JJMIE Jordan Journal of Mechanical and Industrial Engineering

Transcription

1 JJMIE Jordan Journa of Mechanica and Industria Engineering Voume 11 Number 4, December ISSN Pages Investigate the Natura Convection Heat Transfer in A PCM Therma Storage System Using ANSYS/FLUENT Maher M. A-Maghaseh a, b, * a Coege of Engineering, Paestine Poytechnic University, Hebron, Paestine b Facuty of Engineering and Environment, Northumbria University, Newcaste, UK Received... Accepted... Abstract The present paper presents four dimensiona modes for simuation of a Latent Heat Therma Energy Storage System (LHTESS). The LHTESS is in the form of a rectanguar container with a centra horizonta pipe surrounded by a Phase Change Materia (PCM). Paraffin wax with meting temperature of 60 oc is used as a PCM whist water is used as a Heat Transfer Fuid (HTF). Thermo physica properties of paraffin wax are assumed to be constant in the modeing process, whereas the density variation is handed by using Boussinesq mode. ANSYS/FLUENT software was used in simuating four dimensiona modes of temperature distribution, meting fraction, and fow fieds during the meting process. Simuations performed provide information on the instantaneous temperature distribution, soidification/meting dynamics and the veocities fied in the storage unit during the meting process. The effects of the natura convection on the charging (meting) process were aso investigated Jordan Journa of Mechanica and Industria Engineering. A rights reserved Keywords: PCM; therma storage system; paraffin; soar energy, natura convection, CFD, ANSYS/FLUENT. 1. Introduction Recenty, soid-iquid phase change of PCM, for various geometric arrangements with natura convection in the iquid phase, has been investigated theoreticay and experimentay. The therma energy is stored in the PCM as atent heat and it is reutiized when it is needed. Therefore, understanding the natura convection inside the iquid PCM pays important rues in the design of therma storage system. PCM is stored by different ways in LHTESS, for exampe, using cyinder encosed with or without fins, cans, pates or sphere encosure [1]. In the foowing study, the storage unit consists of a rectanguar container with a horizonta pipe and PCM fied around the pipe. During the period of storage, a working fuid fows inside the horizonta pipe and the heat is transferred through the pipe was to the PCM. Later, PCM reaches its fusion temperature and meting is started and then, natura convection motion appears. Consequenty, the bouncy driven become strong enough to drive the meting process. Many authors have tacked the natura convection during meting of PCM. Tan [2], experimentay investigated the buoyancy and natura convection phenomena during the meting process of PCM inside a spherica capsue. Indeed, vaidate these resuts with a numerica soution obtained with the CFD FLUENT. The most interesting finding was that the conduction heat transfer dominates during the eary stage, whereas the buoyancy driven convection becomes more sufficient as the iquid fraction voume is increased. However, the moten PCM ascends upward to the upper regions of the sphere because of natura convection phenomena. Consequenty, the upper region of the sphere mets first and then the other regions. Another interesting finding from the computationa resuts was that some chaotic fuctuation of the temperature distribution in some points inside the sphere and this is due to the unstabe fuid ayer in these points. However, the findings of the this study support the previous researches by the author [1, 3]. Nsofor [4] aso experimentay investigated the heat transfer and natura convection phenomena in packed bed therma storage system for high temperature. However, positive correations were found in terms of Nusset number, Prandt number as we as Reynods number and comparisons were made with existing correations deveoped with simiar storage media. Wu and Lacroix [5] anaysed numericay the natura convection of meting PCM in a vertica cyindrica capsue heated from beow. The mode was soved using finite-difference method and compared the numerica prediction resuts with the numerica and experimenta resuts of other authors [6-8]. It was reported that the heat transfer rate at the top surface was dominated by conduction, whie it was decreased to zero when meting progressed and natura convection was fuy deveoped. It was aso observed that the highest heat transfer was at the bottom surface of the capsue. Rieger and Beer [9] examined the effect of natura convection fow on heat transfer during meting process of ice inside an isotherma horizonta cyinder. They predicted numericay the overa and oca heat transfer coefficients, temperature fieds, interface positions and fow pattern and * Corresponding author e-mai: maherm@ppu.edu.

