www.seipub.org/ce Construction Engineering (CE) Volume 2, 2014 Thermal Characteristics of Concrete in the Vicinity of Embedded Cooling Water Pipe Yulin Lu *1,2, Xiaoran Chen 3 * 1. Institute of Engineering Mechanics, China Earthquake Administration, Heilongjiang Harbin, China 2. Department of Disaster Prevention Engineering, Institute of Disaster Prevention, BeiJing, China 3. College of Resources, Hebei University of Engineering, Hebei Handan, China *1 yllu@cidp.edu.cn; 2 lylcxr@163.com; Received 7 March 2014; Accepted 31 March 2014; Published 14 May 2014 2014 Science and Engineering Publishing Company Abstract This research is to reveal the thermal characteristics of concrete with the cooling water pipe. The temperature distribution is simulated by the finite element method, and an approximated method and a coupling method are used to analyze the concrete three-dimensional temperature field. The results of two methods show that the concrete temperature contours distribution in the vicinity of water pipe are similar, but the maximum temperature is different. The monitoring point temperature compared reasonably well at the same hydration time in these two methods, the coupling method temperature is larger than results of approximate method, and the maximum temperature difference is 0.9. The influence of cooling w ater on temperature distribution is achieved by loaded the convective heat transfer coefficient in approximate method, thus the coupling method results can be better quantitative analysis the concrete temperature distribution and the effect of the cooling water than approximate method. These results will provide a data support for study the concrete thermal field with cooling water pipe system. Keywords Temperature; Concrete; Cooling Water; Approximate Method; Coupling Method Introduction The concrete structure is a very popular structural type in the civil engineering, such as raft foundation, retaining wall and massive dam which are almost built by this material. These structures usually require lots of concrete and mass of heat will release in the cement pouring process. A major factor generated heat is the cement by hydration, then the higher temperature will gather in the concrete. Temperature cracks on concrete are often produced by this temperature difference that existed inside and outside, it not only affect the structure safety, but also affect its durability and waterproof. Therefore, it necessary to take effective measures to control the temperature changes to prevent the cracks in the hydration time. There are many different reducing and preventing the thermal cracks methods are used in the engineering. Two types of technology, about water cooling and air cooling pipe embedded in concrete can reduce the temperature. The cooling water or air can significantly reduce the temperature near the pipe, and these results are confirmed by previous study. So the quantitative analysis of a hydration heat problem with pipe cooling by circulating cooling water or air has been developed a complex problem, and how predict the thermal field distribution is also an essential part of the control temperature cracks. The numerical method is a valuable tool for predicting temperature in concrete during the cement hydration. However, the temperature distribution may be very complicated due to the existence of cooling water in the pipe. Two methods contains the approximate method and coupled fluid flow and solid heat transfer method (coupling method) have been successfully used in this research to solve this problem. The approximate method has been employed in flow transfer heat, which is applied the average convective transfer coefficients on interface of the concrete and cooling pipe. This method is widely used to compute the temperature field distribution and predict the temperature cracks. It have some advantages, such as simple calculation and less computed time. With these advantages, it also brought some troubles, for instance, the results calculated by this method were not match well with the factual results in both the temperature contours near the water pipe and the individual point 10
Construction Engineering (CE) Volume 2, 2014 www.seipub.org/ce temperature, because the convective coefficient distribution along the pipeline was not a constant. The another numerical method is the coupling method based on the discrete water pipe actual physical model, which is a more precise way to solve the water flow between the concrete and pipe compared with the approximate method. In the discrete pipe model, the thermal impact of water flow on temperature distribution can be calculated analytically with the water pipe system configurations. It focused on the detailed interaction of complicated water flow in the pipe with the heat transfer in the adjacent concrete by simultaneously treating the coupled water flow and heat transfer, so that the heat energy balance between the water and the concrete could be attained automatically and accurately. This method not only reflect the temperature field is whether affected by the dynamic water flowing and the bending pipe curvature, but also avoid solving the complex convective coefficient between the cooling water pipe and the ambient concrete. This present study focused on analysis of hydration heat in concrete structures with cooling pipe to quantify its thermal characters. The three-dimensional temperature distribution in the concrete was simulated by the finite element method, the temperature field was also calculated by the approximate method and the coupling method. Two numerical results would compare with the same conditions, and the difference of results would analyze in this study. The results data would provide a help for predicting the thermal cracks in concrete hydration process. Mathematical Model of Temperature Therapy The concrete could be seen as a continuous medium with internal heat source in the concrete hydration heat period, the three-dimensional transient heat transfer equation was as seen the equation (1). 