PressurHnthalpy cha s for mixtures of oil and refrigerant R 12

D. W. Hughes, J. T. McMul lan , K. A. Mawhinney and R. Morgan

Key words: refrigerant, oil, mixture, pressure, enthalpy

Diagrammes pression- enthalpie pour les m61anges d'huile et de R 1 2

Les donn#es de I'ASHRAE pour le R 12 sont combin#es avec celles de Bambach pour les m#langes d'huile et de frigorig#ne en vue de I'#tablissement, ~ /'aide d'un ordinateur, de diagrammes pression-entha/pie des mglanges. Les

diagrammes prgsent#s concernent des m#langes contenant O. 1, 2.0 et 15.0% en masse d'hui/e.

L'emploi de ces diagrammes permet de calculer /a performance des circuits de machine frigorifique et de pompe ~ chaleur en particulier dans /es cas d'utilisation de compresseurs ~ vis noyds d'huile ou de compresseurs rotatifs ~ aubes. On montre que /a pr#sence d'hui/e aboutit g#ndralement ~ la r#duction du d#bit et du COP. Ces effets ne seront pas forc#ment apparents si /'on essaie les compresseurs en uti/isant /a m#thode de 'boucle de vapeur'.

PressurHntha lpy charts are presented for refrigerant R 12-oil mixtures, for a range of oil concentrations,

The effect that the existence of this mixture

has on evaporator performance and COP of a refrigeration/heat pump system is discussed and some results are compared with the pure refrigerant cycle normally considered.

• The solubility of fluorinated refrigerants in lubricating oil is a phenomenon which has been recognised for a long time. It is widely recognised that the presence of oil in a vapour compression refrigeration or heat pump system reduces its capacity. This is because, at the evaporator outlet, liquid refrigerant remains dissolved in the oil. thus being unavailable for evaporation, and hence to carry latent heat. Previous work on the subject has looked at the influence of oil on the evaporator heat transfer performance, 4 and at its effect on system capacity. 2 However it has not been generally appreciated that the presence of oil in a refrigeration or heat pump system would have a serious deleterious effect on the coefficient of performance.

In a previous paper s it was shown how, theoretically, the COP can be reduced by as much as 30% in systems which require relatively high oil- refrigerant ratios, such as those using rotary sliding- vane or screw compressors. This results from the

The authors are from the Energy Study Group, New University of Ulster, Coleraine. Northern Ireland, UK. Paper received 4 January 1982.

fact that, while evaporator capacity is reduced, compressor power is largely unaffected by the oil circulation. The analysis set out in this previous paper was based on the oil-refrigerant solubility equation developed by Bambach 1 for R 12- paraffinic oil mixtures. Subsequent experimental work by the authors 6 verified the theoretical analysis.

It is worth summarizing the effects of oil with various fluorocarbon refrigerants. Liquid R 1 2, and certain other refrigerants, are totally miscible with oil, the mixture forming a single phase at all temperatures and pressures. Other refrigerants, such as R 22, display limited solubility, and separate oil- rich and refrigerant-rich phases may exist. Even with these refrigerants, however, there is a critical solution temperature above which complete miscibility again appears. The effects of this miscibility are to: one, change the working fluid from a pure refrigerant with well known properties to a poorly understood mixture with properties that depend on the oil type and concentration. Two, affect the heat transfer processes in the evaporator and condenser. This can either improve or impair

0140-7007/82/0401 99-0483.00 © 1982 Butterworth & Co (Publishers) Ltd and IIR Volume 5 Num6ro 4 Juillet 1982 199

heat transfer, depending on the oil concentration 4. Three, elevate the boil ing point of the mixture above that of the pure refrigerant, as described by Raoult's Law. Four, reduce the heat carrying capacity of the mixture during evaporation. This is because the oil holds a proportion of the refrigerant in the liquid phase.

Our interest in these phenomena was aroused when it proved to be impossible to reproduce the manufacturer's published performance data for a particular range of compressors which required a high oil charge (about 10%). Several effects were observed, including a clearly defined optimal evaporator superheat setting and a compressor coefficient of performance some 25% below that indicated by the manufacturer's data. This latter observation was very worrying and led to our undertaking a lengthy and exhaustive recalibration programme verifying the performance of all our test equipment. After this we had confidence in our experimental test rig, which was confirmed by successful reproduction of manufacturer's data on a different compressor type wi th a much lower oil charge.

