Oil-free centrifugal refrigeration compressors: from HFC134a to HFO1234ze(E) Joost J. Brasz Danfoss Turbocor Compressors Inc. Syracuse University CASE Incubation Center 2-212 Center for Science and Technology Syracuse, New York 13244-4100 USA jbrasz@turbocor.com

ABSTRACT

The high global warming potential of HFC134a (GWP=1300) has led to the

development of a new family of man-made refrigerants with much lower global

warming potential. HFO1234ze(E) is one of those fluids. It has a GWP value of 7

and has been in commercial production as a blowing agent for a few years. Based

on cost and availability this new fluid has been selected a potential candidate to

replace HFC134a in commercial chillers.

A new family of oil-free direct-drive centrifugal compressors with HFO1234ze(E) as

working fluid has recently been introduced commercially, covering a cooling

capacity range from 200 – 300 kWthermal. These new compressors are a spin-off of

an existing platform of oil-free HFC134a products. Being oil-free eliminated oil-

refrigerant compatibility issues, which were a major stumbling block during the

transition from CFC12/HCFC22 towards HFC134a in the early 1990’s.

Due to its somewhat lower pressure and vapor density HFO1234ze(E) requires in a

slightly larger fluid module to achieve equal capacity as HFC134a. Impeller tip

speed is reduced at equal temperature lift (=difference between condenser and

evaporator saturation temperatures) as a result of the lower sonic velocity of

HFO1234ze(E). Overall compressor efficiency improves as a consequence of these

two effects.

1. INTRODUCTION

During the nineties, the air-conditioning and refrigeration industry saw a transition

from CFC’s towards HCFC’s and HFC’s. The main driver for this transition was the

discovery of stratospheric ozone layer depletion by the chlorine atoms found in

CFC’s and HCFC’s.

After the resolution of the ozone layer depletion problem by the introduction of

chlorine-free HFC refrigerants, the environmental concern started to focus on the

global warming impact of these man-made refrigerants. Legislation intended to

Tcond,sat = 35.6 0C

Tevap,sat = 5.6 0C

29.4 0C 34.7 0C

12.2 0C 6.7 0C

Refrigerant ODP GWP

CFC12 1 8500

HCFC22 0.05 1700

HFC134a 0 1300

HFO1234yf 0 4

HFO1234ze 0 6

limit the future use of refrigerants in systems with a high direct effect on global

warming has led to the development of new systems using natural refrigerants such

as CO2 as well as the development of new man-made refrigerants with much lower

global warming potential than the currently used HFC refrigerants.

Table 1 shows a number of medium pressure refrigerants used in water-cooled

chillers. The refrigerants have transitioned from fluids with high ozone depletion

potential (ODP) such as CFC12 and HCFC22 towards fluids with zero ODP but still

a large global warming potential (GWP) such as HFC134a. Today there are

candidate refrigerants with both zero ODP and very low GWP. HFO1234yf is the

result of a joint effort by DuPont and Honeywell to develop a drop-in replacement

fluid for HFC134a for mobile air conditioning (MAC) applications [1].

HFO1234ze(E) is a low GWP fluid developed by Honeywell and currently used as a

foam blowing agent [2].

Table 1. Ozone layer depletion (ODP) and global warming potential (GWP) of

various fluids.

2. CENTRIFUGAL CHILLERS AND THEIR OPERATING CONDITIONS

In the refrigeration industry centrifugal compressors are predominantly used in

water-cooled chillers although smaller capacity direct-drive centrifugal compressors

are now also applied to air-cooled chillers. Figure 1 shows a simplified equipment

diagram of a water-cooled chiller. A representative full-load operating condition for

a commercial water-cooled chiller is to reduce the temperature of the water returning

Figure 1. Simplified equipment diagram of a water-cooled chiller

from the air-handling equipment. Typical full-load entering and leaving evaporator

water temperatures as specified by the ARI rating code [3] are 2.220C (54

0F) and

6..670C (44

0F), respectively The heat absorbed by the evaporating refrigerant is

rejected in the condenser together with the heart of compression. This heat increases

the temperature of the water returning from the cooling tower and entering the

condenser at 29.440C (85

0F) to a leaving water temperature of around 34.72

0C

(94.50F) – the exact number depending on the [deal cycle efficiency of the

refrigerant as well as the compressor efficiency. The evaporator saturation

temperature will be slightly below the leaving chilled water temperature of 6.670C

(440F), say 5.56

0C (42

0F) and, similarly the saturation temperature of the refrigerant

in the condenser will be just above the leaving condenser water temperature of

34.720C (94.5

0F), say 35.56

0C (96

0F).

