Phase Equilibrium in Oil-Refrigerant Solution R 134a / SW22

Transcription

1 Purdue University Purdue e-pubs International Refrigeration and Air Conditioning Conference School of Mechanical Engineering 1996 Phase Equilibrium in Oil-Refrigerant Solution R 134a / SW22 V. V. Zhidkov NORD Association V. P. Zhelezny Odessa State Academy of Refrigeration P. V. Zhelezny Odessa State Academy of Refrigeration Follow this and additional works at: Zhidkov, V. V.; Zhelezny, V. P.; and Zhelezny, P. V., "Phase Equilibrium in Oil-Refrigerant Solution R 134a / SW22" (1996). International Refrigeration and Air Conditioning Conference. Paper This document has been made available through Purdue e-pubs, a service of the Purdue University Libraries. Please contact epubs@purdue.edu for additional information. Complete proceedings may be acquired in print and on CD-ROM directly from the Ray W. Herrick Laboratories at Herrick/Events/orderlit.html

2 PHASE EQUILIBRIUM IN OIL-REFRIGERANT SOLUTION Rl34ajSW22 V.V. Zhidkov 1, V.P. Zhelezny 2, P.V. Zheleznl 1 "NORD" Association Zhucovsky Ave., 2, Donetsk, , Ukraine phone (0622) , fax (0622) Thermophysical Engineering Department Abstract The Odessa State Academy of Refrigeration Petra Velikogo St.l/3, Odessa, , Ukraine phone (0482) , fax (0482) In this work the results of experimental study of phase equilibria in the refrigerant R134a-oil SW22 mixture were discussed. It was shown that this mixture has limited solubility with upper critical point. Method of approximation of the exfoliation line was proposed. Results of studies of liquid-vapor phase equilibria in this mixture were analyzed. General features of saturated vapor pressure temperature and concentration dependence of the refrigerant-oil solution were analyzed too. Method of description of the experimental data was proposed which did not need a big volume of the initial information. System of the equations was proposed for describing the experimental data. Nomenclature Ps Pressure, 10 5 Pa T Temperature, K x Concentration of 134a (wt %) p Density, kg/m 3 --c ln(tjt)- reduced temperature Subscript c s nb " Introduction Critical point Saturated point Normal boiling Boiling point Dewp::>int At present time the refrigerant Rl34a is widely used in domestic refrigerators. Its thermodynamic properties were well known. By this time some firms produce refrigerating oil for domestic refrigerators, freezers and air-conditioners. But Rl34a has big value of the Global Warming Potential GWP=l300 and lower, as to compare with Rl2, energetical effectiveness of cooling equipment. Marking this shortcomings it must be said that refrigerator's working media is not pure RI34a but its mixture with refrigerating oil. So an optimal choice of the oil determines the energetical effectiveness of refrigerator to a considerable extent. In spite of that there are few literature data on phase equilibria in real working media. List of refrigerating oils produced by 447

3 various firms is sufficiently wide. But almost complete absence of information on thermodynamical properties of oil-refrigerant solutions prevents from scientifically grounded at choice of the optimal sort of refrigerating oil In order to calculate thermodynamical properties of an oil-refrigerant solution one needs proper information on the phase equilibria in it At present time one can obtain this information by experimental method only. Results of Research In this way the exfoliation line of the oil-refrigerant solution Rl34a/SW22 was studied firstly. Measurements were made with the help of glass measuring unit by the visual checking on the appearance of the interface boundary in the solution. Temperature of exfoliation was determined during the lowering the temperature of the thermostat Mixing of the solution was made to obtain the reliable results. It was discovered that the exfoliation line had an upper critical point with parameters: Tc=237.25K; xc=0.82wt%. They were determined with the straight diameter rule of Matias. There are no theoretically grounded relations describing exfoliation lines of such complicated systems as an oil-refrigerant solution. So experimental data, as a rule, are not fitted analytically. Due to sufficient steepness of the left boundary curve any extrapolation with the help of polynomial relations to low temperature region results in big errors. But it is this interval of parameters that covers the region of the most interest for designers of refrigerators. On the other hand, it is very difficult to do measurements here. In our work an attempt to theoretically grounded approximation of data on exfoliation line with upper critical point was made. In the frames of the proposed method there are restrictions: first, thermodynamic system consists of non-associated substances; second, refrigerant and oil does not chemically interact; third, degree of polymerization of the refrigeration oil components is stable during the refrigerant solution in it. Only the region of the exfoliation line parameters for which theoretically grounded universal relations between temperature and concentration exist is the one of the small vicinity of the critical point. Universal character of critical fluctuation, and equivalence of simple polynomial models describing the phase transitions of various nature lead to conclusion that the critical phenomena are isomorphic [ 1]. When the variables are choiced correctly (isomorphic ones) then thermodynamical potentials of various systems have the same fundamental correlation with temperature and the order parameter in the frames of the second order phase transition. Difference between concentrations x-xc on the left and right branches of the exfoliation line may be used for it as an order parameter. In this case the temperature dependence of concentration in the vicinity of the critical point may be written as it follows or ln(x/xc) = ~0 1:~ (2) where xc and Tc- concentration and temperature in the critical point of exfoliation; Xo and 1; 0 - critical amplitudes; {3- critical index. (1) 448