2 2 2 2 ( T T T k + + ) = cγ T q (1) 2 2 2 2 x y z t Where, T was concrete transient temperature; k was concrete thermal conductivity; c was concrete specific heat; γ was concrete bulk density; q was the heat rate generated by unit volume concrete. Physical Model of Concrete In the concrete engineering, the water pipe was usually arranged for S type, which connected with horizontal pipe and vertical pipe. The water cooling system was embedded in the concrete, it can be seen Fig.1. The horizontal pipe spacing was 300 mm, and the length of pipe was 1000 mm. The water pipe parameters were as follows, pipe diameter was 30 mm, and the bend angle was 90 at the bend junction. The monitoring points were used to analyze the concrete temperature change in the vicinity of cooling pipe. FIG.1 GEOMETRIC MODEL OF WATER COOLING SYSTEM Numerical Simulation and Methodology Approximate Method In this method, the water cooling effect on concrete was simulated by the convective heat transfer coefficient, Nusselt number could be defined as the strength parameters of convective transfer heat, which reflected the flow and heat exchange between the cooling water and the concrete. Therefore, the Nusselt number calculation can be expressed equation (2)and (3). Re = vd / υ (2) ; 0.8 0.4 u 0.023 e r N = R P (3) Where, Re was Reynolds number; v was water velocity; υ was viscosity of water; d was diameter of the pipe; Pr was the Prandtl number. Coupling Method For this numerical method, the concrete was modeled as solid region and the water was modeled as fluid region, the conservation of mass within these regions was expressed by the continuity relation in equation (4). div( U ) = 0 (4) The momentum equation was expressed by the equations (5). 11
www.seipub.org/ce Construction Engineering (CE) Volume 2, 2014 div( ρ u U ) = P + div( µ gradu ) + S (5) i i i The energy equation was given by the equation (6). ρ c div( TU ) = div( kgradt ) +Ψ (6) b Where, ρ was density; U was velocity; ui(i=x,y,z) was the velocity components; T was temperature; k was the thermal conductivity; cb was heat capacity; Si was the generalized source; µ was the viscosity and the concrete viscosity was infinity. Heat Source The heat was generated by the concrete hydration, and so every position in the concrete had the same generate heat rate. The equation as follow q= dq / dt = mq0e mt (7) Q was the concrete hydration heat; t was the hydration time; Q0 was the finial hydration heat; m was the hydration coefficient. Material Properties Thermal properties of the concrete and the water used in this study were shown in Table1. TABLE1 PHYSICAL PARAMETERS OF CONCRETE AND WATER Parameters Material Water Concrete Density (kg/m3) 998.2 2100 Specific heat (J/kg ) 4183 960 Thermal conductivity (W/m ) 0.599 1.74 Viscosity (m 2 /s) 1.006 10-6 0 Nusselt number (40cm/s) 82 0 Boundary Conditions In the approximate method and the coupling method, it had some assumptions in the simulation were as follows. (1) The water flow was assumed to be turbulence and the water was taken as an incompressible Newtonian fluid with constant density and viscosity. (2) The pipe was an rigidity material and not deformed. (3) The pipe wall thickness was very thin and not considered its effect on the temperature distribution. The boundary conditions for the simulation were as follows. The thermal boundary condition of the cooling water was maintained at 20 in these two methods. The concrete initial temperature also was 20. In the coupling method the inlet velocity of the water was uniform at 40 cm/s, and at the outlet of the pipe the reference pressure was set as zero. All other surfaces of the water pipe domain were considered noslip boundary conditions along the pipe wall that the velocity components were set as zero. In the approximate method the Nusselt number was loaded on the coupling area of the concrete and pipe to simulate the cooling effect. Methodology The geometric model was meshed with eight-node hexahedral element in these two methods. Mesh independence was assessed by comparing the temperature distribution in the finial working mesh, after increasing the number of elements by 10%, the maximum temperature values changed by less than 1%. In this study, about coupling method, the finite element model consisted of 455426 computational fluid elements and solid heat transfer elements, and the total the nodes were 78863, which ensured that the calculation results were independent of the finite element mesh. The mesh of fluid near the pipe wall was refined to suit