Computer Modeling of a Single-Stage Lithium Bromide/Water Absorption Refrigeration Unit FL. Lansing DSN Engineering Section Solar-assisted refrigeration systems have been proposed for application in Goldstone energy consercation projects. This article describes and analyzes the dynamic simulation and the computer modeling procedure of one of these systems, namely a lithium bromide/water absorption refrigeration system. A new ana- lytical expression that fits the three-dimensional surface of LiBr concentration, refrigerant temperature and solution temperature in the range of interest from 0.50 to 065 kg LiBr/ke solution is presented with a standard deviation of 0.2 percent. This will save considerable computing time and effort required for evaluation of system performance. A numerical example from typical running conditions is added to show the relative weight of each parameter used together with the sequence of program steps followed. The results from this simulation are heat rates, line concentrations, pressures and the overall coefficient of performance. |. Introduction nonvolatility ofthe absorbent (LiBr), allowing only ‘water vapor to be driven off the generat In recent years, the lithium bromide/water absorption ae riven ee system has become prominent in refrigeration for air (3) Less pump work is needed compared to other units conditioning. It possesses several advantages over the due to operation at vacuum pressures other types of absorption systems, such as: (1) Ithas the highest coefficient of performance (COP) _On the other hand, the lithium bromide/water absorp- compared to other single-stage absorption units at __tion system has some drawbacks such as: eee ae (1) Its limited to relatively high evaporating tempera- (2) It is composed of simpler components since it can tures since the refrigerant is water. This means that work efficiently without the need of rectification ‘evaporation temperature above 0°C must generally columns. A basic generator is sufficient due to the be satisfied to prevent flow freezing, JPL DEEP SPACE NETWORK PROGRESS REPORT 42:32 287ystallization of LiBr salt at moderate concentra- s (065 kg LiBr/Kg sokution) will tip off the eycle range of operation. (3) The systems have to be designed in hermetically they operate at vacuum pres- sures, Improper operation would result if leakage of air into the system occurred Inrespective of its drawbacks, the LiBr/water unit is still considered the most economical for this kind of refrigeration technique. It has been selected as a candi- date refrigeration system in connection with proposed solar assisted equipment for application in Goldstone energy conservation projects. This article describes and analyzes the computer ‘modeling of such units. The modeling procedure is gen- eralized to enable those concemed with use or evaluation of cycles employing this material to save considerable time and effort required for calculations. Il, Thermodynamic Cycle ‘The system components and the working fluid states are shown in Figs. 1 and 2. There are four basic heat exchanger surfaces: the evaporator, the absorber, the generator, and the condenser, in addition to a liquid- liquid heat exchanger. Lithium bromide is, basically, nothing more than salt water. However, lithium bromide is a salt with an especially strong attraction for water. The cycle of operation may be started as shown ii Fig. 1, from the evaporator. The refrigerant (water) is evaporated while it is taking heat from the fluid being chilled (air for instance). The water vapor (state 10) is then sucked up by lithium bromide spray injected into the absorber, thus the name absorption system. Due to the exothermic reaction taking place in the absorption process, heat has to be removed, and the mixture of lithium bromide and refrigerant vapor at this stage is called the “strong solution” (state 1). “Strong” and “weak” refer to the amount of refrigerant present. The strong solution is then pumped (state 2) through a liquid-liquid hheat exchanger (state 8) to the generator. This heat ex- changer will improve the cycle performance, as will be shown later. In the generator (sometimes called the con- centrator) the strong solution is heated and boiled by an external heat source to release the refrigerant vapor (state 7), leaving behind