4 Because of universality of critical phenomena the value of ~ for the exfoliation lines equal to that for density on the saturation line, [3=0,325. Equations (1), (2) are valid only in small vicinity of the critical point. Their value is in theoretically grounded using of the scaling principles for prognosing thermophysical properties of substances on their saturation lines [2,3], according to which a density on boiling and condensation lines is described by equations having the following shape ln(p' / Pc) = B1 -r 13 F 1 (-r) (3) where p' and p"- density of liquid and vapor on the saturation line; Bb Br constants which have the meaning of critical amplitudes; F 1 (t) and F 2 (t)- universal (for non-associated substances) cross-over functions, which can be calculated with the help of the equations [3,4] F 1 (-r) = r 0.4/ln -r (5) F2(-r) = r 0 " 6 /ln T (6) Taking into account isomorphism of critical phenomena, for the exfoliation line one may write the following equations: (4) ln(x' /xc) = ~ 0 ' -r13 F 1 (-r) (7) ln(xjx") = ~o" 1"/3F2(c) The amplitudes for the mixture RI34a-SW22 obtained by experimental data approximation have the following values: ~ 0 '= \ ~ 0 "= at Tc= K and [3=0,325. Carried-out studied point on that the equations (7) and (8) correctly describe experimental data within +0,5% for concentration and+ IK for temperature. This level of discrepancies corresponds to methodical and experimental errors of determination of these values. Calculation of thermodynamic properties of oil-refrigerant solution requires correct data on phase equilibria "liquid-vapor". At present time this data can be obtained only by experimental studies. Study of phase equilibria in oil-refrigerant solution Rl34a/SW22 has been done by optical method. The experimental device has been described in detail in [5]. The metal cell with membrane zero-indicator and hot valve was placed in liquid thermostat. That's why there no ballast volumes of vapor phase in measuring cell. The magnetic mixer was used in the cell to shorten the time of thermodynamic equilibrium's establishment. Temperature was measured by platinum resistance thermometer with the error +O,OlK and pressure- by cargo plunger manometer MTI-6 and MIT-60 with accuracy 0,05. Mixture composition has set by the weight method. The experiment was carried out with decreasing and increasing temperature of the thennostat. Characteristic feature of the phase equilibria studies of oil-refrigerant solution is a big time of thermodynamic equilibrium's establishment (up to 3-4 hours). Measurements were made in intervals 275~::::345K and 0.25:-s;P:Q.2 MPa of temperatures and pressures at five values of concentration of Rl34a: xl = ; X 2 =0,483; X 3 =0,256; ~=0, 1786; X 5 =0, (8) 449

5 Analysis of experimental data shows that far from critical point of solution the lines x=const can be approximated by polynomial in the lnp-f(l!f) coordinates. Approximation of the experimental data was made on the base of the equation ln (Pc/P) = ar 't + b i 64 (10) which has been proposed for description of saturated vapor pressure by one of the authors [3,4]. In the Eq.(ll) Pc - critical pressure; and b -constant which were obtained from the experimental data. Analysis of Eq.(lO) shows that cxr is Riedel's similarity criterion. Two-constant equation (11) has wide extrapolation possibilities both in the region of low and high temperatures. So for the fitting of experimental data a system of equations was used: ln Ps =In Pc(x) - ar(x) - b(x) 't2.64 (11) ln Pc(x) = x (x 0 ) (x 0 ) 3 (12) 2 3 ar(x) = x x x 0 (13) (9) b(x) = x x x 0 3 (14) where x 0 -oil concentration: x 0 = 1 -x. The values of critical temperatures of the oil-refrigerant mixtures with various concentrations needed for calculation were obtained by the modificated Guldberg's rule where constant C was determined for the refrigerant R134a. TcTnb-1 = C (15) The values of critical temperatures of the oil-refrigerant mixtures were obtained by the experimental method. The data obtained in this manner were fitted by the equation. Numerical values of coefficients in approximation model (11)-(14) are equal to: pl Mean-square deviation of calculated values from experimental ones does not exceed at 0.1:::; x:::; 1 (16) where Ao=5.5093; A 1 = ; A 2 = " 2 ; A 3 = " 2 ; ~= Mean-square deviations of the calculated (Eq.(ll)) and experimental results lie within+ 1.5%. Conclusion This work contains the results of experimental studies of the exfoliation line and phase equilibria in the mixture R134ajSW. New method of experimental data treatment is proposed. Equations obtained permit to calculate vapor pressures of the oil-refrigerant mixture in the 450

