ASHRAE OR-16-C017-2016 Refrigerant R513A as a Replacement for R134a in Chillers.pdf

1、 Kenneth Schultz is a thermal systems development engineer for Ingersoll Rand, La Crosse, WI. Steve Kujak is the Refrigerants Technology leader for Ingersoll Rand, La Crosse, WI. Julie Majurin is a chemistry team leader for Ingersoll Rand, La Crosse, WI. Refrigerant R513A as a Replacement for R134a

2、in Chillers Kenneth Schultz, PHD Steve Kujak Julie Majurin Member ASHRAE Member ASHRAE Member ASHRAE ABSTRACT Regulatory and voluntary actions are beginning to limit the direct global warming potential (GWP) of refrigerants used in many applications. A new class of fluids called unsaturated hydroflu

3、oro-carbons or hydrofluoroolefins (olefin HFCs) has been developed to address this concern. These new fluids are being blended with existing HFCs to obtain lower GWP replacements or substitutes for todays refrigerants. This paper provides an overall assessment of the application of R513A, an azeotro

4、pic blend of R1234yf and R134a (56%wt/44%wt), as an alternative to R134a. R513A provides specific environmental and safety features of interest including: no impact to stratospheric ozone, 56% reduction in GWP compared to R134a, no significant secondary adverse environmental impacts, low toxicity, a

5、nd non-flammability. This paper will review thermodynamic cycle performance and heat transfer characteristics of R513A relative to R134a. Material compatibility and chemical stability have been demonstrated to be favorable and test results will be compared to R134a. INTRODUCTION Societal demands to

6、limit climate change are driving new regulatory policies and voluntary actions to lower the direct GWP impact of refrigerants. These regulations have spurred both the development of alternative lower GWP F-gas chemistries and renewed consideration of so-called “natural” refrigerants, such as water (

7、R718), ammonia (R717), carbon dioxide (R744), and hydrocarbons. R513A is an azeotropic blend of R1234yf and R134a (56%wt/44%wt) with a GWP100of 572, a 56% reduction from R134as 1300 (IPCC, 2013). This makes R513A compatible with existing application standards and building codes and therefore allows

8、its immediate use. R513A is a near-design-compatible alternative to R134a, offering essentially equal capacity to R134a, similar operating pressures, and very comparable material compatibility and stability. ASSESSMENT OF R513A IN CENTRIFUGAL CHILLER PRODUCTS Environmental and Safety The environment

9、al characteristics of R134a are well documented, so this section will focus on the environmental features of R1234yf. A recent article summarizes the atmospheric chemistry and features of a number of short-chain haloolefins (Wallington, et al., 2014), including R1234yf. R1234yf is documented to have

10、 an atmospheric lifetime of 10.5 days, a photochemical ozone creation potential (POCP) of 7.0, a GWP of 1, and an ozone depletion potential (ODP) of zero. These values indicate R1234yf will have limited impact on ground level ozone formation and its direct contribution to GWP will be very low. R1234

11、yf has been extensively studied for formation of noxious degradation products. R1234yf can form trifluoroacetic acid (TFA). The World Meteorological Organization concluded “that global replacement of HFC-134a with HFC-1234yf at todays level of use is not expected to contribute significantly to tropo

12、spheric ozone formation or produce harmful levels of the degradation product TFA” (WMO, 2011). R513A has been recently classified in ASHRAE Standard 34 (ASHRAE, 2015) as “A1” (lower toxicity/non-flammable). R513A has an occupational exposure limit of 650 ppm v/v compared with 1000 ppm v/v for R134a

13、and a refrigerant concentration limit of 72,000 ppm v/v versus 50,000 ppm v/v for R134a. This makes R513A compatible with existing application standards and building codes and therefore allows its immediate use. ASHRAE Standard 15 allows Class A R513A to be used with fewer restrictions in regards to

14、 mandatory mechanical equipment room requirements compared with other next generation refrigerants R1234yf and R1234ze(E) that have an A2L (lower toxicity/mildly flammable) classification, and will allow the refrigerant to be used in direct expansion heat exchangers in an occupied space if warranted

15、. Thermodynamic Properties The operating pressures of R1234yf are very similar to R134a, hence R1234yfs use as a low GWP alternative to R134a in automobile air-conditioning. Blends of R1234yf and R134a therefore also have very similar operating pressures; see Figure 1. The azeotropic nature of R513A

16、 results in slightly higher pressures than R134a. Glide (the difference between dew and bubble point temperatures) is less than 0.04 F (0.02 C) down to 32 F (0 C), then increases to 0.32 F (0.18 C) at 40 F (40 C) The slope of the R513A saturation curve, nullnullnullnullnullnullnull , is slightly sha