a concentrated LiBr/water sohu- tion (state 4). The latter is ealled “weak solution” since 248 it contains a smaller amount of refrigerant. The refrig- cerant vapor leaving the generator is condensed (state 8) in the condenser and is directed to the evaporator through ‘an expansion valve (state 9). The weak solution flows bback to the absorber through the liquid-liquid heat ex- changer as a spray (state 6) to complete the cycle. The thermodynamic eycle is sketched on the equilib- rium temperature-pressure-concentration chart as shown, in Fig. 2. It is bounded by two constant concentration lines: X, and X, for the absorber and generator concen- tration, respectively, and two constant pressure lines: P, and P, for the evaporator and condenser pressures, respectively. For an efficient air conditioning application, the evapo- rator temperature t. should be low enough to dehumidify the air, In practice, it ranges from 4.5 to 10°C according to the load conditions. The heat rejection temperatures t, and t- for the absorber and the condenser, respectively, vary according to the type of cooling medium (air, water), ambient conditions of dry and wet bulb temperatures, and whether their cooling lines are connected parallel with each other. The generator temperature t, depends on the source of heat available, ic. solar, gas or steam, However, a mini- ‘mum temperature of 80°C should be maintained to pro- vide efficient operation ‘The operational function of the liquid-liquid heat ex- changer in the cycle will be the reduction of the weak solution temperature t,, leaving the generator and increas- ing the strong solution temperature ¢, leaving the absorber. ‘The result is that it reduces the heat input necessary to the generator while reducing the heat rejected from the absorber by heat exchanging between the two solutions. This heat exchanger does increase the coefficient of per- formance of the system and is therefore, always desirable. Ill. Thermodynamic Properties Enthalpies of the water (refrigerant) and LiBr (absor- bent) solutions were calculated with reference tempera- ture at 25°C (Refs. 1-4), The following expressions were found essential to the calculation ofthe cycle performance: (1) The enthalpy of pure water liquid at temperature PC kcal/kg @ JPL. DEEP SPACE NETWORK PROGRESS REPORT 42-32(2) The enthalpy of saturated water vapor at tempera ture #°C = (5728404171) keal/kg @ (latent heat = 597.8 — 0.583¢ -+0.4 kcal/kg) (8) The enthalpy of superheated steam at temperature #°C and at a pressure equal to the saturation pres- sure of steam at temperature t,°C = (5728 + OAITt,) + 0.48 (¢ ~ 1) = 5728 + 046¢~ 0.0481, kcal/kg @) by taking the specific heat of water vapor = 0.46 eal/kg°C at very low pressures (0.01 to 0.1 atm). (4) ‘The specific heat of lithium bromide/water solution of concentration X kg LiBr/kg solution is given by 01-1234 048K! keal/kg solution a (8) The enthalpy of LiBr/water solution of concentra- tion X kg LiBr/kg solution at 25°C is H15 = 68.08 — 456.67 X + 416.67X? kcal/kg solution ) (6) The enthalpy of LiBr/water solution of concentra- tion X kg LiBr/kg solution at temperature #°C is given by Hy.as + Ca(t ~ 25) = (4281 ~ 495.92 X + 404.67 X*) + (101 — 1.28X + 048.x2) (0) ® (7) Im the range of concentration from 050-065 kg LiBr/kg solution, the author found that it is possi- ble to fit the refrigerant temperature te, the satu- rated solution temperature fy, and the concentration X by the relation (te°C) = (49.04 — 134.65 X) + (1.125 ~ 04TX) (ta°C) JPL DEEP SPACE NETWORK PROGRESS REPORT 42:32 with a standard deviation 0.2%. This may be rewritten as 49.04 + 1,125 ty — te 1365 + OTF, m ie, X is completely defined by the set {teitw}. (8) The saturated vapor pressure P in mm Hg corre- sponding to saturation temperature T°K for pure water is given by (Ref. 5) 1555 logePnm ig = 75550 — 1555 _ HBX I0% 7 8) IV. Performance Calculations ‘The determination of the thermodynamic properties of each state in the cycle, the amount of heat transfer in each component, and the flow rates at different lines de- pend on the following set of input parameters: Generator temperature t,, °C. Evaporator temperature f., °C Condenser temperature t., °C Absorber temperature fe, °C Liquid liquid heat exchanger effectiveness E,, Refrigeration load Qr, tons The above