6 REFRIGERATION CYCLE SIMULATION PACKAGE: REFSIM The mathematical models for the main components of the vapour compression based refrigeration circuitry, namely compressor, condenser, capillary tube-suction ljne intercooler. assembly and evaporator were developed from the fundamental conservation principles. Special test facilities were designed and fabricated for the model validation studies of the individual components. The component models were subsequently incorporated into a robust and efficient computer algorithm for the performance simulation of entire refrigerator unit under the steady-state regime. The steady-state simulation model were then extended to include on/off working conditions by adapting the cycle-average strategy. The final software product were verified using the experimental measurements taken on specially intstrumented refrigerator units. The component models are described briefly in the following subsections. Compressor Model: The biquadratic curve fits as a function of condensing and evaporating temperatures are obtained from the compressor map data and used for both compressor mass flow rate and power predictions. The compressor discharge temperature under different operating conditions is determined from the first-law energy balance which relates input power, heat rejection from the canister and enthalpy change of refrigerant across the compressor. The rate of heat loss from the compressor canister is calculated in terms of the convection heat transfer equation. This equation is based on the refrigerant discharge temperature for simplicity. The overall conductance value of the compressor is estimated from the test data measured at the standard rating point. Condenser Model: A simple model based on the measured values of air side heat transfer coefficient is established for the tube-and-wire type condenser unit. Experiments were conducted for a number of relevant geometric configurations and wide range of operating temperatures by circulating hot water inside the condenser tubes. The measured values of natural heat transfer coefficients are correlated in terms of temperature difference between the condenser and the ambient air for each unit tested. The relative position and the inclination of the condenser unit were changed systematically to study effects of these parameters and determine the optimum configuration. The thermal efficiency of condenser wires were calculated from the standard expressions of fins given in common literature. The tube side heat transfer coefficient for the condensing refrigerant is obtained from the correlations given in [1]. Capillary Tube-Suction Line Heat Exchanger Assembly Model: The REFSIM package includes a comprehensive and a rigorous numerical model for the performance prediction of the capillary tube expansion device. The present numerical model was intended to account for the effects of suction line cooling on the performance characteristics of capillary tubes in order to capture the true system performance [2]. The governing equations for steady one-dimensional homogeneous equilibrium two-phase flow through the capillary tube are obtained by applying basic conservation principles to differential control volumes. The subsequent numerical solution is facilitated by transforming these equations into a more convenient form in terms of numerical integration. The transformed equations along with the energy equation of for steady one-dimensional gas flow through the suction line forms an initial value problem. Forth order Runge-Kutta algorithm is used to march advance the solution along the capillary tube length until the critical conditions is reached. Mass flow rate is continuously adjusted until the calculated tube length equals to the prescribed tube length. Additional iteration loop is required on suction line exit temperature to match the calculated suction inlet temperature to the prescribed conditions at the evaporator outlet. Evaporator Model: The evaporating refrigerant flowing inside the tubes exchanges heat with the cabinet air. The mathematical model for the tube-on-sheet type cold wall evaporator is formed in a way similar to the condenser unit. A boundary layer type flow exists in the vicinity of cold evaporator surface inside the cabinet. The air side heat transfer coefficient in the boundary layer were determined from the numerical experiments conducted on full cabinet geometry. The details of the CFD studies performed in the present work is discussed in the following section. The calculated values also includes the effects of surface-to-surface radiation heat exchange which takes place inside the cabinet enclosure. The refrigerant side heat transfer coefficients were obtained from the correlations given in [3]. The thermal resistances due to heat conduction and the thermal contacts of the tube-on-sheet assembly were estimated by using the experimental data obtained from the existing refrigerator unit. 460