We were now left with the problem of resolving the original 25% discrepancy. By al lowing for the high oil concentration we could produce crude theoretical performance data which approximated to what we observed, but we could not reconcile this wi th the manufacturer's figures. Finally we realised that the manufacturer's calibration rig was not a calorimeter, but was in fact a 'gas loop' as is commonly used in the refrigeration industry. In this type of test rig, the refrigerant is always in the gas phase with the condenser and evaporator being simulated by thrott l ing and cooling the refrigerant gas to the required suction conditions. The heat extraction rate is calculated from the measured temperatures, pressures and f low rates using published pure refrigerant property tables, not those of the oil-refrigerant mixtures that are obtained in a practical refrigeration system. Further, since the refrigerant is assumed always to be in the gas phase, any problems which would arise in the evaporator (or indeed in the condenser) because of the oil- refrigerant interaction cannot be observed.

With increasing interest in rotary and screw compressors, in many cases requiring a high oil charge in the system, it is essential that more accurate methods are available for handling experimental and test data. This paper presents thermodynamic property (pressure--enthatpy) charts for a range of R 12-oil mixtures. These charts can be used directly for heat pump cycle analysis using measured state point properties in the normal way.

A n a l y t i c a l a p p r o a c h

The fundamental approach adopted in the present work was to take the solubil i ty equation derived by Bambach ~ for R 12 in a paraffinic oil, and to use it, together wi th the ASHRAE thermodynamic data for

35 r~ r~ ~ o 0 o 0 0 0 0 0 r~ ~ {e ~ c~ 0 o

~o I / / / / / / / / / / , z ~ z " / ,/ ,' JI,'J~'R, ~ i ~o25 l .~ / / / / // ii I

" l I ( /IJ, '" ! ~ / ~ / / / / ,// " / / / / //11/\\

'/ / ///11,1 - ~ ' ~ ' ~ ' ~'o '~o',~o '~ 'o '4o ' . . . . oo ~o ~o

Entnolpy KJ kg i of m i x t u re

Fig. 1 Pressure-enthalpy chart for mixture involv ing 0.1% oil

Fig. 1 Diagramme pression-entha/pie pour un m#/ange contenant O. 1% d'huile

10

£

Fig. 2

? ° °o o ~o ~ o o ~ o oo

/ /

I ~ / / / v / / /

/ ,/

/ /

/ /

/ /

/ /" /

/

210 4~- I 60

/ /

I I I I I / 80 1OO 120 140 16Q Enthulpy kJ kg i of m i x t u re

o o o o o O / , o o I " ~" / / f I"i"~ o~ I / / /_LI ' ! '4\o ~ I

, / / / - w , , / \ X J

/

Pressure-enthalpy chart for mixture involv ing 2% oil

Fig. 2 Diagramme pression-enthalpie pour un m#lange contenant 2% d'huile

R 12, to modify the calculation of the state point properties of an oil-refrigerant mixture. The results are then presented in a graphical form similar to the well known refrigerant pressure-enthalpy charts.

The particular paraffinic oil considered by Bambach is not really typical of currently used refrigeration oils, but Mawhinney 8 has shown that the solubil i ty of R 12 in Shell Clavus 33, a commonly used naphthenic oil, is remarkably similar. Consequently we felt that the use of Bambach's data was justified and could serve to illustrate the new approach.

The computer programs and data used for calculating the state point properties of pure refrigerants are based on the approach of Kartsounes and Erth 7 wi th improved convergence routines 6. The refrigerant data is as published by Downing 3. This program was modified to take account of the o i l - refrigerant solubil i ty data and used to generate the pressure-enthalpy charts shown later.

