[000007]

7/27/2019 [000007]

http://slidepdf.com/reader/full/000007 1/24

Simulation and Performance Analysis of a 4-Effect

Lithium Bromide-Water Absorption Chiller

*Technion-Israel Institute of Technology

submitted toASHRAE Symposium

Chicago, IllinoisJanuary 28 - February 1, 1995

Prepared by theOAK RIDGE NATIONAL LABORATORY

managed byMARTIN MARIETTA ENERGY SYSTEMS, INC.

Oak Ridge, Tennessee 3783 1-2008for the

U.S. DEPARTMENT OF ENERGYunder Contract No. DE-AC05-840R21400

T h e subnrittcd m u s c r i p t has been authored by a contractorofthe U.S. govcmmcnt under Con trac t No. DE-ACOS-

840R21400. Accordingty, the U.S. Govcmmcnt retains anonexclusive, royally-fizc license to p u b M or reproduce the

p u b W orm of t iscontribution, or allow othm to do so, for

U.S. Govemmcntpurposes.

7/27/2019 [000007]


Dl SCLAlMER

Portions of this do cum ent m ay b e illegible

in electronic imag e products. Im ages areproduced from t he best available original

document.

7/27/2019 [000007]


PAPER COVER SHEET

Simulation and Perform ance Analysis of a 4-Effect Lithium Brom ide-Water A bsorptionChiller

G. Grossman, Sc.D., ASHRAE Member; A. Zaltash, Ph.D.; and R.C. DeVault

G. Grossma n is Professor, Faculty of Mechanical Engineering, Technion-Israel Instituteof Technology, Haifa 32000, Israel;A.Zalta sh is Dev elopm ent Staff Mem ber, Energy Division, Oak Ridge NationalLaboratory, Oak Ridge, Tennessee 37831, USA;R.C. DeVault is Rese arch Staff Mem ber, Energy Division, O ak Ridge N ationa lLaboratory, Oak Ridge, Tennessee 37831, USA.

Perfo rma nce simulation has been conducted for a 4-effect lithium brom ide-water chiller,cap able of substan tial performance improvement over state-of-the-art double-effectcycles. Th e system investigated includes four condensers and four deso rbers coupledtogeth er, forming an extension of t he conventional double-effect cycle; based on prioranalytical studies, a paralle l flow system was pre ferred over series flow, an d double-condenser coupling was employed, to further improve performance. A modular computercode fo r simulation of ab sorption systems (ABSIM) was used to investigate theperformanc es of th e cycle. Th e simulation was carried out to investigate th e influence ofsom e major design para me ters. A coefficient of perform ance a roun d 2.0 (cooling) wascalculated at the design point, with a heat supply temperature of 600°F (315°C) at th esolution outlet from he high temperature desorber. With some optimization of the weak(pum ped ) solution flowrate and of the solution split among the fou r desorbers, this COPmay be raised above 2.2.

Key W ords: Absorption, H ea t Pum p, Chiller, Gas-Fired, Simulation

7/27/2019 [000007]


Sim ulation and P erform ance Analysis of a 4-Effect Lithium Bromide-W ater AbsorptionChiller

ABSTRACT

Perform anc e simulation has bee n conducted for a 4-effect lithium brom ide-water chiller,capab le of substantial performance improvement over state-of-the-art double-effectcycles. The system investigated includes four condensers and four desorbers coupledtoge ther , forming an extension of the conventional double-effect cycle; bas ed on priorexperience, a parallel flow system was preferred over series flow, and Double-CondenserCoupling (DCC) was employed, extending from triple-effect cycles, to further improveperformance. A modular computer code for simulation of absorption systems ABSIM)was used to investigate the perform ances of the cycle. Th e simu lation was carr ied out toinvestigate the influence of some major design parameters. A coefficient of performance

around 2.0 (cooling) was ca lculated a t th e design point, with a heat supply temp eratureof 600°F (315°C) at the solution outlet from the high temperature desorber. With someoptim ization of the weak (p um ped ) solution flowrate and of the solution split among thefour d esorbe rs, this COP may be raised above 2.2.

INTRODUCTION

All current gas-fired residential absorption cooling systems are based on the well-knownsingle-effec t or doub le-effect cycles. Single effect systems (COP =0.7) ar e severely limitedin their ability to utilize high t em pera ture heat sources, and a re particularly suitable for

was te h eat or solar applications. Th e double-effect cycle (C OP =1.2) represents asignificant ste p in perform ance im provement over the basic single-effect cycle.