7 PERFORMANCE OPTIMIZATION OF A LARDER TYPE REFRIGERATOR UNIT USING COMPUTER AIDED ANALYSIS TOOLS Engin Dirik, Harun iz, Cezmi Aydm ARCELiK A.$,. Research and Development Center, Tuzla 81719, istanbul, TURKEY ABSTRACT This paper describes the utilisation of computer aided analysis tools in the design and performance optimisation process of domestic refrigerators. A larder type table-top refrigerator unit is considered in the present study. In the first phase of the project, the existing refrigerator using CFC-12 as a working fluid. was converted into a prototype with environmental benign refrigerant isobutane, R600a. The conversion process involved compressor stroke volume and refrigerant charge adjustments to achieve the same cooling performance. in the second phase, in-house developed refrigeration cycle simulation code and commercial Computational Fluid Dynamics (CFD) packages were used for the optimisation with respect to energy consumption. Design parameters of the main components of the refrigeration circuit, namely compressor, condenser. evaporator and capillary tube~ suction line intercooler assembly were modified based on the results of computer simulations. Three prototypes were subsequently built and tested under standard climatic conditions. Redesigned unit has been found to have significantly lower energy consumption compared with the existing refrigerator. INTRODUCTION The ongoing phaseout process of ozone depleting clorofluorocarbons (CFCs) and the increasing emphasis on global warming considerations which demand the production of more energy efficient appliances have been forcing domestic refrigerator manufacturers for the design changes and improvements on their existing product lines. To help to facilitate this need, a flexible component based computer simulation package for the performance prediction of the household refrigerator/freezer units has been developed. Extensive use of this inhouse developed analysis capacity is being made in the product development environment. Commercial CFD packages are also being used routinely to support the refrigerator design process. In this study, utilisation of the established capacity is described for the redesign case of a larder type refrigerator unit. The redesign objective was twofold: (i) conversion to an environmentally friendly hydrocarbon refrigerant, R600a from CFC-12 and (ii) improvement in energy consumption. Since the beginning of 1990's, appliance manufacturers in Europe have started using hydrocarbons refrigerants in domestic refrigerators. Hydrocarbons are environmentally safe substances since they have no ozone depletion and negligible global warming potential. Among their advantages, superior transport properties (high thermal conductivity and lower viscosity), requirement for lower amount of charge compared with CFC and HFC type refrigerants, solubility in traditionally used lubricant mineral oil can be stated. These advantages are so attractive for refrigeration system designers and the people who care for nature that the efforts spent for the design of a risk-free system are easily justified. Although it requires some design changes in the cooling circuit components, isobutane became the choice of domestic refrigeration producers among other hydrocarbon refrigerants. The reason for its selection is that it allows the compressor to operate under less loads thus causing more reliable and a quieter system. In the first part of the paper, the establish simulation capabilities are reviewed. The refrigerant replacement and the energy optimisation methodology followed in the redesign process are presented in the remainder of the paper after a brief description of the existing refrigerator unit. 459

8 Conclusions Our investigation allows to infer that the mixture forms positive azeotrope and that azeotropic concentration changes its value from x=0.39 at 180 K to x=0.33 at 285 K. References 1. Zhelezny, Yu.V. Semenyuk and Yu.A. Chemyak, Proc. 13th European Conf. on Thermoph. Prop., (Lisbon, Portugal, 1993), pp Vladimirov, Yu.F. Shvets, Pressure of the saturation vapor ofr218,r329 and azeotropic mixture Rl16/R23. Thermoph. Prop. of Substances and Materials, 28: 28 (1989). (in Russian). Experimental apparatus A, optical cell; B, magnetic stirrer; C, main heater; D, stirrer; E, subheater; F,G, platinum Fig. resistance termometres; H, windows; I, thermostat; J, vacuum pamp; K, thyristor amplifier; L, PID controller; M, thermometer bridge; 0, potentiometer; P, dead-weigh pressure gauge; Q, nitrogen gas; R, separators; S, Hg-cylinder; T, precise volume controller; U, vacuum gauge; V 1 - V 12, valves. 458

9 Table 2 Thermodynamic properties on the saturation line for composition X= 0.36 T P' P" p' p" h' h" s' s" K MPa MPa kgm- 3 kgm- 3 kjkg-1 kjkg-1 kj(kg Kfl kj(kgkr ) Table 3 Parameters used in Eqs. (7) i ~ pi Pi

10 Table I (continued) Coefficient R23 R116 Bo Co cl Cz l l Do Dt o- 1 Et E:z E F F Tc, K Pc, MPa kg -3 Pc m The thermal properties of the mixture on the vapor-liquid coexistence curve were calculated by the following equations: ln(pjps) = Erln(TJI) + E2(ln(TJT))E 3 (5) ln(p' /Pc) = F 1 (ln(tjt))f 1 r, (6) were f=l-l.ll3(ln(tcft)) 0.4(ln(ln(TJT))r 1 ; EI> E2, E 3, and Fr. Fr coefficients dependent on individual properties of substances. The coefficients and critical parameters used for R23 and Rll6 are listed in Table 1. Thermodynamic properties on the saturation line for composition of 36 mol.% Rl16 were calculated by using Eq.(4) and listed in Table 2. The results for the critical parameters of the mixture can be represented by the following equations: Tc = L ti xi; PC= L pi xi; Pc = L Pi xi. (7) The coefficients of Eqs. (7) are listed in Table

