MODELLING OF VENTILATION AIRFLOW RATES OF SOLAR CHIMNEYS FOR BUILDING INTERGRATION OF RENEWABLE ENERGY DEVICES

Transcription

1 MODELLING OF VENTILATION AIRFLOW RATES OF SOLAR CHIMNEYS FOR BUILDING INTERGRATION OF RENEWABLE ENERGY DEVICES A.G. L, Ph.D., P. Jones, Ph.D., P.G. Zhao, L.P. Wang ABSTRACT Heat transfer process and natural ventlaton effectveness drven by a solar chmney attached to a sdewall of buldng are nvestgated wth CFD technque n detal. In ths paper, condtons and parameters studed n the modellng study are the cavty wdth of the solar chmney, the wall temperature, the heght and breadth of the solar chmney, the rato of outlet area to nlet area as well as the outlet locaton of the solar chmney. The ranges of calculaton parameters focused on a solar chmney wth sngle-sded solar collector (sngle-sded heated wall) cover followng condtons: solar chmney length L =.5m~5.m,breadth B=.1m~.5m,heght H=2.m~ 5.m,and B/H=.5~.25. Heated wall surface temperature Tw s changed n the range of 3 ~7 ; the rato of the outlet area to nlet area A r s changed n the range of.6~.8. It s found that for gven buldng geometry and nlet areas, there s an optmum cavty wdth at whch a maxmum arflow rate can be acheved. Based on the predcton, the arflow rate reaches maxmum when B/H s approxmately 1/1. It s also found that for gven chmney geometry, solar chmney ventlaton flow rate can be ncreased wth the enhancement of chmney heght only the cross sectonal area no more than the crtcal area, because cross secton area has a strong effect on the transtonal and/or turbulent convectve heat transfer n an enclosure. From the vew of economy technology, the optmzed heght of a solar chmney can be determned accordng to the optmzed secton rato of breadth to heght and avalable breadth n practce. It s also noted that optmzed ventlaton flow rate can be obtaned when the outlet area takes the same area as the nlet area. Generally, there s good agreement between the numercal results and avalable expermental data from lterature for a solar chmney. INTRODUCTION In the desgn of new buldngs and retroft of old buldngs, the attenton s turnng towards a more ntegral energy desgn on optmal use of sustanable technologes such as natural ventlaton (daytme comfort ventlaton and nght coolng)(bre/cibse 1997;Brarozz et al. 1999; L et al 2). The current trend s to desgn buldngs responsve to the clmate wth an acceptable ndoor envronment and hgh energy effcency. For example, a recent survey n UK ndcated that 9% of drectors and senor managers preferred natural ventlated buldngs wthout ar-condtonng (Lddament 1997). The lmtatons of conventonal energy sources, n terms of cost and avalablty, and an ncreased awareness of envronment ssues, have led to renewed nterest n passve buldng desgn. One mportant applcaton of ths technology s to ntensfy nteror ventlaton by usng a solar chmney to nduce buoyancy-drven flow. Solar chmney natural ventlaton has hgh ground n both developng and developed countres. For nstance, nnovatve solar chmney natural ventlaton desgn technques have an hstorcal bass and have been advanced n recent years (L 2;Lddament 1997;Gvon 1994; Khedar 2). There are many examples usng chmney effect to ventlaton, coolng and smoke exhaust n Chna: courtyard, Xn Jang arch buldng, and cave dwellngs (L 21). In UK, there has also been a development of solar chmney stack-nduced cross ventlaton desgn for non-domestc buldngs, such as Audtora and Offces, Inland Revenue Headquarters Buldngs located n Nottngham. However, dfferent heated sdes and ar outlet locatons can result n dfferent temperature felds and velocty felds. As a result, the arflow rate and exhaust temperature are also changed. Maor uncertanty n the desgn of solar chmney s the effect of the geometrcal structures, solar radaton and outdoor temperature. Unfortunately, we have lttle knowledge n determnng those parameters for an effcent and effectve desgn of the solar chmney natural Angu L s Professor, School of Envronmental & Muncpal Engneerng, X'an Unversty of Archtecture & Technology, X'an, PR. Chna, Angu L was a Royal Socety KC Wong Fellow from 22 to 23 at Cardff Unversty, U. K.; Phllp Jones s Professor, Head of Welsh School of Archtecture, Cardff Unversty, Cardff, Wales, CF1 3 NB, U. K.; Pngge Zhao and Lpng Wang 1