In ord er to further im prove utilization of the high temperature hea t available fromnatu ral gas, a variety of triple-effect cycles have been proposed, capable of substantialperformance improvement over equivalent double-effect cycles. In a recent study(Gros sm an e t al., 1994), severa l of these cycles were simulated and analyzed in detail.Am ong th e cycles considered were 1) the three-condenser-threedesorber(3C3D) triple-effec t cycle (Oo uch i et al., 1985), forming an extension of the conventiona l double-effectcycle, comprising one evaporator, on e absorber, three condensers and th ree desorbers,recovering heat from each high te mp erature condenser to the next low er temperatu redesorber; (2) a variation .of the 3C3D cycle with Double- Condenser Coupling (DCC)(Miyoshi et al., 1985; DeVault and Biermann, 1993; DeVault and Grossman, 1992) whereheat is recove red from the hot cond ensate leaving the high tem pera ture condensers andadd ed to the lower tempe rature desorbers; and (3) the dual loop triple-effec t cycle(DeVault, 1988) comprising two complete single-effect loops, recovering heat from thecond enser and absorb er of one loop to the desorber of the oth er loop and generating acooling effect in the evaporators of both loops. Other triple-effect configurations are alsotheoretica lly possible (Alefeld, 1985; Ziegler and Alefeld, 1994). Imp ortant considerations

7/27/2019 [000007]


in com paring th e various systems include not only the energy efficiency of th e cycle butalso its practicality and potential initial cost.

Th e pu rpose of th e pre sent study has been to investigate the possibility of furt herimproving utilization of the high temperature heat available from natural gas combustion.Performance simulation is c onducted for a C effect lithium brom ide-water cycle including

four conde nsers and fo ur deso rbers coupled together, forming an extension of t heconventional double-effect cycle. Based on prior experience, a parallel flow system isused in prefe rence over series flow, and D ouble-Cond enser Coupling (DCC ) is employed,extending from triple-effect cycles, to further improve performance. One goal of thestudy is to investigate the effect of various design param eter s on th e cycle's per form anc e.Som e pa ram etric analysis is conducted which indicates perform ance tre nds.

DESCRIPTION O F T H E 4-EFFECT CYCLE

Figure 1 describes schematically the 4-effect lithium bromide-w ater chiller unde r

investigation. Th e system comprises an evaporator, an a bsorbe r, and fo ur pa irs ofdesorbers/condensers coup led together for internal heat recovery. Th e cycle forms anextension of the conventional double-effect cycle, or of the three-condenser-three-desorb er (3C3D) triple-effect cycle (Grossman et al., 1994). Th e system has 24comp onents o r sub -units (indicated by the circled nu mb ers) and 62 state points (indicatedby the uncircled numbers). Absorber (2) and condenser 5) are externally cooled;desorber (22) is externally heate d. Chilled water is prod uced in evap orator 1). Heatrejected from condenser (6) powers desorber (3), heat from condenser (14) powersdesorber (4) and heat from condenser (23) powers desorber (13). The coupling betweeneach condenser-desorber pair is through a circulating hea t transfer fluid loop, as shown,but may a lso be achieved by physically combining the two components, such that th erefrigeran t condensing on on e side of a heat exchange surface would h ea t u p th e solutiondesorbing on the oth er side of th at surface. The absorb ent solution is in parallel flow,where the wea k (weak in L iBr concen tration) solution from the absorber is split anddivided am ong t he four desorbers. According to simulation results of double-effect cycles(Go mm ed an d G rossm an, 1990) and triple-effect cycles (Grossman et al., 1994), the

parallel flow arrangement is superior in performance to the series flow in terms ofincreased COP an d a lower risk of crystallization. The condensate leaving the condensers(6),(14) and (23) is mixed with th e superheated vapor leaving the desorb ers (3),(4) and(13), respectively before proceeding from each to the next lower-temperature condenser.This method, known as Do uble-Conde nser Coupling DCC)DeVault and Biermann,1993) helps subcool each condensate stream and reject the heat t o a correspondingdesorber. It was shown in an earlier study of triple-effect cycles (Grossman et al., 1994),that the main e ffe ct of this heat recuperation is in providing extra cooling capacity to th eevaporator through the now subcooled refrigerant, at no additional expenditure of highgrade heat. n adde d benefit is a somewhat increased generation capacity of t hedesorbers (3) and (4).