17、llower than for R134a, meaning slightly lower compression ratios for a given temperature lift. Figure 2 and Figure 4 show the pressure-enthalpy and temperature-entropy charts, respectively, for a single-stage vapor compression cycle operating with an evaporator saturation temperature of 40 F (4.4 C)

18、 and no leaving superheat, a condenser saturation temperature of 115 F (46.1 C) with 15 F (8.3 C) of exit subcooling, and a compressor isentropic efficiency of 0.7. Note that the width of the R1234yf dome, representing the latent heat of vaporization, is narrower than for R134a; R513A is midway betw

19、een the two. The performance of several refrigerants operating at the above conditions assuming the compressor has a fixed volumetric displacement is listed in Figure 3. For these conditions, R513As narrower dome (smaller heat of vaporization) but higher vapor density combines to deliver cooling cap

20、acity essentially equal to that with R134a with efficiency about 2% lower than R134a. This efficiency shortfall can be made up by raising the evaporator saturation temperature and lowering the condenser saturation temperature by a combined 1.3 F (0.7 C) (e.g., by 0.6 F (0.35 C) each) through an incr

21、ease in heat transfer surface area. Such an increase in evaporator saturation temperature enhances capacity by about 2%. Using R1234yf by itself results in losing over 6% in capacity and over 3.5% in efficiency. The thermodynamic properties of R1234ze(E) offer the potential for nearly matching the e

22、fficiency of R134a. However, capacity with R1234ze(E) is 26% less than with R134a, requiring a compressor with 34% larger displacement to match capacity. R1234ze(E)s lower vapor density may also necessitate use of larger diameter heat exchanger shells and interconnecting piping. The properties of R5

23、13A are similar enough to R134a that R513A can be considered for using in existing chiller designs with little need for modification. Figure 3 and Figure 4 highlight that the compressor discharge temperatures for all of the lower GWP alternatives are significantly lower than for R134a. This is a con

24、sequence of the steeper saturated vapor curve on the right side of the Ts dome. The lower compressor discharge temperatures need to be considered in the quality of (amount of refrigerant in) the lubricating oil returned from the compressor discharge back to the compressor. Chiller Test Results Perfo

25、rmance testing of R513A in a water-cooled screw chiller and two air-cooled screw chillers has been run in our laboratory. The results of those tests are summarized here. Figure 1. Saturation pressure versus temperature curves. The symbols indicate critical points. Figure 2. Pressure-enthalpy chart f

26、or R134a, R513A, and R1234yf. Figure 3. Performance of a single-stage vapor compression cycle. Top chart shows differences in capacity and efficiency relative to R134a. Bottom chart shows compressor discharge superheat. Figure 4. Temperature-entropy chart for R134a, R513A, and R1234yf. Water-Cooled

27、Chiller The equipment tested was a development prototype of a water-cooled water chiller product with a nominal capacity of 230 tons (800 kW). This product comprises twin screw compressors with two separate refrigerant circuits and shell-and-tube heat exchangers; the evaporator utilizes a falling fi

28、lm design. Only one circuit was run to minimize the amount of refrigerant needed and to more clearly compute shell heat transfer performance. Tests were run with R134a as a baseline, R513A; blends of R134a, R1234yf, and/or R1234ze(E) labeled N-13a and N-13b; a blend of R1234yf, R134a, and R152a labe

29、led ARM-42, and R1234ze(E). R1234yf was not tested because a sufficient quantity was unavailable at the time of testing. Only the R134a baseline and R513A results are reported here. See Schultz and Kujak (2013) and Schultz (2014) for further details on the other refrigerants tested in this chiller.

30、Initial testing indicated that performance was optimized with the same refrigerant charge for both R134a and R513A. The COP and capacity obtained with each refrigerant from full load to minimum load at the test conditions listed above are shown in Figure 5. With both refrigerants, the COP first rose

31、 as the chiller was unloaded to 80%. This is due to the heat exchangers becoming effectively larger, resulting in smaller approach temperatures and lower compressor lift. A decrease in compressor efficiency as capacity drops below 80% of full load results in a decrease in overall performance. As sho

32、wn in Figure 6, the measured maximum capacity with R513A was essentially the same as with R134a. The simple thermodynamic cycle model described above predicted about 1% higher capacity for R513A. The measured efficiency with R513A was about 4.5% lower than with R134a; about a 2.5% decrease was predi

33、cted by the simple thermodynamic model. The predicted performance is based on the thermodynamic properties of the refrigerant with the assumption that the evaporator and condenser saturation temperatures are the same for all refrigerants, along with the same isentropic efficiency of the compressor.