2 are Graduate students, X'an Unversty of Archtecture & Technology, X'an, PR. Chna. Correspondng author. Tel/Fax.: ; E-mal: ventlaton systems (Pashnamurth 1998;Bansal et al.1993). Ths paper s focused on arflow and convectve heat transfer nsde the solar chmney. The emphass s lad on the comparson between dfferent heated sdes, outlet locatons and the relatonshp between the outlet and nlet area, the temperature dfference, and the optmum structural sze. TURBULENCE MATHEMATIC MODEL AND MITFLOW PROGRAM Ths research s carred out through CFD calculaton wth theoretcal analyses and MITFLOW program, whch was developed by Chen, etc n MIT [1]. The temperature feld and velocty feld are showed by FORTRAN coded program and EXCEL2. About 6 cases have been calculated to fnd out the optmum condton for solar chmney ventlaton desgn. The rate of arflow can be calculated from these equatons n varous szes of solar collector, seeng whether ventlaton requrements can be satsfed. Arflow nsde solar chmney belongs to turbulent flow. In ths research, MITFLOW equaton model based on tme average method s solved. Every grd s n conformty wth the conservaton of mass flow,.e. the nlet mass flow s equal to the outlet flow. For one-sded heated solar chmney turbulent convectve dffuson, ts governng equatons are concluded from the basc equatons n flud dynamcs (contnuty, momentum, energy) and tme average method. When zero governng equaton s concluded, the prmary assumptons are as the followng (1) Arflow nsde solar chmney s low speed, ncompressble and conforms to deal gas law. (2) Indoor ar belongs to Newton fluds whose surface stress s conform to broadly Newton vscosty stress equaton. (3) Boussngq assumpton s appled. Turbulence tme average equatons are obtaned based on the assumptons above. Contnuty Where V average velocty n x coordnate x drecton V X = (1) Momentum Where ρ ar densty ρv ρ VV + t V average velocty n x drecton P pressure µ effectve vscosty coeffcent eff P = + µ β coeffcent of ar thermal expanson T O temperature at the referent pont T ar temperature g acceleraton of gravty n drecton The last rght tem n equaton s buoyancy. V V eff( + ) + ρβ( T T) g (2) Energy In order to determne the temperature dstrbuton and buoyancy tem n (2),energy conservaton equaton should be solved. 2