7/27/2019 [000007]


METHODOLOGY OF SIMULATION

A modular computer code for simulation of absorption systems ABSIM)was used to

investigate the perfo rman ce of the cycle und er study. The code, dev elop ed specifically forflexible cycle simulation, has be en described in detail by G rossman and Wilk (1992) andin a related report (Grossman et

al.,1991) containing a user's manual. The modular

structure of th e c ode makes it possible to simulate a variety of a bso rptio n systems in

varying cycle configurations and w ith d ifferent working fluids. The code is based on unitsubro utines containing the governing equations for the system's com pon ents and onprop erty subro utines containing thermodynam ic properties of the w orking fluids. Thecom pone nts are linked together by a main program which calls the u nit s ubrou tinesaccording to the user's specifications to form the complete cycle. When all the equationsfor the entire cycle have been established, a mathematical solver routine is employed tosolve them simultaneously. The code is user-oriented and requires a relatively simpleinput containing the given o perating conditions and th e working fluid a t e ach state point.The user conveys to the computer an image of the cycle by specifying the differentcom pon ents and their interconnections. Based on this information, the code calculatesthe temp erature, flowrate, concentration, pressure and vapor fraction at each state pointin the system and the heat duty at each unit, from which the coefficient of performancemay b e determ ined. The code has been employed successfully to sim ulate a variety ofsingle-effect, double-effect and du al loop absorption chillers, hea t p um ps an d hea ttransform ers employing the w orking fluids LiBr-H 20, H20-NH3,LiBr/ZnBr,-CH,OH,N aO H -H 20 and more. Recently, the s ame code was used to simu late the rath er complexGenerator-Absorber Heat Exchange (GAX) cycle employing amm onia-w ater, in severalcycle variations, an d a variety of triple-effect chillers employing lithium bro mide-wa ter(Grossman et al., 1994).

T he simulation methodology in the present study has followed an ap pr oa ch taken in

earlier studies of single- and double-effect cycles (Gommed and Grossman, 1990), andtriple-effe ct cycles (G rossman et al., 1994). Since the performance of eac h systemdepends on many parameters, the ap proach has been to establish a design point for thesystem, and vary the relevant para m eters a round it. In particular, a perfo rm anc e map of

COP and cooling capacity as functions of desorber heat supply temp eratu re wasgen erated for each system. Thus, th e perfo rm ance of systems in single, double and triplestages could be compared not only a t a single point but over the e ntire tem peraturedom ain applicable to the cycle.

The system's performance under a given set of operating conditions depends, of course,on the design characteristics and particularly on the size of the he at trans fer surfaces in

its exchange units he evapo rators, absorb ers, condensers and desorb ers. As arefer ence case, a p ractical system was considere d with economically reason able, if notoptimized, heat transfer areas. In the earlier study of simpler systems (Gom me d andGrossman, 1990) a single-effect solar-powered lithium brom ide-water chiller known asSAM-15 Biermann, 1978) was selected as a refere nce case. SAM-15 as been testedextensively. xtension of this study to triple-effect systems (Gro ssm an e t al., 1994) hasemployed the sam e approach. He re, a re ference case have been cr eated f or a 4-effect

7/27/2019 [000007]


lithium bromide-water chiller according to Figure 1,with SAM-15 size (specified in termsof its UA) of the evapo rator, absorb er, condensers, desorbers and heat exchangers(recupe rato rs), and with SAM-15 flows of the external fluids. Selec ting the re ferencecase in this mann er m ade it possible to use the results of the presen t sirnulation forcom parison with tho se of the simpler, single-, double- and triple-effec t cycles (Gom me dand Grossm an, 1990; Grossm an et al., 1994), on an equivalent basis. The design

characteristics of the 4-effect reference system a re listed in Tab le 1, including theexterna lly imposed flow rates of cooling and chilled wate r; the weak abs orbent circulationrate; the UA s (overall heat trans fer coefficient times are a), which charac terize the heattransfer performance of the exchange units; and design point temperatures of theexterna l fluids and of the solution outlet from the gas-fired desorber (for this de sorber,unit 22, the externa l fluid loop is redundant). With these values as input, the simulationcode calculates the internal temp eratures, flowrates, concentrations, and oth er operatingparam eters at all the system's sta te points from which overall performan ce param etersmay be derived.

Un fortunately , me asured prope rty data for lithium brom ide-water a re not a vailable in the

litera ture at te mp eratures beyond 210°C (410°F). Prope rties of lithium bromide-w aterfor the simulation were taken from the ASHRAE Handbook (1985) and extrapolated,wh ere necessary, to the high tem pe rature range required by the 4-effect cycle. Th eextrapolation was don e by employing the same correlations given in th e ASHRAE