2 Jordan Journa of Mechanica and Industria Engineering. A rights reserved - Voume 11 Number 4 (ISSN ) compared their numerica prediction resuts with the experimenta resuts. It was reported that the heat transfer was enhanced at the ower part of the ice body when the wa temperature ess than 8 because of the density effect, whie the ice body move downward when the wa temperature exceed 8. The same procedures have been appied by Rieger et a. [10] to investigate the heat transfer during meting n-octadecane as a PCM inside horizonta tube. Saitoh and Hirose [11] numericay investigated the natura convection inside a horizonta circuator cyinder capsue packed with PCM during meting and soidification process in the case of high Rayeigh numbers. The mode obtained the transient soid-iquid interface, soid-iquid temperature, streamines, isotherms, and heat stored. The numerica resuts was vaidated through comparison with resuts of Pannu [12]. In genera, there was quite different from theirs both quantitativey and quaitativey. It is interesting to note that the natura conviction controed the meting heat transfer inside the capsue as we as the therma instabiity happened at the bottom portion of the capsue. Trp [13, 14] studied the transient heat transfer in the she-and-tube therma storage system in an experimenta and numerica study. He deveoped a mathematica mode based on non-isotherma phase transition and it has been impemented with a FORTRAN computer code. The numerica resuts were vaidated through a comparison with experiment data. He concuded that heat transferred from HTF to the PCM was sow; this is because of the arge Prandt numbers of the HTF. Therefore, a arge amount of heat was carried out downstream with the HTF, whies a sma amount of heat was transferred to the PCM upstream. The same author [15] numericay investigated the effect of severa geometric parameters and different HTF operation conditions on the heat transfer during both meting and soidification processes. They measured the transient temperature distribution of the HTF, PCM and tube wa. The objective of this study is to investigate numericay the natura convection dominated meting of PCM fied around horizonta pipe within a rectanguar container. Four different modes are investigated (Fig. 1). First, the centre of the horizonta pipe is ocated at the centre of the rectanguar container (Fig. 1.A). Second, the centre of the horizonta pipe is ocated by 5mm down from the centre of the rectanguar container (Fig.1.B). Third, the centre of the horizonta pipe is ocated by 10mm down from the centre of the rectanguar container (Fig.1.C). Last, the centre of the horizonta pipe is ocated by 15mm down from the centre of the rectanguar container (Fig.1.D). The CFD findings are presented and discussed in the present study. 2. Simuation Mode Fig. 1 shows the cross section of the four different rectanguar therma storage configurations with horizonta pipe. The test unit has dimensions 75mm (width) 50mm (height) 500mm (depth). The rectanguar container is fied with Paraffin wax which has a meting point of 60, atent heat storage capacity of 200 kj/kg, density of 800 kg/m3, specific heat of 2050J/kg K, dynamic viscosity of kg/m-s, and therma conductivity of 0.25 W/m K. The horizonta pipe was 10mm of inner diameter with a thickness of 1mm. The HTF is circuated inside the pipe with inet temperature of 353 K in order to charge the PCM storage unit. The mass fow rate of water is Kg/s. The computationa grid of the system was buit using ICEM CFD 13.0 software by ANSYS. Meshing of the mode was generated by using hexahedra eements and boundary ayers were created around the pipe. Upon severa trias it was found that the hexahedra computationa grid with eements woud be sufficient for accurate 3D modeing of the heat transfer in the PCM system. Figure 1: Schematic view of the four different modes, (A) case 1, (B) case2, (C) case3, and (D) case4

3 2017 Jordan Journa of Mechanica and Industria Engineering. A rights reserved - Voume 11 Number 4 (ISSN ) 219 The fow inside the pipe was described using transient simuations with k-epsion turbuence mode depoyed. The soidification/meting mode was used in order to examine the phase change phenomena in paraffin wax. The time step used in cacuations was set to 0.1 second. To obtain numerica resuts the first-order upwind spatia discretization and the pressure sover with PRESTO agorithm for pressure-veocity couping were used. Convergence criteria were defined by setting the absoute residua vaue at 10-6 for energy and at 10-3 for a other variabes. The mathematica formuations for soving PCM reated probems have been categorized [16] as fixed grid, variabe grid, front-fixing, adaptive grid generation, and enthapy methods. Two methods are used to anayse the heat transfer in soid-iquid PCMs. These are the temperature-based and enthapy-based methods. In the former, temperature is considered to be a singe dependent variabe. The energy equations for both soid and iquid are formuated separatey; and thus the soid-iquid interface positions can be tracked easiy to achieve an accurate soution