Overall chiller efficiency is defined using of a coefficient of performance (COP)

number which is defined as follows:

��� � ��������� ��������������

������������ ���!��������"��#"�

For refrigeration systems this number is always larger than , which is the reason that

the word efficiency has been replaced with the term coefficient of performance. In

the US is is customary to specify chiller performance in terms of a kW/ton number

defined as:

��/�� � ������������ ���!��������"��#"�

��������� ��������%��%���������

Since a ton of refrigeration (defined as the cooling capacity obtained from the

melting of a so-called short ton (=2000 pounds) of ice in 24 hours) is equivalent to a

cooling capacity of 3.52 kW, the relationship between these two quantities is as

follows:

��� � 3.52

����

Actual state of the art COP numbers for water-cooled centrifugal chillers at ARI

full-load conditions vary 6.18 (0.57 kW/ton) to 6.77 (0.52 kW/ton). This variation is

caused by the choice of refrigerant, compressor efficiency and refrigeration cycle

details such as sub-cooling and economizing. Advertised centrifugal chiller

performance figures can be up to 7.5% percent better as a result of the measurement

tolerances allowed by the ARI-550 test code, including a 5% deviation in power and

flow and a 0.28 oC (0.5 F) deviation in temperature.

3. NEW REFRIGERANTS IN OIL-FREE COMPRESSORS

3.1 Oil-free compressors allow easy transition to new refrigerants

Oil-free centrifugal compressors can relatively easy be converted to alternative

refrigerants. The oil-free compressor operation eliminates the identification or

development and subsequent qualification of lubricants that are compatible with the

new refrigerant. The new refrigerant compatibility studies are limited to the

elastomer (O-rings) and motor insulation materials.

3.2 Chiller global warming impact of HFO1234yf and HFO1234ze(E) The global warming impact of a refrigeration system consists of at least two effects:

- a direct effect (representing the global warming due to the leakage of refrigerant

molecules into the atmosphere) and

- an indirect effect (representing the amount of global warming as a result of energy

utilization of the refrigeration system which is related to its efficiency).

The industry has accepted the TEWI concept (Total Equivalent Warming Impact) [4]

to account for both effects. For systems notorious for their high refrigerant leak

rates such as commercial refrigeration and automotive air conditioning systems the

direct effect is the predominant cause of global warming. For residential and

commercial air conditioning equipment with its much lower leak rates the indirect

effect is the main contributor of global warming, with the exact amount of global

warming being determined by the percentage of power being generated by coal, oil

and natural gas versus renewable and nuclear power generation. As a result the

cycle efficiency of a low GWP refrigerant relative to the one it replaces is an

important characteristic. It can be categorically stated that for commercial and

residential HVAC equipment the cycle efficiency of a low GWP alternative

refrigerant should not be inferior to the existing refrigerant it is intended to replace.

Table 2 compares the ideal (= assuming 100% compressor efficiency) coefficient of

performance (COP) for typical water-cooled chiller conditions, using the latest

Refprop 9.0 refrigerant properties [5]. As can be seen from that table HFO1234yf

has a 3.2% lower COP than HFC134a and should therefore to be eliminated as a low

GWP alternative for commercial HVAC applications where the energy consumption

is a major contributor to global warming.

HFO1234ze has an ideal COP that is equal to that of HFC134a. A transition from

HFC134a to HFO1234ze does not increase the indirect global warming – like a

transition towards HFO1234yf would – and dramatically reduces the direct global

warming effect.

Table 2. Ideal cycle COP comparison for HFC134a, HFO1234yf and

HFO1234ze(E) showing an almost equal COP for HFC134a and HFO1234ze(E)

and a 3.2% lower COP for HFO1234yf

Input values in blue. Black input values derived from blue ones. Output values in red.