set can be determined from actual running measurements or assumed by a first reasonable estimate to cycle performance. ‘Together with the assumptions of neglecting the pump- work and neglecting the pressure drop in components and lines and assigning saturation conditions to states numbers 1, 4,8, and 10 in Fig. 1, the properties are deter- mined as follows: ‘A. Absorber Concentration This is determined by Eq, (7) using t, for the solution temperature and t, for the water temperature correspond- ing to the evaporator pressure P.: _ 49.04 + L195 ~ te Xstrong = joluten 3408 + OATE X=X =X kg LiBr/kg solution (9) 289B. Generator Concentration This is evaluated, from Eq, (7), using f, for the solution temperature and t. for the refrigerant temperature cor- responding to the condenser pressure P, Xweak = 1904 +15 =f “Solution BEBE OATE kg LiBr/kg solution (10) Te may be noted that X, is always larger than X,, and MaMa ero ky LiBr/kg solution ayy for the pure water flow in the condenser and evaporator. C. Pressure Limits in the Cycle Using Eq, (8), it is possible to evaluate the pressure in every line as follows: Pangoraton Pe = Py = Py = Po = Pr in-mm Hig is given by eS 11.2414 X 106 ge Pe = 7, TFRs G+ TIE (12) and the condenser pressure, P. P, = Py in mm Hg is given by hog? = 1855 -—_2585__ 24d x 10° O80 EF) h FTF as) D. Flow Rates Enthalpy of saturated liquid water, state 8, is given by aq, (1) at the condenser temperature te as H, = (te 25) keal/kg as) The throttling processes from 8 to 9 and that from 5 to 6 give (15) Enthalpy of saturated water vapor, state 10, is given by Eq. (2) at the evaporator temperature te as Hy = 572.8 + OAITt, (a6) 250 Applying the first law of thermodynamics to the evapo- rator will give Qe = me (Hie ~ He) where me is the refrigerant fow rate, equals the differ- ence between the strong and weak solution rates. By using Eq. (15) “He -e an On the other hand, the lithium bromide mass balance in the absorber gives myXe + maXwo = MX, = (Me + me)X, and by using Eqs. (11) and (17), then (atin) (REx as) "=o (as) ry Since the concentrations X, and X, are restricted not to exceed certain limits to avoid erystallization problems, and if the temperatures ofthe eycle are set to vary according to ambient and load conditions, the mass flow rates in the different lines will be varied accordingly. This necessi- tates the existence of LiBr and water solution inventories to be used for Low compensation, especially at times when variations of load, hot water temperature, and cool- ing water temperature do occur, E. Liquid-Liquid Heat Exchanger Temperatures Once the heat exchanger effectiveness Ey, the mass flow rates (my.m,) and the concentrations (X,,X,) are «given, it is possible to determine the solution tempera- tures f, and f, from Fig, 3 as follows: Based on the weak solution side, which has the minimum heat capacity, the effectiveness B,, is defined by (Ref. 6) neato (208) or based on the strong solution side y= Lt Cn) t= (et (iw Cxa) (te = te) JPL DEEP SPACE NETWORK PROGRESS REPORT 42:32where Cx, is specific heat of the strong solution whose concentration is X;, and Cx, is the specific heat of the weak solution whose concentration is X,, Both Crs, Cre are determined from Eq, (4) as Cx, = LOL ~ 1.28 X, + 048 Xt | (2) Cx. = 101 ~ 128K, +0483 J Equations (20) and (21) are rewritten using Eqs. (18) and (19) to give the temperatures t, and f, as: t= te — Bully ~t (20b) and ty “Lee The enthalpies H, and H, are then calculated using Eq, (6) as follows: 12.81 ~ 495.92 X, + 404.67 X1) + (LO1~ 123%, + 048X3)-(64) (28) H, = (42.81 — 495,92, + 404.67 X3) + (LO1~ 128%, +048X2)-(¢,) (286) F. Heat Transfer in Condenser, Generator, and Absorber ‘The enthalpy of water vapor leaving the generator and entering the condenser (state 7) is determined by Eq. (3) H, = 5128 + 0.481, ~ 00131, (m) The heat balance of the condenser gives, Qe = malt — Hs) (25a) or using Eq, (17) for my expression, it becomes Q Qe = oa apy (Hs ~ He) (25b) Heat balance for the combined generator and heat exchanger control volume gives Qe = Mel, + mgH, — mH, (26) JPL DEEP SPACE NETWORK PROGRESS REPORT 42:32 Since the pumpwork is negligible, then H.