for the probem [17]: Ts T k k Lv n n s n where Ts denotes the temperature in the soid phase; T is the temperature in the iquid phase. k s, and k are the therma conductivity of the soid phase and iquid phase, respectivey; n is the unit norma vector to the interface; L is the atent heat of the freezing; and v n is the norma component of the veocity of the interface. However, an enthapy-porosity method is used for modeing the soidification/meting process [18]. This technique is described in detai by Voer and Prakash [19]. The energy conservation equation for this case is written as: t r H vh kt S.. where H, is the enthapy, ρ is the density, v r is fuid veocity and S is the source term. The enthapy of the materia is cacuated as the sum of the sensibe heat, h, and atent heat, ΔH: H h H (3) The sensibe heat is cacuated as: T ref p (4) Tref h h c dt where, h ref is the reference enthapy, T ref is the reference temperature and c p is the specific heat at constant temperature. The atent heat is aso cacuated as: H L (5) The iquid fraction,, can be cacuated as: (1) (2) 0, whent T 1, whent T soid soid T T if T T T soid soid iquid Tiquid Tsoid The soid and iquid temperatures are aso cacuated as: T T K my (7) soid met i i i soutes Tiquid Tmet my i i (8) soutes where, K i is the partition coefficient of soute i, which is the ratio of the concentration soid to that in the iquid at the interface; Y i is the mass fraction of soute i, and m i is the sope of the iquid surface with respect to Y i [18]. The source term in the momentum equation can be written as [18]: 3 1 r r S Amush v v 3. Resut and Discussion Numerica simuation is carried out for cycic meting process of Paraffin wax fied in rectanguar therma storage unit invoving horizonta pipe and the resuts are presented and evauated in this section. Numerica simuations investigate the natura convection dominated meting of PCM for different mode geometries (Fig.1). At the start of the meting cyce, the inet temperature of HTF inside the pipe is maintained at fixed temperature of 353 K and the initia temperate of PCM is 300 K. Further, mass fow rate of HTF is maintained at Kg/s. Figs. 2-3 distincty show temperature distribution and meting/soidification fieds aong the four different geometries after 10000sec. The resuts show that the temperature starts to rise graduay in the region of the storage container cose to the pipe wa and then ascends upward to the upper region at the centre of the container. In this stage, sensibe heat was transferred from the pipe wa to the PCM soid by pure conduction, and then a thin iquid ayer was created between the pipe and the soid PCM. The soid-iquid interface expanded graduay over the axia and radia directions with respect to time. Thereafter, the meting fronts were dominated by natura convection heat transfer in the meted regions of PCM. Consequenty, the convection heat transfer drives circuation in the meted PCM due to the buoyancy force. The moten PCM ascends upward from the bottom to the upper regions at the center of the container and returns downward to compete the natura convection circe, since the moten PCM has the ower density and viscosity. The convection circe became more sufficient as the iquid fraction voume is increased. p (6) (9)

4 Jordan Journa of Mechanica and Industria Engineering. A rights reserved - Voume 11 Number 4 (ISSN ) Figure 2: Temperature distribution process in the domain, (A) case 1, (B) case2, (C) case3, and (D) case4. Eapsed time is s. Two factors affected the natura convection in the meted PCM; these are the temperature difference as we as the distance between the pipe wa and soid-iquid interface. This is ceary expained in Figs It can be aso deduced that the moten regions of PCM is remarkaby bigger in case 4, and then in cases 3 &2 with respect to that of case 1. Figure 3: Soidification/Meting process in the domain, (A) case 1, (B) case2, (C) case3, and (D) case4. Eapsed time is s

5 2017 Jordan Journa of Mechanica and Industria Engineering. A rights reserved - Voume 11 Number 4 (ISSN ) 221 This is because the natura convection in the PCM is affected by the temperature difference as we as the distance between the pipe wa and the soid PCM. This therefore increases the amount of PCM, resuting of increase the distance between the pipe wa and soid PCM, and so increases the natura convection mechanism. In order to evauate the performance of the therma storage unit 132 monitoring points were set inside the domain with PCM to record the variation of the temperature as a function of time. These points are ocated in 11 measurement panes perpendicuar to the axis of the domain (see Fig. 6). Each pane contains 12 monitoring points, as shown in Fig. 7. A monitoring points are divided into three groups. The first group of 44 points was set in the upper part of the computationa domain (u1-u44). Figure 4: Meting process in the front section of the domain, (A) case 1, (B) case2, (C) case3, and (D) case4. Eapsed time is s. Figure 5: Veocity profie in the front section of the domain, (A) case 1, (B) case2, (C) case3, and (D) case4. Eapsed time is s.