H and boo k (1985) at th e high tem peratures , beyond their s tated range of validity. A

comp arison of th e properties thu s obtained was carried out later with the higher-temperature LiBr-water data developed recently under the ASHRAE Research Program(Jeter et al., 1992;Lenard et al., 1992), which are valid up to 210°C (410°F). Thedifferences in vapo r pressures a nd specific heat were on the o rder of a few percents, andhence the extrapolations were considered adequate for a first evaluation of the 4-effectcycle. A m ore detailed evaluation leading to actual design will have to rely on more

accura te prope rty da ta that may becom e available in the future.-

RESULTS OF SIMULATION

In conducting th e simulation to g ene rate the ope rating curves of the 4-effect system, thesolution outlet tempera ture from th e gas-fired d esorbe r (22) (sta te point 57) was vanedwhile all the o the r design param eters w ere kept constant. F or the exchange units, it wasassumed that the values of the UA's remain constant while the temperatures and all theother unspecified parameters change. In reality, this is no t strictly accu rate ; although theheat transfer areas (A) remain consta nt, the heat transfer coefficients (U) vary somewhat

with th e te m pera ture s as well as with the loading conditions. How ever, this variation isrelatively sm all in most cases and th e assumption of co nstant UA is a reasonably goodapproximation. B ette r funda me ntal understanding of the com bined hea t and masstransfer process in absorption and desorption would allow taking the variation of UA

with temperature into consideration.

Th e coefficient of performance has been defined here as the ratio of the h eat

7/27/2019 [000007]


quantity in the evap orator producing the desired cooling effect, to that supplied to theexternally hea ted high tempera ture d esorber. The effect of pumping and oth er parasiticlosses is not considered.

Figure 2 describes the COP of the 4-effect cycle as a function of the heat supplytemperature to the externally heated desorber (22), for different cooling water inlet

temperatures, and for a fixed chilled water outlet tem peratu re. The weak solution splitamong the four d esorbers remains even. COP curves for the equivalent double- andtriple-effect, D C C parallel-flow systems with SAM-15 size com ponents (specified in termsof their UA's, per Table l), are plotted along for comparison. It is evide nt tha t allsystems exhibit the same typical, qualitative behavior, with the COP increasing sharplyfrom zero at som e minimum tem peratu re, then levelling off to som e con stant value at ahigher temperature and even decreasing slightly with further increase in temperature.The reason for this behavior is well understood an d is explained in detail in the abovereference (Gommed and Grossman, 1990). Th e 4-effect system has a C OP higher thanth e d oub le- and triple-effect cycles bu t requires a higher minimum hea t sup plytem peratu re in o rder to begin operating. Figure 2 indicates that the double-effect systemperforms best at the hea t supply tem peratu re range of 300-350°F (150-180°C ). Abovethat, from the COP point of view, it is beneficial to switch to the triple-effect system,which perfo rms be st a t the heat supply tempe rature range of 400-450°F (200-230°C).With a still higher hea t supply temperature, a 4-effect system is mo re desirab le.

Figure 3 describ es the cooling capacity of the 4-effect cycle as a function of the heatsupply tem peratu re to th e externally heated desorber (22), for d ifferent cooling waterinlet temperatures, and for a fixed chilled water outlet temperature. The curves for theequivalent double- and triple-effect, DCC parallel-flow systems with SAM-15 izecom pon ents, a re p lotted along for co mparison. It is evident that all systems exhibit thesame typical, qualitative behavior, with the capacity increasing almost linearly with theheat supply temp erature . For each system, the lower the cooling water tem peratu re, thehigher the capacity. Note that unlike the COP which increases with the number ofeffects, the capacity is highest for t he double-effect system and lowest fo r t he 4-effectsystem for the same temperature. This is a direct result of the way the three systemswere created, wit SAM-15 size components, for comparison to e ach o ther. T he sametotal amount of weak solution is distributed more thinly among more desorbers, thehigh er th e num ber of effects, thus producing less refrigerant out of each desorber. U nderthese conditions, a lower capacity is the price one must pay for the higher COP.How ever, th ere is am ple room for optimization of th e solution flowrates an d of the hea ttransfer a rea among th e system's com ponents to improve upon the capacity or the COP,

as will be shown next.

Th e so lution flowrate distribution among the four desorbe rs in the 4-effect system hasbee n selecte d e qu al at th e design point. However, an e qual distribution of solution is notnecessarily optimal. Based on t h e simulation of double-effect systems (Gom me d andGrossman, 1990) and triple-effect systems (Grossm an et al., 1994), an im provement maybe gained by deviating from an equa l distniutio n both in increasing the COP andreducing the risk of crystallization. H ere , t h e effect of varying the solution flow rate to t he