11 The numerical constants in Eq.(2) are tabulated in Table 1 for pure components of the mixture. The mixing rules ofthe coefficients ofeq.(l) were given in the following equations: (3) 0.6%. The binary interaction parameters Kij for the present system were determined to be: K/ = (X2 ~ 0.36),[X2::;; 0.36]; Kii a= (X2 ~ 0.36), [X2. > 0.36]; Kit = 0.4; K/ = 0.4. Deviation ofthe experimental data from those calculated by means ofeq.(l) did not exceed Data [2] (Ps of composition X= 36 mol. % R116 in temperature range from 187 to 280 K) and our experimental VLE data of the mixture were used to determine parameters of the cubic equation of state (the equation used to calculate vapor~ liquid equilibria): P = RT[l/(V ~b)- ajv(v + c)], (4) were a, b, c are coefficients determined from the conditions of vapor - liquid equilibria of the components of the mixture for each value of the temperature on the saturation line. be: Using the same mixing rules (Eq.(3)), the binary interaction parameters Ktj were determined to Kit= X2; Kit= 0. The binary interaction parameters for the calculation of the saturated-liquid densities of mixture were determined to be: Kit=O.l, [Xi~ 0.44]; Kit=O.l +0.8(Xr0.44),[X2>0.44]; Kijc=O. Deviation of the measured bubble-point and dew-point pressures from values calculated using Eq.(2) did not exceed 1%. The maximum deviation between the measured saturated -liquid and saturated-vapor densities of mixture and those calculated from Eq.(2) was 0.4% and 1% respectively. Critical parameters and constants in Eqs.(2),(5),(6) Coefficient R23 R116 Ao A A A " Table 1 455

12 another 1 OOohm platinum resistance thermometer (F) was mounted into the thermostated bath. The temperature fluctuation was detected by thermometer (F) and temperature control was carried out by the 1 kw main heater (C), 40 W subheater (E), thermometer bridge (M) (Model M0-62), PID controller (L) (Model VRT-2) and thyristor amplifier (K). The pressure inside the balloon (Q) had been transmitted to mercury and to oil through a visual separators (R). The dead-weight pressure gauge (P) (Model MP-60) was directly used to measure the pressure. The level of mercury and vapor-liquid interface was determined by visual observation. The sample density was calculated from the inner volume of the cell and the mass of the confmed mixture. The composition of the mixture was detennined by weighting the mass of each pure components before mixing. PVTx properties were measured on the isotherms. Determination of the saturation state of a mixture with a prescribed composition was performed by observing the appearance of a bubble in the liquid-phase sample confmed in the cell The parameters on the dew curve were determined in the points, where isotherms of gaseous phase crossed the isotherms of double-phase region. For analysis of the results of measurements near the dew curve a method of investigation the PVT properties in the region of blurred phase transition considered in [ 1] was used. The critical point was determined by observing disappearance and reappearance of the meniscus. The purities of each component supplied was 99.79% R23 and 99.80% R116. The experimental uncertainty in determining the concentration was estimated as mol/mol, temperature - at 15 mk. Taking into account the systematic and accidental errors, full error in pressure value was equal to % in all pressure region, the one ofliquid density was about %, for vapor density %. The uncertainties in the critical temperature, critical pressure and critical density have been estimated to be within 20 mk, 8 kpa and 6 kgjm 3, respectively. Results The measurements were performed at compositions of x: ; ; ; ; ; ; and mol/mol Rll6 in the following ranges: for temperature from to K, for pressure from 1.0 to 3.4 MPa and for density from 50 to 250 kgjm 3 in the vapor phase. In addition, the bubble-point pressures and saturated-liquid densities were measured at these compositions in the temperature range from to K. The shape of the critical curves was studied at all compositions sudied. Discussion We developed an equation of state in the vapor region based on the present measurements using for it a functional form derived from the cubic equation of state: P = RT[ 1/(V-b)- af(v(v+c)+d)]. (1) Temperature dependences of the coefficients ofeq.(l) are following: a= Ao(l + AI (1/ 't - 1) + A2(1/ r - 1) + A3(1/ t 4-1)); c = C 0 + C 1 (1 - -c)+ C 2 (1 - -c) 2 ; d = D 0 + D 1 (1 - t). b = const; (2) 454