3 ρt ρvt + t X Where Γ temperature effectve dffusvty T, eff = X q heat source C p specfc heat at constant pressure In ths research, the equaton below s used to estmate the temperature effectve dffusvty. µ eff Γ T, = (4) eff Preff Pr eff general Prandtl number Turbulent effects are unted to turbulent effectve dffusvty, whch s the sum of turbulent dffusvty µ and lamnar vscosty coeffcent. t Γ T, eff µ µ + µ T X + q c p (3) eff = t (5) In the assumpton of Prandtl-Kolmogorov,turbulent dffusvty µ t s the result of turbulent fluctuaton momentum energy and turbulent fluctuaton dmenson.l s used to denote turbulent length proportonal scale. 1/ 2 µ t = C v ρk l (6) Where C v = s emprcal constant (Chen; Glcksman 21). Accordng to the dfferent methods to solve the unknown parameters K and L,the turbulent vscosty models are dvded nto many forms. The smplest model s Prandtl mxng length model. In ths research, turbulent vscosty s denoted wth local average velocty and length scale by the smple algebrac equaton (Chen 21;Afonso et al 2) µ t =. 3874ρVl (7) l length scale, the dstance between envelope and nearest surface. Those equatons can be dscretzed nto algebrac equatons, and a seres of algebrac equatons can be solved by MITFLOW program. Algebrac equatons n up-wnd dfference form are used durng modellng process as used by Chen and Glcksman (21). NUMERICAL SIMULATION RESULTS AND ANALYSES The theoretcal model to descrbe sngle-sded solar chmney s dealzed. Amongst other smplfcatons, t assumes that there s no external wnd; that the densty of the ar outsde the chmney s unform; that the solar collector s derved from a plate of unform temperature; and that the adacent walls of the solar chmney are nsulated. These studes are focused on 4 parameters, that s, H, B/H, A and L. The case geometrcal structures studed for sngle-sded r heated solar chmney are presented n Fg.1. The calculatons on arflow rate m and outlet temperature t out are performed respectvely. The ranges of calculaton parameters are reported below. Solar chmney model sze:l(length) B(breadth) H(heght), L=.5~5.m,B=.1m~.5m,H=2.m~5.m,B/H=.5~.25 Heated wall surface temperature T w = t w +273, t w =3 ~7, t n =2 The rato between outlet area and nlet area A r =.6~.8 In consderng practcal archtecture condtons,the dstance between the upper edge of outlet and the top of solar chmney s.1h. The outlet s located at the heated wall. In ths paper, about 6 cases coverng the nfluence of the solar collector wall temperature t w, nflow temperature t n, solar chmney breadth B, length L, heght H, as well as the rato between 3

4 outlet area and nlet area A r on arflow and outlet (exhaust) temperature t out are smulated n order to understand better the mechansm of solar chmney natural ventlaton. The relatonshp between the arflow rates and solar chmney heght parameter are demonstrated n Fgure Inflow Fgure 1 Vertcal solar chmneys for buldngs. 2 tw=3 Arflow rate (KG/S) Exhaust temperature ( ) H( m) F gur e 2 The r el at on bet ween a r f l ow r at e and ch mney he ght H( m) tw=3 tw=4 tw=5 tw=6 tw=7 tw=4 tw=5 tw=7 F gur e 3 The r el at onsh p bet ween exhaust temperature and chmney he ght 4

5 The arflow rate of solar chmney s notably affected by chmney heght. When the solar collector (wall) temperature s n the range of 3 ~5, arflow rates can reach the maxmum at the heght of H=4m. Meanwhle, from the Fgure 2, when the heght s ncreased to 4 m or more, the ncrease of arflow rate s gradually lower down. Ths phenomena can be partly explaned by the natural convecton n a channel formed by a sngle sothermal plate and an nsulated plate. Only the solar collector surface, whch s an sothermal surface, s nvolved n heat transfer, whle the assocated adacent surfaces beng adabatc. The Nusselt number based on the temperature dfference between the wall and the ambent ar was once proposed by Bar-Cohen (1984). 2 4 c p ρ gβb ( t Nu = 24 µ kh 1 w t ) 1 e (1 ΓH ΓH )(1 e ΓH ) (8) The hghest convectve heat transfer coeffcents can be expected n vertcal channel formed by an sothermal plate and an nsulated plate at a certan heght and wdth (Bar-Cohen, Rohsenow 1984). The arflow rate nsde the solar chmney s exactly produced by the densty dfference or temperature dfference caused by heat transfer. As a consequence, the optmum heght and wdth could be antcpated (see Fgure 6). It s thus possble to select the chmney heght or/and breadth whch wll maxmum the arflows or, alternatvely, choose the heght or breadth whch wll yelds the maxmum heat transfer from the entre solar collector. On the other hand, n fact, the solar chmney to be nvestgated here s three-dmensonal rather than two-dmensonal as a channel mentoned above. The nfluence of three-dmensonal flow,.e., sde nflow, or lateral edge effects, on the arflow should not be gnored. Clearly such effects can be antcpated to become progressvely greater as the surface temperature s ncreased, seeng the curve at the wall temperature t w =7 shown n Fgure 2. The exhaust temperature or outlet temperature s enhanced as the heght or wall temperature s ncreased, as presented n Fgure 3. Arflow rate(kg/s) F gur e 4 The r el at onsh p bet ween a r f l ow r at e and r at o of out l et ar ea t o nl et ar ea From the vew of energy conservatve, the curve tendences are easly to be understood. The exhaust temperature/outlet temperature wll gradually ncrease as the heght or solar collector temperature enhances The relatonshp among arflow rate m, the rato of outlet area to nlet area Ar s shown n Fgure 4. In Fgure 5,the exhaust temperature changes brought about by the rato of outlet exhaust temperature ( C) Ar Ar tw=3 tw=4 tw=5 tw=6 tw=7 F gur e 5 The r el at onsh p bet ween exhaust temper at ur e and r at o of out l et ar ea t o nlet area tw=3 tw=4 tw=5 tw=6 tw=7 5