6 Jordan Journa of Mechanica and Industria Engineering. A rights reserved - Voume 11 Number 4 (ISSN ) The second groups of 44 points are ocated in the bottom part of the domain (b1-b44), and the ast groups of 44 points are ocated at the side of the domain (e1- e44). Figs. 4-6 show the temperature variations monitoring points in the first section of the computationa domain (the measurement pane 1). Fig. 3 shows the temperature variation at u4, which ocated at the upper part of the domain. It can be seen that the temperature in case1 is the highest compared to that in cases 2, 3 and 4. The reason behind this is that the distance between the point u4 at the upper part (Fig. 8) and the outer surface of the pipe is ess than those in cases 2, 3 and 4. Figs. 11 and 12 show the iquid fraction on PCM against time variation and tota meting time. It can be deduced from this figures that the tota meting time was reduced by approximatey by 16.57% with case2, 31.83% with case3 and 41.3% with case 4, compared to that of case1. As mentioned above, the average heat is transferred to the PCM by conduction heat transfer from the pipe wa to soid PCM and convection heat transfer in the meted PCM. Indeed, conduction heat transfer was maintained by using the same operation and boundary conditions on the pipe and HTF in a cases. Thus, tota meting time was significanty reduced in cases 3&4 due to the natura convection in meted PCM. The anaysis further indicated that the appropriate configuration and height for the tube as in case 4 which provides the shorter meting time of the PCM. Figure 6: Measurement panes aong the tube. Figure 9: The temperature variation at point b4 (the bottom of the domain, first measurement pane). Figure 7: The ocation of 12 monitoring points in the measurement pane 1. It can be contributed to the conduction heat transfer from the pipe wa to the PCM is maintained constant due to the same inet parameters and boundary conditions in a cases. However, the temperature distribution in case1 is more significant on the points at the upper part of the storage unit due to the effect of natura convection. On the contrary, the temperature at the points at the bottom parts in is significanty increased as the distance between the PCM and the HTF decreased. It can be seen in Fig. 9 that the temperature variation at the bottom parts in case 4 is higher with respect to that of other cases. Figure 10: The temperature variation at point e4 (the side of the domain, first measurement pane). Figure 11: Meting fraction on PCM. Figure 8: The temperature variation at point u4 (the top of the domain, first measurement pane).