~H, 7) Using Eqs. (17), (18), (19), and (27), itis possible to write Qo as _ Oe fk Oe amg ee S Heat balance of the absorber gives Qy as Qa = mH + malo ~ malls Using Eqs. (15), (17), (18), and (19), Qu is rewritten as (29) Equations (25b) and (29) are governed by the first law of thermodynamics in the form Qe + Qn = Qo + Oe (30) G. Coefficient of Performance (COP) This is defined as _ refrigeration effect _ Qe COP = Stemal heat input ~ Qr It is simply derived from Eq, (28) as (Hy — Hs) (Xe — Xs) CORR RR XA] &) H. Ideal Coefficient of Performance ‘The maximum coefficient of performance of the above absorption eyele is given by: (32) where T., To, Te, and T, are the absolute temperatures of the evaporator, absorber, condenser, and generator, respectively. 251‘The ratio (COP Jeena (COP nae is called the ‘relative performance ratio,” to show the deviation from reversible cycle operation. Main Components Modeling Each of the four basic heat exchangers (condenser, absorber, generator and evaporator) is considered as a “constant temperature heat exchanger” as shown in Fig, 4. This is due to the fact that the heat transfer mechanism involves a change in phase while the temperature of one of the heat transfer fluids is kept constant. ‘There are two basic approaches to determine the heat transfer characteristics of each heat exchanger, namely, (1) Using the conventional logarithmic mean tempera- ture difference expression, (2) Using the combined effectiveness/number of ex- changer heat transfer units N,,, the latter defined as (Ref. 6) UA ‘Smaller Teat capacity of the two heat transfer fluids (33) Both approaches give a straightforward solution and there exists a one-to-one correspondence between the two sets of parameters in each case, Because of lack of sufficient and reliable information| about the overall conductance coefficient U for a LiBr- ‘water solution, an approximate modeling trial was made by Wilber et al. (Ref. 7) in terms of a “characteristic product” (UA) associated with each heat exchanger. The latter was determined from the temperature pattern at nominal design conditions, in spite of the fact that the ‘conductance U is a dominant function of flow rates and uid properties for a laminar flow. (On the other hand, Lackey (Ref. 2) indicated that an increase of LiBrawater flow rate by as much as 350 per- cent resulted in a change of the overall temperature pat- tern of only 6 percent and the product (UA) was no longer considered to be a fixed property at all operating conditions. The above suggests that the main components, modeling procedure would be best characterized by fixed temperature differences AT, and ATs, as shown in Fig. 4, rather than a fixed (UA) product. Consequently, the flow 252 rates of the externally heating or cooling fluids would be self-controlled to suit the variations in heat transfer rates. ‘A good practical estimate of the inlet temperature differ- ence AT, and the outlet temperature difference AT is 10°C and 3°C, respectively, each for the generator, absorber and condenser and 20°C and 6°C, respectively, for the air-cooled evaporator. VI. Liquid-Liquid Heat Exchanger Modeling This is considered as an “unbalanced counter flow heat exchanger,” since the two relevant streams possess un: equal heat capacities. The exact effectiveness expression (Ref. 6) as applied to LiBr-water solutions indicated that extreme off-design changes of flow rates or concentrations can change the design effectiveness value of Ex by only £5 percent. The consequent effect on the sohution enthalpies leaving the heat exchanger, Eq. (23), is negli- gible. This suggests that the effectiveness E, may be taken as a constant in the analysis without great loss of accuracy. Vil. Summary The sequence of program calculations is summarized as follows: Input Data (1) ty, °C, generator temperature 2) te °C, evaporator temperature (8) te, °C, condenser temperature (A) ta, °C, absorber temperature (5) Ex, exchanger effectiveness (©) Qr, keal/hr, load Steps of Analysis _ (49008 + 1.195%, ~ t,) () X= eres oat) ME LB4/kg solution _ (49.04 + 11254, @) x. Toa KE MBr/g solution * IFO5 < (X,and X,) < 0.65 proceed, else stop. (8) Hy=(t.