6 area to nlet area s presented. The exhaust temperature decreases wth the ncrement of outlet areas or decrement of nlet areas. When B/H s vared but parameter condtons L,H,T n,v n and T w reman constant,arflow rate m and outlet temperature t out s predcted and shown n Fgure 6 and Fgure 7. Moreover, arflow rate ncreases wth the heated wall temperature ncreasng. the bgger the rato of outlet area and nlet area, the more the ncreasng value.. 25 In Fgure 6,arflow rate vares wth the varaton of rato of breadth to heght. Arflow rate s up to the maxmum when the rato of breadth to heght s about 1/1.But when the rato s ncreasng further, exceedng 1/1, and arflow rate wll fall down a lot. Ths trend shown n the solar chmney could be explaned by the buoyant plumes characterstcs. Ths mean the reasonable breadth (space) could nduce maxmum ar from the nlet at the bottom. In other words, too much breadth does not create the effectve arflow rate enhancement. Fgure 7 shows that temperature trends at the outlet whle the ratos, breadth to heght, as calculated from MITflow program. When B,L,H,T n,v n and T w reman constant, however, the A r s vared,arflow rate m outlet average temperature t out s nvestgated. The studed cases parameter condtons and outcome are reported n Fgure 8. In fgure 8,arflow rate vares wth the varaton of rato of breadth to heght. Arflow rate s up to the maxmum when the rato of breadth to heght s about.1 ( or 1/1). However, when the rato s ncreasng further to the.15, arflow rate notably fall down a lot. These phenomena also totally concde wth the fact presented n Fgure 6. Agan there s evdence to show that the szes of solar chmney are essental for the heat transfer and the arflow rates. arflow rate(kg/s) exhaust temperature(oc) arflow rate (KG/S) B/ H F gur e 6 The r el at onsh p bet ween a r f l ow r at e and r at o of br eadt h t o he ght tw=3 tw=4 tw=5 tw=6 tw=7 tw=3 tw=4 tw=5 tw=6 tw= B/ H F gur e 7 The r al at onsh p bet ween exhaust temperature and rato of breadth to he ght B/ H=. 5 B/ H=. 1 B/ H=. 15 B/ H=. 2 B/ H=. 25 L(m) F gur e 8 The r el at onsh p bet ween a r f l ow r at e and ch mney he ght 6

7 In Fgure 9, exhaust temperature decreases wth the ncrement of rato of breadth to heght. It mght be the result of ar mxng effect, as the rato of breadth to heght ncreases, there wll be more space (cross sectonal area) for arflow passage gradually. It s also noted that effects on arflows from the breadth change outweghs the effects from heght change. In the range of heght.5 to 5m, the exhaust temperature does not vary much when the heghts enhance but the breadth remans constant. So the sutable solar chmney heght should be selected n desgnng the solar chmney, based on the vew of ventlaton enhancement and economcs. exhaust temperature ( ) L( m) F gur e 9 The r el at onsh p bet ween exhaust temper at ur e and ch mney l engt h B/ H=. 5 B/ H=. 1 B/ H=. 15 B/ H=. 2 B/ H=. 25 The effects of temperature dfference between nlet and outlet on arflow rates are shown n Fgure 1. The bottle neck effect caused by the cross secton appears agan. The arflow rate does not always ncrease wth the enhancement of the temperature dfference between nlet and outlet. The arflow rate wll reach the maxmum at a certan temperature dfference between nlet and outlet for a gven solar chmney szes.. 16 arflow rate(kg/s) T/T n F gur e 1 The r el at onsh p bet ween a r f l ow rate and T/Tn 7