the boundary flow and the solid mesh near the pipe was also refined to capture the rapid temperature variation in the vicinity of the cooling pipe. In the approximate method, the concrete was meshed 280071 solid heat elements and the nodes were 50732, and the mesh grid was also refined near the water pipe. The thermal field of concrete was calculated by finite element method and the multi-physics simulation tool ANSYS was chosen in present study. In the approximate method, the temperature field was calculated through the thermal analysis and the water cooling effect was simulated by applied the Nusselt convective coefficient. In the coupling method, the water velocity and pressure were solved first, and then the transient temperature field of concrete was solved with the flow solutions. When solving the water flow equations, the SIMPLEF algorithm, which has been extensively validated, was used for velocity-pressure coupling. A Tri-Diagonal Matrix Algorithm (TDMA) solver was adopted for the discretized momentum equations, whereas a Pre-Conditioned Conjugate Residual (PCCR) solver was used for the pressure equation. After the convergence of the flow field calculation, a transient conjugate heat transfer process in both solid concrete and fluid water was simulated with a PCCR solver. The initial time step was 0.1h, and the convergence criterion was kept at 10-4 ~10-3, which ensured the validity of the numerical results. Results Fig.2 showed that the concrete thermal characters in the vicinity of the cooling water pipe with different compute methods. It could be seen that the 12
Construction Engineering (CE) Volume 2, 2014 www.seipub.org/ce temperature contours distribution in approximate method were very similar to the results of coupling method, but the maximum temperature was discriminated. The contours gradient of temperature changed obviously which were located near the pipe, and the temperature around the pipe was attenuated and the curves density decreased gradually from the pipe to the concrete. were larger than the results of approximate method and the value was 0.9, this was due to the difference of the heat transfer coefficient which was employed on the area which connected concrete and water pipe. An average convective coefficient was loaded to simulate the water cooling effect in approximate method, but in coupling method, this effect was computed though the interaction of complicated water flow in water pipe with the heat transfer in the adjacent concrete. Therefore, the cooling water convective along the pipeline was not a constant and it was changed with the water velocity. (a) COUPLING METHOD RESULT Z=0 (a) (b) APPROXIMATE METHOD RESULT Z=0 (c) COUPLING METHOD RESULT Z=0.5 (b) (d) APPROXIMATE METHOD RESULT Z=0.5 FIG.2 TMEPERATURE DISTRIBUTION OF TWO METHODS Fig.3 showed that the temperature changes at different hydration time in two methods. From these figures, we could see that the results of the coupling method compared reasonably well to the approximated results in temperature profiles over the same hydration time range. The highest temperature appeared at the hydration time of 24h, the temperature gradient increased quickly before this time range and then the temperature declined gradually. Ultimately, they were more likely to reach a steady heat transfer state after the hydration time of 240h. Moreover, it was noteworthy that the temperatures of coupling method (c) (d) FIG.3 TEMPERATURE COMPARED AT DIFFERENT HYDRATION TIME 13
www.seipub.org/ce Construction Engineering (CE) Volume 2, 2014 Fig.4 was the result of coupling method that described the velocity changed along the pipeline, the inlet velocity was the default 40 cm/s, and it declined after the bend junction. In this method, the flow field was turbulence and the fluid direction would change when the water through the junction which connected the horizontal and vertical pipe, so the velocity would be decreased by this bend pipe. However, in the approximate method the convective coefficient was assumed the steady constant accord with the velocity, it was all the initial value of 40 cm/s, this means that the cooling water convective transfer heat was the same in anywhere of the water pipe system. This convective effect difference would lead to the thermal distribution around the water pipe and temperature sizes were not same, especially with the increased the number of the junction in the cooling system. It can be seen that the velocity declined with the angle increased, the losing maximum difference was 9.08cm/s (see Fig.5(a)). Fig.5(b) was the velocity losing compared with velocity constant results of the approximate method, the losing could affect the heat exchanged between concrete and water, therefore, temperature results of coupling method would larger than approximate method results. FIG.6 CONCRETE TEMPERATURE COMPARED WITH TWO METHODS FIG.4 VELOCITY OF WATER ALONG PLPELINE (a) VELOCITY DISTRIBUTION In order to verify the accuracy of the two methods numerical results, the concrete temperature difference at different hydration time and position were compared in Fig.6. These temperature difference followed different laws along the Z axis. At the position of z=0 m, the maximum value was 0.77, appeared at the hydration time of 24h, the difference