrete surface with an asphalt layer 5. Increasing wind speed cools the surface, and vice versa. The maximum temperature difference of the upper concrete surface is 7.78 and 12.60 °C when comparing the temperature curves of the wind velocity of 2 and 4 (m/s) with the wind velocity curve temperatures equal to zero, respectively. The influence of wind speed is less on other areas of the compound section. 6. From the parametric study the maximum effect can be obtained during the hot day when the solar energy absorbed by the concrete surface is higher while the steel part is shaded by the overhanging parts, the wind speed is lower, and the air temperature range is high. Based on these conditions, an effective temperature of 60 ° C may be recommended for bridges whose concrete slab thickness is 200 mm or less. While for bridges with concrete slab greater than 200mm, 55 °C is recommended as the effective temperature.