When a heat pump is charged wi th refrigerant and oil the proportions of oil, l iquid refrigerant and

200 International Journal of Refrigeration

vapour refrigerant in the circulating fluid wil l vary in different parts of the circuit. Thus, at the condenser outlet the vapour refrigerant fraction is zero, whi le at the evaporator outlet, the liquid refrigerant fraction is limited to that amount which is dissolved in the oil. This can be represented by the fol lowing definitions.

w = refrigerant fraction in the liquid mixture

mass of refrigerant mass of l iquid refrigerant + oil

x = o i l fraction in the total mixture mass of oil

total mass of refrigerant + oil

z= t iqu id fraction in total mixture

mass of l iquid refrigerant + .oil

total mass of refrigerant + oil

(1 - z ) = v a p o u r fraction in total mixture

mass of refrigerant vapour total mass of refrigerant + oil

y = refrigerant vapour fraction (quality)

mass of refrigerant vapour total mass of refrigerant

Hence

x 1 - z z - , • y =

w " 1 - x I

Bambach's equations for the solubi l i ty of refrigerant R 12 in the paraffinic oil are

for t < 0°C

log 1 oP= 4.9972 - 0.558 w -°~

1177.67 - 98.753 w -°5

for t > O°C

I o g l o P = A - ( T - 2 7 3 . 1 6 )

[0.002338 ( w - 0 .6 )2 -0 .000075]

where P is the absolute pressure in bar, T is the absolute temperature in K.

From these equations the maximum amount of refrigerant, w, dissolved in the oil can be determined if the temperature and pressure are known. This allows the fraction of the total refrigerant which is dissolved in the oil to be determined from the known oil fraction, x. Hence the total l iquid fraction, z, can be calculated. Under condit ions in which pure refrigerant would be liquid in any case, the liquid fraction of the oil-refrigerant mixture must obviously be unity and the solubi l i ty of the refrigerant in the oil has little significance. For our purposes the fluid can be treated as a mixture rather than as a solution.

In normal practice this condit ion only appears at the outlet from a subcooled condenser.

In the more interesting case where refrigerant vapour is also present the solubi l i ty effect takes on a much greater significance. Under condit ions in which pure refrigerant exists only in the vapour phase, an oil-refrigerant mixture may have a substantial fraction of the refrigerant dissolved in the oil and consequently still in the liquid state. This is the effect which causes the dramatic changes • observed in heat pick up at the evaporator. The enthalpy of the mixture is the sum of four components: the enthalpies of the refrigerant liquid, the refrigerant vapour, and the oil, together wi th the heat of solution of the refrigerant in the oil. In fact the last quantity has been shown 2 to be small enough to be neglected.

The enthalpy of the refrigerant liquid is calculated by extrapolation of the subcooled liquid refrigerant properties at the particular pressure to the required higher temperature. The enthalpy of the oil can be approximated by integrating its specific heat and taking the reference temperature as - 4 0 ° C for consistency with the refrigerant data. The equation used was

ho~= 67.1 2 + 1 . 7 5 4 t + 0.001 9t 2

Hence the total enthalpy of the mixture can be calculated as

h mix = Zh f iq -I - (1 - z ) h rw~

where

hliq= (1 - w )ho i l - I - Whrliq

The specific volume of the mixture was determined in an analogous manner from the sum of the partial volumes of each of the components. The density of a typical paraffinic oil is given by

do ,=932 .47-O.6298 t

The refrigerant densities d,~q and d~,ap are determined from the refrigerant property equations in the usual way. The mixture specific volume, Vmix, is then given by

z 1 - z Vmix = dzz -'1" drva p

where d z, the liquid mixture density, is given by

1 w 1 - w

d--~ = d,,,q + do,---~

Pressure -entha lpy charts

Sample pressure-enthalpy charts are shown in Figs 1 to 3, for oil fractions ranging from 0.1% to 15% The influence of the oil is readily discernible.

Volume 5 Number 4 July 1982 21)1

Probably the most noticeable effect is the disappearance of the saturated vapour line wh ich is characteristic of the pure refrigerant curves. This occurs, of course, because the boi l ing l iquid is no longer pure, so that no unique boil ing point exists. What is not obvious from these charts is the relationship between these curves and their pure refrigerant equivalent. This is shown in Fig. 4, where the curves for an oil fract ion of 8% are superimposed on the pure refrigerant set. The differences are immediately obvious. Boil ing begins in the mixture at a lower pressure and a higher enthalpy; there is no dist inct boi l ing point (the slope of the solid curves is more apparent in this diagram); and a signif icant level of superheat appears at much lower mixture enthalpies. The effect of this on performance is best il lustrated by an example.