declined gradually with the increased time. At the position of z=0.5 m, the value was, 0.89the difference increase at beginning time and decreased after the time of 48h, this was due to the position far away the water pipe and the cooling water effect was not obviously. At the position of z=0.75 m, this position almost near the concrete outside and the temperature difference was basically nothing, especially after the time of 100h. Conclusions This study investigated the thermal distribution with the cooling water pipe embedded in concrete and the cooling water effect was included in the employed FEM by using the approximate method and coupling method. Temperature of concrete at the different time was compared and thermal distribution around the water pipe was analyzed. (b) VELOCITY COMPARISON FIG.5 WATER VELOCITY DISTRIBUTION AND COMPARISON In order to quantitative analysis the velocity losing along the pipeline in coupling method, the velocity of the bend junction position was calculated in the Fig.5. Firstly, temperature field size calculated by these two methods were almost same, but the maximum temperature of concrete was not same, the coupling results were larger than results of approximate method. Secondly, the cooling water effect on thermal field was mainly dependent on the convective 14
Construction Engineering (CE) Volume 2, 2014 www.seipub.org/ce coefficient, the coupling method was more accuracy due to the coefficient was computed by the water flow. Thirdly, the Nusselt number was not same because of the cooling water velocity losing. The difference of temperature and velocity were all lower in coupling method. The coupling method concrete temperature results were more conservative than approximate method, so this method would better predict the thermal distribution in the hydration process and the results were safe. This study results would provide a reference for research the concrete temperature field in future work. ACKNOWLEDGEMENTS This research is funded by Teachers' Scientific Research Fund of China Earthquake Administration 20120103 and Special Fund of Fundamental Scientific Research Business Expense for Higher School of Central Government (Projects for creation teams) ZY20110102 and ZY20120103. REFERENCES Ballim Y. A numerical model and associated calorimeter for predicting temperature profiles in mass concrete. Cement and Concrete Composites, 2004, 26(6): 695 703. C.-X. Huang. The three dimensional modelling of thermal cracks in concrete structure. Materials and Structures, 1999,32:673-678 Hans Hedlund, Patrik Groth. Air cooling of concrete by means of embedded cooling pipes- Part I: Laboratory tests of heat transfer coefficients. Materials and Structures, 1998,31:329-334 Jin Keun Kim, Kook Han Kim, Joo Kyoung Yang. Thermal analysis of hydration heat in concrete structures with pipe-cooling system. Computers and Structures, 2001, 79:163-171 Kovler K L,David A,Stang H.Early age concrete-- properties and performance.cement and Concrete Composites,2004, 26(5): 413-415 Patrik Groth, Hans Hedlund. Air cooling of concrete by means of embedded cooling pipes- Part I1: Application in design. Materials and Structures. 1998,32:387-392 Shiming Yang; Wenquan Tao. Heat Transfer, Beijing, High Education Press,2006 T. G. Myers, N. D. Fowkes, Y. Ballim. Modeling the Cooling of Concrete by Piped Water; Journal of Engineering Mechanics; 2009,135(12): 1375-1383 Wang Jiachun, Yan Peiyu. Analyzing of cracking risk of early-age concrete structure. 2006, 14(2): 262-267(In Chinese) Yulin Lu, Tao Lu, Li Wang, Zhenyu Wang, Peipei Zhao. Numerical study on thermal field in concrete with cooling pipe based on the coupling method. Applied Mechanics and Materials;2011,94-96:388-392 Yulin Lu, Xiaoran Chen, Jinli Ding. Analysis of concrete temperature in early age with field test and numerical simulation methods. Journal of Beijing University of Technology, 2013,39(12):1843-1848 Yulin Lu, Yongduo Liang, Gaihua Yu, Wenqian Li, Li Wang. Numercial calculated the 3D temperature field of concrete conjugated conduction method. Sichuan Buliding Science, 2013,39(1):317-321 Yulin Lu;Jiaqi Li;Li Wang. Influence analysis of the concrete temperature field with different computed methods. Advanced Materials Research; 2012,446-449:517-521 Yun Lee, Myoung-Sung Choi, Seong-Tae Yi, et al. Experimental study on the convective heat transfer coefficient of early-age concrete. Cement & Concrete Composites, 2009,31:60-71 Yulin Lu was born on December 21, 1983, Beijing. He has a background in civil engineering and mechanics courses at the Beijing University of Technology with master level. Currently he is a student of Institute of Engineering Mechanics in China Earthquake Administration for doctor degree in rock and soil engineering. Yulin Lu is also a teacher on Institute of Disaster Prevention, and the research field is Multi-physics fields coupling calculation in engineering, and has published more than 15 scientific papers. Xiaoran Chen was born on December 23, 1988, Hebei provience. She has a bachelor degree of civil engineering at 2012, and currently she is a student in College of Resources, Hebei University of Engineering for a master degree in geologic engineering. She has publishs 5 scientific papers about engineering in the related journals. 15