7/27/2019 [000007]


four desorbers has been investigated, with the system operating otherwise at the designcondition, per Table 1. Table 2 lists the results of several runs with different flowdistribution am ong the four desorbers (units 3, 4, 13 and 22), showing in each case thecooling capacity and the COP. Note that t h e sum of the four flowrates is kept co nstant atthe design value of 60 lbs/min (27 kglmin). While Table 2 does not cover the entire rangeof possibilities, it ind icates an optim al maximumCOP) distribution of solution to the

high-, medium- and low-temperature desorbers of approximately 40, 10, 5 and 5 lbs/min(18, 5, 2 and 2 kg/min), respectively. Un der this condition, the CO P reaches 2.177,instead of 2.013 at equal distribution; the solution concentration at the absorber inlet(state point 1) is reduced to 59.2 w t LiBr, compared to 63.5 wt LiBr at equaldistribution. Th e capacity is reduce d som ewhat due to the lower concentration, to 2567.7from 3964.5 Btu/min (45.1 kW from 69.7 kW) at equal distribution. No te tha t theoptimum flow distribution at the design temperatures is not necessarily preserved in off-design conditions. Also in the extrem e cases where e ither of t he four desorbers is starvedfor so lution, the entire system goes out of balance and both the CO P and capacity tendto zero.

It is known from earlier work (Gomm ed and Grossman, 1990) that the flowrate ofsolution has an im por tant effect on performance and an optimum value, since too large asolution flow rate leads to excessive circulation losses and too l ittle is insufficient to supplythe required amount of refrigerant. Figure 4 shows the cooling capac ity norm alized withrespect to the total UA in the system's components (QwaP/vp4,,)nd t he cooling CO Pof the 4-effect system as functions of total weak (pum ped) solution flowrate at stat e point5. Th e system operates otherwise at the design condition (Ta ble l , with equaldis tni uti on of the so lution among th e four desorbers. It is evident that the total solutionflowrate yielding maximum COP is approximately 8.0 lbs/min (3.6 kg/min), deviatingconsiderably from the design condition, with a COP of 2.431 and a capacity of 1556Btulmin (27.3 kw); (QevapjCJAtota, 0.40 F or 0.22 C). However, if the cooling capacity is

to b e maximized, th e optimal solution flowrate is approximately 60 Ibs/min (27 kdm in) asselected for the design condition, with a COP of 2.013 and a capacity of 3964.5 Btu/min(69.7 kw); (QmpjLTAtob, = 1.019 F or 0.566 C).

In addition to capacity and CO P, th e value of QmpjU&w s an interesting performancecriterion, mak ing it possible to com pare systems of different sizes. Th e he at exchange sizefor the components of the 4-effect system may be characterized not only in terms of theirUA's, but also using the effectiveness EFF) r the closest approach temperature (CAT).Figure 5 shows the variation of the cooling COP and Q w a p j U ~ o t l lith the effectiveness,assumed the s am e for all the system's comp onents, units 1-8, 12-14, 21 and 23. Thecooling COP increases with increased effectiveness. However, Qe v a p j L J ~ o b oes through a

maximum at an effectiveness of approximately 0.7. The reason for this maximum is that ahigh effectiveness yields better performance, but at the sam e time re quires larger UA's,

the return for which diminishes at high effectivenesses. Figure 6 describes the effect ofthe closest approach tem perature on the cooling COP and Qwap./Ubobl.t is evident thatthe COP decreases with increasing CAT, quite substantially with CAT'Sgreater than10°F (55°C). The Qwsp/U&b,eaches a maximum approximately at CAT=7.5 F(4.2 C)with a COP of 2.231. The results of these runs suggest tha t th e desorbers of the

7/27/2019 [000007]


base case SAM-15 ize) have been oversized and the absorber undersized for the 4-

effect cycle.

As mentioned e arlier, t h e system's' performanc e und er a given set of opera ting cond itionsdepends on t he design characteristics and particularly on the size of the heat transfersurfaces in its exchange units. As a base case, a practical system was considered with

economically reasonable, if not optimized, heat transfer areas. In search of th e o ptimumsize of the comp onents, several runs were made with d ifferent UA's of the components,presented in Table 3. Th e results show case #6 to give the best COP, cooling capacityand Qevap./UA,ola,mong the test cases studied. Performanc e map of C OP a nd coolingcapacity with ca se #6 UA's as functions of d esorber heat supply tem pe ratu re (Figures 7and 8) show the significant improvem ent over the base case. As can be seen, with someoptimization of the UA's, the COP was raised above 2.2, with approxim ately half th eheat transfer surface of the base case system's components.