I l lustration

An example of the appl icat ion of these charts is superimposed on Fig. 4, wh ich shows the implications of using the same measurements for pressure, temperature and mass f l ow rate to determine system performance using pure R 12 refrigerant curves (dotted and cycle 1 ) and the present oi l -refr igerant curves for an 8% oil mixture (solid and cycle 2). The effect is immediately obvious; evaporator capacity is seriously reduced and compressor work is increased. Using the idealised cycle shown, w i th an evaporat ing temperature of 0°C, 5°C evaporator superheat, discharge temperature of 75°C, condensing temperature of 50°C, and no condenser subcool ing, the heat transferred at the evaporator is reduced from 107 kJ kg -1 to 79 kJ kg -~, and the diagram COP from 3.97 to 2.55. When a motor eff iciency of about 80% is al lowed for, these become 3.18 and 2.07. The latter f igure is very close to the value observed experimental ly under similar condit ions.

It is wor th noting that this example and the pressure-enthalpy charts presented earlier are

4 O

2(?

15

10

i

i

20 40 60 80 100 120 140 160 180 200 220 240 Entholpy kJ kg 1 of m i x t u re

Fig. 3 Pressure-enthalpy chart for mixture involving 15%oil

Fig. 3 Diagramme pression-enthalpie pour un m#lange contenant 15% d'hui le

o I 35

30

25

20

15

k 4L

3 i i

2i- 1

Fig. 4

20 40 60 80 100 120 140 160 180 200 220 240 Enthe~py kJ kg 1 of m i x t u re

Influence of oil on system performance. Comparison of pure refrigerant cycle (broken lines) with that appropriate to a mixture with an 8% oil fraction (solid lines)

Fig. 4 Influence de/ 'hu i le sur la performance du syst#me. Comparaison du cyc/e 2, frigon~7#ne pur (/ignes en pointi l l#) avec le cyc/e correspondant ~ un m#lange contenant 8% d'hui/e (lignes pleines)

referred to total mass f l ow as wou ld be measured by a f l ow meter in the liquid line. If it is preferred to have the data expressed in terms of refrigerant mass f l ow alone, then the horizontal axes for the o i l - refrigerant data must be rescaled by dividing the numbers by (1 - x ) . This wil l have no effect on the performance, but gives the evaporator capacity in kJ kg -1 of refrigerant. In our example above, the evaporator capacity then appears as 85.3 kJ kg -1. but the COP is unaffected as the compressor work is increased by the same ratio.

This work was partly supported by the Commission of the European Communit ies under its First Energy R&D Programme.

References

1 Bambach, G. Das Verhalten von Mineralol ~F12- Gemischen in Kaltemaschinen, C. F Muller, Karlsruhe (1955)

2 Cooper, K. W., Mount, A. G. Oil circulation - its effect on compressor capacity, theory and experiment'. Proceedings of 1972 Purdue Compressor Technology Conference, Purdue University (1972)

3 Downing, R. C. Freon refrigerant equations ASHRAE Trans 78 (1972) 158-169

4 Green, G. H., Furse, F. G. Effect of oil on heat transfer from a horizontal tube to boiling refrigerant 12-oil mixtures ASHRAE Journa/ (October, 1963)

5 Hughes, D. W., McMullan, J. T., Mawhinney, K. A., Morgan, R. Lubricant related problems with heat pumps. Proceedings of 1980 Purdue Compressor Conference, Purdue University (1980) 156-163

6 Hughes, D0 W., McMullan, J. T., Morgan, R. Calculation of the influence of lubricating oil on the performance of refrigeration and heat pump systems/nt J Energy Res 6 (1982) (to be published)

7 Kartsounes, G. T., Erth, R. A. Computer calculations of the thermodynamic properties of refrigerants 12, 22 and 502 ASHRAE Trans 77 (1971 ) 88-103

8 Mawhinney, K. A. On the performance of vapour compression heat pumps D Phi/Thesis New University of Ulster, Coleraine, Northern Ireland (1981)

202 Revue Internationale du Froid