8 CONCLUSIONS By the CFD predcton on the mechansm of natural ventlaton, the relatonshp between the arflow rate and parameters for sngle-sded heated solar chmney s concluded. In the ranges of, solar chmney length L =.5m~5.m,breadth B=.1m~.5m,heght H=2.m~5.m,and B/H =.5~.25, t s found that for a gven buldng geometry and nlet area, there s an optmum cavty wdth at whch a maxmum arflow rate can be acheved. There s a bottle neck effect zone nsde the chmney for the natural ventlaton rates, and natural ventlaton rates are restrcted by ths bottle neck phenomena. Based on the present studed ranges, the arflow rate reaches maxmum when B/H s approxmately 1/1. It s also found that ventlaton mass rate may be ncreased wth the enhancement of chmney heght only the cross sectonal area no more than the crtcal area. From the vew of technology and economy, the optmzed heght of a solar chmney can be determned accordng to the optmzed secton rato of breadth to heght and avalable practcal feld condtons. ACKNOWLEDGEMENTS Ths research proect s sponsored by Royal Socety K C Wong Fellowshp foundaton (U.K.), Chna Natural Scence Foundaton (No ), and SRF for ROCs, SEM. REFERNCES 1 BRE/CIBSE Ventlaton and ar polluton: Buldngs located n urban and cty centres. Buldngs Servces Engneerng Research and Technology. 18 (4), pp Brarozz, G.S., M.S.E.Imbab, E.Noble, and A.C.M.Sousa Physcal and numercal modellng of a solar chmney-based ventlaton system for buldngs. Buldng and Envronment, 27(4),pp L, A.G., and P.J.Jones. 2. Developments n strateges used for natural and mechancal ventlaton n Chna. Indoor + Bult Envronment. 9, pp Lddament, M BRE/CIBSE dscusson and round-up. Buldngs Servces Engneerng Research and Technology. 18 (4), B16-B18 5 L, A.G. and P.J.Jones. 21. Improve passve ventlaton effectveness of buldngs through open-sded ar cowls. IAQVEC. Changsha, Chna 6 Gvon, B Passve and Low Energy Coolng of Buldngs, Van Nostrand Renhold, New York 7 Khedar, J., B.Boonsr, and J.Hrunlabh. 2. Ventlaton mpact of a solar chmney on ndoor temperature fluctuaton and ar change n a school buldng, Energy and Buldngs, 32, pp Pashnamurth N Expermental and theoretcal performance of a demonstraton solar chmney.j. Energy Res., 22, pp Bansal, N.K., R. Mathur, and M.S. Bhandar Solar chmney for enhanced stack ventlaton, Buldng and Envronment, 28 (3), pp Chen Q., and L.R. Glcksman.21. Smplfed methodology to factor room ar movement and the mpact on thermal comfort nto desgn of radatve, convectve and hybrd heatng and coolng systems, ASHRAE RP-927,ASHRAE. 11 Afonso, C. and A. Olvera.2. Solar chmneys: smulaton and experment, Energy and Buldngs, 32 (1), pp Chen, Z.D. and Y. L. 22. Dsplacement natural ventlaton n a sngle-zone buldng wth three level openngs, Buldng and Envronment, 37, pp Bar-Cohen, A. and W.M. Rohsenow Thermally optmum spacng of vertcal, natural convecton cooled, parallel plates, J. Heat Transfer, 16,pp