TECHNICAL OUTLOOK

The results of th e p resen t sim ulation have shown the 4-effect cycle capa ble of providing aCOP increase on th e ord er of 15% over the equivalent triple-effect cycle (Gro ssm an etal., 1994) - 2.013 vs. 1.724, respectively , at th e design point. The UA investment relativeto the equivalent triple-effect in the base case is an ad ditional 27% (4158 Btu/min. F or131.8 kW/ C to tal UA vs. 3261 Btu/min. F or 103.4 kW/ C, respectively). There is stillroom for optim izing the flow split among the four desorbers, the UA distribution in thesystem etc. which have n ot been fully investigated. However, there are several practicalconsideratio ns which will dete rmine th e comm ercial feasibility of the 4-effect and itscapability to rep lac e the triple-effect cycle:

1. Flue losses: The need to provide a higher firing tem peratu re is associated wit a lowercomb ustion efficiency due to higher flue gas losses. While some of the exhaust heat maybe recovered through an economizer (air preheater), the usefulness of doing this is notclear and must still be determined.2. Corrosion: A higher corrosion rate is expected at the high temperature components(Desorber 22 and Recuperator 21) which may require more expensive materials ofconstruction a nd corrosion inhibitors.3. Heat/mass transfer enhancement additives: The ability of the commonly used additives,suc h,a s 2-ethyl-1-hexanol, to survive at the high temperature is very limited. This is also aproblem, for th at ma tter, in triple-effect cycles and requires further study.

CONCLUSION

Perform ance sim ulation has been carried out for a lithium bromide-w ater chiller basedon the 4-effect cycle. A reference condition was established based on th e com ponentsizes and flowrates of the single-effect SAM-15 system. Performanc e sim ulation wascam ed out over a range of operating conditions, including some investigation of the

7/27/2019 [000007]


influence of the design parameters. A COP of 2.103 was calcula ted a t the design point.The study showed a mp le room for substantial optimization of the C OP, capacity andQmap./U~ota,y varying the flow and UA distribution among th e com pon ents with littleincrease in potential cost.

ACKNOWLEDGEMENT

This work has been supported in part under Oak Ridge National Laboratory SubcontractSOX-SK033V.

LIST O F SYMBO LS AND ABBREVIATIONS

CATCOPDCCE F F

s.p.TH

Q m p .

T C

UA

W O l a l

Closest Approach Tem peratureCoefficient of PerformanceDoub le-Conden ser CouplingH ea t transfer effectivenessEvaporator (cooling) capacityState PointTemperature of solution leaving the ex ternally hea ted , gas-fireddesorber, characterizing the heat supply tem per atu re (e.g. Ts7 nFigure 1)Cooling water supply (inlet) temperature (e.g. T3 nd TB n

Figure 1)Overall heat transfer coefficient times areaTotal UA of exchange units (units 1-8, 12-14, 21 and 23 in

Figure 1)

REFERENCES

Alefeld, G. 1985: Multi-stage apparatus having working fluid and absorption cycles, andmethod of opera tion thereof.'' U.S. Pa tent 4,531,372, July 30.

ASHRAE Handbook of Fundam entals, 1985 :Thermodynamic Propertie s of Lithium

Brom ide-W ater. pp. 17.69-17.70.

Biermann, W. J., 1978: Proto type energy retrieva l and solar system, Bonneville PowerAdministration , Proceedings 3rd Workshop on the Use of Solar Energy for ooling ofBuildings San Francisco, CA, pp. 29-34. Also personal comm unication regarding Can ierSAM-15 solar-po we red water-lithium bromide absorption chiller, July 1986.

DeV ault, R.C. an d Biermann, W.J., 1993: Triple-effect abs orp tion refrigeratio n systemwith dou ble-co nden ser coupling . US Pa ten t 5,205,136.

7/27/2019 [000007]


DeVault, R.C., 1988: Triple-effect absorption chiller utilizing two refrigerant circuits .US Patent 4,732,008, March 22.

DeVault, R.C. nd Grossman, G., 1992: 'Triple-effec t absorption chiller cycles .Presented at the International Gas Research Conference IGRC92, Orlando, Florida,November 16-19.

Gommed, K. and G rossman, G., 1990: Perform ance analysis of staged ab sorption heatpumps: Water-lithium bromide systems . ASHRAE pape r No. AT-90-30-6, presented atthe 1990 ASHRAE Winter Meeting, Atlanta, GA, February 10-14.

Grossman, G., Gommed, K. and Gadoth, D., 1991: A comp uter model fo r simulation ofabsorption systems in flexible and modular form, ORNL,/sub/90-89673, Oak RidgeNational Laboratory. Also ASHR AE pape r No. NT-87-29-2, presented at the ASHRAE

Annual Meeting, Nashville, TN une 27-July 1, 1987.

Grossman, G. and Wilk, M., 1992: Advanced modular simulation of absorption systems .Presented at the 1992 ASM E W inter Annual Meeting, Anaheim, CA, November 8-13.

Grossman, G., Wilk, M. and DeVault, R.C., 994: Simulation and perform ance analysisof triple-effect absorp tion cycles. Presented at th e ASHRAE Winter M eeting, NewOrleans, L ouisiana, January 22-26.