7 2017 Jordan Journa of Mechanica and Industria Engineering. A rights reserved - Voume 11 Number 4 (ISSN ) 223 Figure 12: Tota meting time of PCM. 4. Concusion In the present study, a numerica simuation has been carried out to determine the effect of natura convection dominated meting of PCM fied around horizonta pipe within a rectanguar therma storage unit. The most obvious finding to emerge from this study is that the natura convection heat transfer has a considerabe effect on both axia and radia temperature distribution aong the storage unit, and so reducing the temperature difference between the pipe was and soid-iquid interface, therefore, reducing the tota meting time of PCM. The tests for investigating four different therma storage configurations were carried out. The resuts indicate that the tota meting time was approximatey reduced 16.57% with case2, 31.83% with case3 and 41.3% with case 4, compared to that of case1. Indeed, heat transfer rate enhanced is significanty more pronounced in natura convection in meted PCM, whie the conduction heat transfer was maintained constant in a cases by using the same operation and boundary conditions on the pipe and HTF. References [1] F. L. Tan and C. W. Chan, "Soidification inside a sphere an experimenta study," Internationa Communications in Heat and Mass Transfer, vo. 33, pp , [2] F. L. Tan, S. F. Hosseinizadeh, J. M. Khodadadi, and L. Fan, "Experimenta and computationa study of constrained meting of phase change materias (PCM) inside a spherica capsue," Internationa Journa of Heat and Mass Transfer, vo. 52, pp , [3] F. L. Tan., "Constrained and unconstrained meting inside a sphere," Internationa Communications in Heat and Mass Transfer, vo. 35, pp , [4] E. C. Nsofor and G. A. Adebiyi, "Measurements of the gaspartice convective heat transfer coefficient in a packed bed for high-temperature energy storage," Experimenta Therma and Fuid Science, vo. 24, pp. 1-9, [5] Y. K. Wu and M. Lacroix, "Meting of a PCM inside a vertica cyindrica capsue," Internationa Journa for Numerica Methods in Fuids, vo. 20, pp , [6] Y. K. Wu, "Numerica Studies of Meting Process in a Cyindrica Encosure," Ph.D., Écoe Poytechnique de Montréa, [7] M. Prud'homme, T. H. Nguyen, and Y. K. Wu, "Simuation numérique de a fusion à 'intérieur d'un cyindre adiabatique chauffé par e bas," Internationa Journa of Heat and Mass Transfer, vo. 34, pp , [8] Y. K. Wu, M. Prud'homme, and T. H. Nguyen, "Étude numérique de a fusion autour d'un cyindre vertica soumis à deux types de conditions imites," Internationa Journa of Heat and Mass Transfer, vo. 32, pp , [9] H. Rieger and H. Beer, "The Meting Process of Ice Inside a Horizonta Cyinder: Effects of Density Anomay," Journa of Heat Transfer, vo. 108, pp , [10] H. Rieger, U. Projahn, M. Bareiss, and H. Beer, "Heat Transfer During Meting Inside a Horizonta Tube," Journa of Heat Transfer, vo. 105, pp , [11] T. Saitoh and K. Hirose, "High Rayeigh Number Soutions to Probems of Latent Heat Therma Energy Storage in a Horizonta Cyinder Capsue," Journa of Heat Transfer, vo. 104, pp , [12] J. Pannu, G. Jogekar, and P. A. Rice, "Natura convection heat transfer to cyinders of phase change materia used for therma storage," American Institute of Chemica Engineers, vo. 76, pp , [13] A. Trp, "A numerica and experimenta study of transient heat transfer in a she-and-tube atent heat storage unit with paraffin as a phase change materia," presented at the Energy and the Environment, [14] A. Trp, "An experimenta and numerica investigation of heat transfer during technica grade paraffin meting and soidification in a she-and-tube atent therma energy storage unit," Soar Energy, vo. 79, pp , [15] A. Trp, K. Lenic, and B. Frankovic, "Anaysis of the infuence of operating conditions and geometric parameters on heat transfer in water-paraffin she-and-tube atent therma energy storage unit," Appied Therma Engineering, vo. 26, pp , [16] N. Özişik, Finite Difference Methods in Heat Transfer,: Tayor & Francis, [17] A. A. A-abidi, S. Bin Mat, K. Sopian, M. Y. Suaiman, and A. T. Mohammed, "CFD appications for atent heat therma energy storage: a review," Renewabe and Sustainabe Energy Reviews, vo. 20, pp , [18] ANSYS FLUENT 14.0 Theory Guide, : ANSYS, Inc, [19] V. R. Voer and C. Prakash, "A fixed grid numerica modeing methodoogy for convection-diffusion mushy region phase-change probems," Internationa Journa of Heat and Mass Transfer, vo. 30, pp , 1987.