34、The lower measured performance observed here was attributed to lower refrigerant-side heat transfer coefficients in the evaporator (by about 15%) and condenser (by about 15%) with R513A compared with R134a. Enhanced tube surfaces might need to be modified for maximal performance with R513A. Sets of

35、tests were also run at full capacity over a range of entering cooling water temperatures. Figure 7 shows the maximum capacity achieved (normalized to the capacity at 85 F (29 C) as a function of entering cooling water temperature. The COP at maximum capacity as a function of entering cooling water t

36、emperature is shown in Figure 8. Capacity and COP are both slightly more sensitive to entering cooling water temperature (as surrogate for condensing temperature) with R513A than with R134a. However, the difference is small and does not hinder application of R513A over the range of conditions curren

37、tly covered by R134a today. Figure 5. COP versus capacity obtained from water-cooled chiller. Figure 6. Capacity and COP relative to R134a at water-cooled chiller full load point. Figure 7. Change in capacity (normalized to cap-acity at 85 F (29 C) as entering cooling water varies. Figure 8. Change

38、in COP (normalized to COP at 85 F (29 C) as entering cooling water varies. Air-Cooled Chiller The performance of a premium efficiency air-cooled chiller with nominal design capacity of 160 tons (560 kW) was measured with both R134a and R513A. This product comprises a variable speed screw compressor,

39、 a flooded-style shell-and-tube evaporator, and an aluminum tube-and-fin condenser with variable speed fans. Sophisticated controls are applied that maximize efficiency by optimizing compressor and condenser fan power. Again, testing showed that the chiller required the same refrigerant charge with

40、R134a and R513A. The nominal design conditions are 44 F (6.7 C) leaving chilled water temperature and 95 F (35 C) ambient air temperature. Figure 9 shows full load capacities with R513A were essentially the same as with R134a, in agreement with the thermodynamic cycle model. As shown in Figure 10, t

41、he measured COP at full load with R513A was down by about 4.5% compared with R134a. This is 1.5% greater than predicted by the thermodynamic cycle model, again, due to heat exchanger effects. For this chiller, the refrigerant-side heat transfer coefficients in the evaporator with R513A were lower th

42、an with R134a by about 5% to 10%, resulting in decreases in overall heat transfer coefficients of about half that magnitude and making a very minor contribution to the degradation in efficiency. The 3% decrement in COP inherent in R513As thermodynamic properties can be offset by increasing evaporato

43、r heat transfer area. Figure 9. Full load capacity of the RTAE chiller at standard design conditions with R134a and R513A. Figure 10. Full load efficiency, relative to R134a, of the RTAE chiller at standard design conditions. The major contributor to the 1.5% incremental reduction in COP beyond the

44、effect of thermodynamic properties comes from the condenser. R513As narrower dome (latent heat of vaporization) compared to R134a results in a higher circulating refrigerant mass flow rate by about 15% for a given capacity. This higher mass flow rate produces a larger pressure drop across the conden

45、ser. The coupling between heat transfer requirements setting the condenser inlet and outlet saturation temperatures (and therefore pressures) and the pressure difference required to drive the refrigerant flow through the condenser result in an incrementally higher condenser inlet (compressor dischar

46、ge) pressure for R513A. In this case, the incremental increase in pressure was equivalent to an increase in condenser inlet saturation temperature of about 1.1 F (0.6 C). This effect can be negated by redesigning the coil to have shorter flow paths and increasing the number of circuits. Figure 11 sh

47、ows that the pressure drop across the condenser depends primarily on the refrigerant mass flow rate, independent of whether the refrigerant is R513A or R134a. Because the air-side resistance typically dominates the overall heat transfer performance of a refrigerant-to-air heat exchanger, differences

48、 between the refrigerant-side heat transfer coefficient with R134a and R513A are expected to make negligible impact on performance. A consequence of R513As inherently lower compressor discharge temperatures is that the amount of refrigerant dissolved in the oil collected by the oil separator at the

49、exit of the compressor is increased; see Figure 12. It has been determined that the viscosity of the oil/refrigerant mixture supplied to the compressor bearings is still sufficient to ensure reliable long-term operation of the chiller. As with the water-cooled chiller, compressor efficiency was observed to be essentially identical between R134a and R513A. A second air-cooled chiller of nominal 50 tons (175 kW) per circuit was tested with R513A. This chiller model has microchannel condensers. At full load under nominal design conditions of