Jeter, S.M., Moran, J.P. and Teja, AS., 1992: Properties of lithium bromide-watersolutions at high tem peratu res and conce ntrations - Part Ill Specific Heat. ASHRAETransactions, v. 98 pt 1, 137-149.

Le nard, J.L.Y., Je te r, S.M. and Teja, A.S., 1992: Properties of lithium bromide-watersolutions at high tem peratu res and concen trations - Part Tv: apor pressure. ASHRAETransactions, v. 98 pt 1, 167-172.

Miyoshi, N., Sugimoto, S. and Aizawa,M., 1985: Multi-Effect Absorption RefrigeratingMachine , US Patent 4,551,991, November 12.

Oouchi, T., Usui, S., Fukuda, T. and Nishiguchi, A, 1985: M dtiS tage A bsorptionRefrigeration System . U.S. Patent 4,520,634, June 4.

Ziegler, F. and Alefeld, G., 1994: Comparison of multi-effect absorption cycles.Proceedings of the biternational Absolption Heat Pump Conference New Orleans,Louisiana, January 19-21, ASME vol. AES-31.

7/27/2019 [000007]


TABLE 1Characteristic Parameters at Design Point for 4-Ef€ectLiBr-H,O Absorption Chiller

Heat Transfer Characteristics UA):Absorber:Desorbers:Condensers:Evaporator:Recuperative Heat Exchangers:

193.0 Btu/min. F (6.118 kW/ C)

268.0 Btu/min. F (8.496 kW/ C)565.0 Btu/min. F (17.911 kW/ C)377.0 Btu/min. F 11.95 1 kW/ C)64.0 Btu/min. F (2.029 kW/ C)

Mass Flow Rates:Absorber (cooling water)Low Temperature Condenser (cooling water)Eva porator (chilled water)Internal Coupling Water Loops, s.p. 10-11, 15-16 and 35-36Weak Solution

Solution split evenly among the four desorbers, each 15.0 Ibs/min (6.75 kg/min)

483.0 lbs/min (219 kg/min)391.0 lbs/min (178 kg/min)

300.0 lbs/min (136 kg/min)400.0 lbs/min (182 kg/min)

60.0 lbslmin (27 kg/min)

Temueratures:Hot solution outlet from gas-fired desorber (22) (s.P. 57)

Cooling wate r inlet (s.P. 3 and 23)Chilled water outlet (s.P. 29)

600°F (315°C)

85°F (29°C)45°F (7°C)

7/27/2019 [000007]


TABLE 2Effect of SolutionDistriiution Among Desorbers in a

4-Effect LBr-H,O Absorption Chiller at TH = 600°F 315°C)from lowest to highest temperature generator (left to right),

Conversion factors: kg/min = 0.454 x Ibs/min and k W = 0.01757 x Btu/min

uni t 3 Unit 4 Unit 13 Unit 22 QWP.. COPmass flow mass flow mass flow mass flow (Btu/min)

(lbs/min) (lbs/min) (lbs/min) (lbs/min)s.p. 8 s.p. 13 s.p. 33 s.p. 53

5 5 15 35 3294.9 1.5578

10 15 15 20 4019.7 1.9250

15 15 15 15 3964.5 2.0131

20 15 15 10 3663.1 2.0750

30 10 10 10 3496.6 2.1374

35 10 7.5 7.5 3129.7 2.1670

40 10 5 5 2567.7 2.1768

45 5 5 5 2419.8 2.1527

35 15 5 5 2600.7 2.1646

7/27/2019 [000007]


Table 3. Effect of UA distriiution among the heatexchange units in a 4-effect DCC arallel flow system at a

fixed total solution flowrate of 60 bs/min or 27 kg/min (equal distriiution) and fixed heat supply temperature of

600°F 315 C), Conversion factors: kW/ C = 0.0317 x Btu/min. F, kW = 0.01757 x Btu/min and A C = A°F/1.8

Unit No. Unit UA UA UA UA UA UA UA

type base case Case #1 Case #2 Case #3 Case #4 Case #5 Case #6

1

2

3

4

5

6

7

8

12

13

1421

23

Evap.

Abs.

Des.

Des.

Cond.

Cond.

HX

HX

HX

Des.

Cond.HX

Cond.