— 25) keal/kg (A) Hag = (5728+ 0417.) keal/kg Qe 8) me= ey keh JPL DEEP SPACE NETWORK PROGRESS REPORT 42:32©) m= mere zy kesbr x, =% (1) t= the x kg/hr (8) ts Exlty— ta) °C (9) Cx, = 101 — 128%, +048 keal/kg*C (10) Cz, = 1.01 - 1.28%, +0482 keal/kg°C an a=n+[ere gee —w)] °c (12) Hy = (42.81 — 495.92.X, + 404.67 Xi) + (L01~ 123%, +0481) +t) keal/kg (13) H, = (4281 — 42592X, + 404.672) + (LOL 128X, +048X1)+(t.) —_keal/kg (14) Hy = (572.8 + 0.46, — 0.0431.) keal/kg. (15) Qe= me(Hy — Hs) keal/ar (18) Qo = (malls + maby — mH) —eal/br (1D) Qa = (tmeHly + migFlse — mH) keal/ar (18) COP = Q1/Qe {te + 278.15) (ty ~ te) (19) (COP wm = = RBI. —F) (20) relative performance ratio = COP/(COP) maz 1555 Tyme (22) Pe anti [78558 . _ M214 X 109 ag GF aay | mts Vill. Numerical Example The following numerical example presents the perfor- mance characteristics of atypical running condition: Input Data = 40°C, JPL DEEP SPACE NETWORK PROGRESS REPORT 42.32 8 ton of refrigeration = 3024 r= kkcal/hr Cooling fluid lines to the condenser and absorber are in parallel. Calculation Steps X, = 0.5672, kg LiBr/kg solution X, = 026233, kg LiBr/kg solution Hy = 15, keal/kg, m, = 59.9199, kg/hr 5268, kg/hr °C 49677, keal/kg°C = 0.42982, keal/kg°C 73.52, °C 49.9124, keal/kg 45,9594, keal/kg HY, = 612.48, keal/kg Qc = 8222.3, keal/hr Qo = 9896.9, keal/hr 1698.6, keal/he cop (COP woe relative performance ratio = 0.664 P.= 7.45 mmHg P. = 55.87 mmHg The temperature profile of each heat exchanger is sketched as shown in Fig. 5. Having established an ana- lytical procedure for the performance characteristics of & lithium bromide/water absorption system enables the study of further changes in the thermodynamic cycle. Examples of these changes are the temperature of entering heating fluid to the generator, the cooling load at the ‘evaporator, and the temperature of entering cooling fluid at the condenser or absorber.Definition of Symbols ‘A. heat exchanger surface area, m* UT overall conductance cocficient, keal/(br-m’-°C) Cy. specific heat of solution of concentration X, X_ solution concentration, kg LiBr/kg solution eal kg solution °C Ey, effectiveness of liquid/liquid heat exchan; ctomied * a : 2A. absorber H_ specific enthalpy, keal/kg m_ mass flow rate, kg/hr eC condenser @E evaporator pressure, mm Hg fC generator © rate of heat transfer, keal/he “strong” solution £ temperature, °C wo “weak” solution References 1, Ellington, R. T,, Kunst, G., Peck, R. E., and Reed, J. F,, “The Absorption Cool- ing Process.” Institute of Gas Technology Research Bulletin No 14, Chicago, ML, Aug. 1957. 2. Lackey, R. S,, “Solar Heating and Cooling of Buildings,” Westinghouse Electric Corporation Report to the National Science Foundation, NSF-74-C584-1-2, Appendix J., May 1974, 8. Threlkeld, J. L., Thermal Environmental Engineering, Prentice Hall, 1970. 4, San Martin, R. L., and Couch, W. A., “Modeling of Solar Absorption Air Con- ditioning,” Institute of Environmental Science Proceedings, Vol. 1, 1975, pp. 186-189, 5. Pennington, W., “How to Find Accurate Vapor Pressures of LiBr-Water Solu tions,” Refrigeration Engineering, May 1955, 6. Kays, W. M, and London, A. L., Compact Heat Exchangers, McGraw Hill Book Co., Inc, 1958, 7. Wilbur, P. J, and Mitchell, C. E., “Solar Absorption Air Conditioning Alterna- tives,” Solar Energy, Vol. 17, 1975, pp. 193-199. 254 JPL DEEP SPACE NETWORK PROGRESS REPORT 42:32Genewaton zeae CONDENSER ® : ‘ 1 6 Ua : to__| evaronaron Fig. 1. Flow diagram for a lithium bromide)water absorption system JPL DEEP SPACE NETWORK PROGRESS REPORT 42:32 255\YAPOR PRESSURE, wn FEFRIGERANT TEMPERATURE, °C FEFRIGERANT PRESSURE, nig Fig. 2. 6 an a itn ® CONSTANT ESNeeNTRATION xy une EES FERIGERANT TEMPERATURE, °C SOLUTION TewPenaTURE, °C Equilibrium chart for lithium bromidelwater solution JPL DEEP SPACE NETWORK PROGRESS REPORT 42:32WEAK" SOLUTION ‘ a g MAXIMUM RANGE & i ete Bs = TStRONG" soLUTION| HEAT EXCHANGER LENGTH = Fig. 3. EMfectiveness of liquid-Hiquid heat exchanger (6) CONDENSE® &) some ‘CONDENSER TEMPERATURE ABSORBER TEMPERATURE ed OSORBER TEMPERATURE Cooune suv iNT ine (0 GeneraTor (a) evaroRaTOR suet wer q vextine we HEATING FLUID (AI) a, , oh Hl ae ah GeReINTOR TPEATIR, F ! Fig. 4. Temperature profiles along the four main components: condenser, absorber, generator and evaporator JPL DEEP SPACE NETWORK PROGRESS REPORT 42.32 (o) CONDENS (&) assomer 12wre 1 2mre { | ref we we we CoouNs VE t L ae Be EoouNe FLUID (e) GeNetaTOR (@ evaromton | eel (6) LIQUID - LQUID HEAT ExcHANGER sore Fig. 5. Temperature patterns along the heat exchangers for the numerical example 257