377.0

193.0

268.0

268.0

565.0

565.0

64.0

64.0

64.0

268.0

565.064.0

565.0

377.0

193.0

150.0

150.0

200.0

200.0

100.0

100.0

100.0

150.0

200.064.0

200.0

377.0

193.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.064.0

100.0

377.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.064.0

100.0

377.0

250.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.064.0

100.0

377.0

300.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

64.0

100.0

377.0

400.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.064.0

100.0

Total (Btu/min. F) 3890.0

2.01303964.5

2184.0

2.11173733.6

1634.0

2.06173322.4

1541.0

1.92122291.9

1691.0

2.09803715.1

1741.0

2.11903977.1

1841.0

2.14604357.1

1.02 1.71 2.03 1.49 2.20 2.28 2.37

7/27/2019 [000007]


LIST OF FIGURES

Figure 1: Schematic description of 4-effect chiller in parallel flow

Figure 2: C OP fo r doub le-effect, triple-effectt and 4-effect DC C parallel flow LiBr-H,Osystems as a fun ction of heat supply tempe rature (TH) for different cooling watertemperatures (TC) and a chilled water temperature fixed at 45°F (7.2 C),Conversion factor: C = ( O F -32)/1.8

Figure 3: Cooling capacity for double-effect, triple-effect, and 4-effect DCC parallel flowLiBr-H,O systems as a function of heat supply tem peratu re (TH) for different coolingwater temperatures (TC) and a chilled water temperature fixed at 45°F (7.2 C),Conversion factors: C = (OF -32)/1.8 and kW = 0.01757 x Btu/min

Figure 4: COP an d normalized cooling capacity for 4-effect DCC parallel flow LiBr-HzOsystems as a function of total solution flo w at e (equal distribution) at a fixed he at supply

temperature (TH) of 600°F (315°C ) and fixed UA's,Conversion factors: kg/min = 0.454 x lb/min and A C = A F/1.8

Figure 5: CO P an d normalized cooling capacity for 4-effect D CC parallel flow LiBr-H,Osystems as a function of effectiveness (EFF) at a fixed total solution flowrate of 60lbs/min or 27 kg/min (equal distribution) and fixed heat supply temperature (TH) of60 0°F (315 C), Conversion factor: A°C = A F/1.8

Figure 6: CO P a nd normalized cooling capacity for 4-effect DCC parallel flow LiBr-H,Osystems as a function of closest approach te mp erature (CAT) at a fixed total solutionflowrate of 60 Ibs/min or 27 kdm in (equal distribution) and fixed h eat supply

temperature (TH) of 600°F (315 C), Conversion factor: A C = A F/1.8

Figure 7: CO P for 4-effect base case and 4-effect optimum case (#6 per T abl e 3) DCCparalle l flow LiBr-H,O systems as a function of hea t supply tem peratu re (TH) fordifferent cooling water temperatures (TC) and a chilled water temperature fixed at 45°F(7.2 C), Con version factor: C = ( O F -32)/1.8

Figure 8: Cooling capacity for 4-effect base case and 4-effect optimum case (#6 perTable 3) D C C pa rallel flow LiBr-H,O systems as a func tion of he at supply temp era ture(TH) for different cooling water temperatures (TC) and a chilled water temperaturefixed at 45°F 7.2 C), Conversion factors: C = (OF -32)/1.8 and

kW = 0.01757 x Btu/min

7/27/2019 [000007]


Figure 1

7/27/2019 [000007]


OWL-DWG 94-1750

2 0 .-

_ _

I

Legend:

..._.._..-..- Triple effect

4-eff ect

Double effect

.._.-----

Figure 2.

7/27/2019 [000007]


7000

6000

5000

4000

3000

2000

1000

100 200 300 400 500

TH (“F)

Legend:

...-..-..-..- Triple effect

4-0ffeCt.-.-.-.-

Double effectI I

Figure3.

600 7 800

7/27/2019 [000007]


3.0 , I 3.0

qr evap. ' * total

0

0 20 40 60 80 100

Total Flow (Ib/min)

Figure 4.

h+

m4

a?dm>Q,

I

7/27/2019 [000007]


CJ)

.-

-l .E-.

COP -

EFF

Figure 5.

7/27/2019 [000007]


3

2I

1

i

5 10

CAT T)

Figure 6.

15

7/27/2019 [000007]


2.5

2.0

1.5

1 o

O R V L - ~ W G 4-17009

I

TC = 75°F

i

TC = 85°F optimum

TC = 85°F

ii*

iJi

TC = 95°F

1 1 1 1 I 1 , , I I I , , . . ' , . . I . , . I , ( ,

100 200 300 400 500 600 7 00 800 900 1000

TH ( F)

................. 4-effed,optimized size of components

..-.-.--- 4-0ffed,base

I

Figure 7

7/27/2019 [000007]


.-

Legend:

................. 4-effect optimized size of components

..---.-.- 4-effect base case

I000

i-

4000 '

20003 t000

TH OF)

Figure 8

[000007]

Documents