CYCLE CONDITIONS

Evaporature saturation Tevap,sat 5.56 [0C] or 42.0 [

0F]

Evaporator superheat ∆∆∆∆Tevap,sup 0.00 [0C] or 0.0 [0F]

Condenser saturation Tcond.sat 36.11 [0C] or 97.0 [0F]

Condenser subcooling ∆∆∆∆Tcond,sub 3.89 [0C] or 7.0 [0F]

Refrigeration capacity 1232 [kW] or 350 [ton]

Compressor specific speed ns 0.76 [-]

Compressor specific diameter ds 3.40 [-]

Fluid R134a R1234ze

Speed of sound m/s 146.71 139.07

∆∆∆∆hevaporator kJ/kg 156.89 144.13

∆∆∆∆hs,compressor kJ/kg 19.55 17.98

mdot kg/s 7.85 8.55

density kg/m3 17.45 14.20

Vdot m3 /s 0.450 0.602

N (Impeller Speed) rpm 17891 14524

D (Impeller diameter) m 0.193 0.228

u2 (tip speed) m/s 181 173

Pr(Pressure ratio) - 2.57 2.60

3.3 Predicted centrifugal compressor capacity change when replacing HFC134a with HFO1234ze(E) The refrigeration capacity of a compressor is the product of the evaporator enthalpy

rise and the compressor mass flow rate. Table 3 shows the inlet volumetric flow

rates required to achieve 350 ton of cooling. As can be seen HFO1234ze(E) requires

a 33% (.601/.450) larger volumetric flow rate. Since compressors are essentially

constant flow machines, drop-in compressor behavior will result in a capacity

shortfall of 25.2% for HFC1234ze(E). However, drop-in of alternative refrigerants

in centrifugal compressors requires a speed adjustment to insure surge-free operation

without over-compressor (choke) in order to maintain compressor peak efficiency

and turn-down capability [6]. As can be seen from Table 3, the isentropic enthalpy

rise needed to compress the refrigerant from the evaporator saturation temperature of

5.56 0C to the condenser saturation temperature of 36.11

0C is quite different for

these refrigerants. The head to be delivered by a compressor designed for HFC134a

(19.55 kJ/kg) is 9% larger than that required for HFO1234ze(E) (17.98 kJ/kg).

Since head is proportional to the square of speed, compressor drop-in applications

require a speed reduction of 4% for HFO1234ze(E). Since for pressure ratios around

2.5 the compressor flow varies with compressor speed to the power 1.6 [7], and

additional capacity reduction of 7% is anticipated as a result of the speed reduction

needed for optimum compressor performance.

As a result the capacity of an existing R134a compressor will be reduced by 32.2 %

(25.2% + 7%) for HFO1234ze(E). Since the platform of the four existing HFC134a

centrifugal compressors come in frame sizes that have upward capacity jumps of

about 50% (= downward capacity jumps of about 33%) most existing R134a

compressor applications could be fulfilled with HFO1234ze(E) by just selecting the

next available compressor frame size.

Table 3. Calculation of thermodynamic properties used to predict the

compressor performance change in terms of capacity and speed when

switching from HFC134a to HFO1234ze(E)

3.4 Maintaining refrigeration capacity by moving up a frame size when replacing HFC134a with HFO1234ze(E) Moving up a compressor frame size has an efficiency advantage. Compressor

efficiency increases with impeller size - both due to Reynolds number effect and

reduced relative surface roughness. For the relatively small, high-speed centrifugal

compressors in the 200 to 700 kWth capacity range, peak efficiency increases about

2% for each jump in frame size. Figure 2 shows the increase in efficiency when

going from one frame size (TT300) to the next larger frame size (TT350).

Figure 2. Comparison of aerodynamic efficiency for two adjacent compressor

frame sizes. Running the larger TT350 compressor with HFO1234ze(E) and comparing its

performance against the performance of the TT300 compressor with HFC134a

results in equal refrigeration capacity with a 2% boost in compressor efficiency for

HFO1234ze(E) as shown in Figure 2.

Figure 3. Comparison of aerodynamic efficiency for a TT300 DTC compressor

with HFC134a against a next frame size TT350 compressor with HFO1234ze(E)

assuming identical aero efficiency for HFO1234ze(E) and HFC134a.

A remaining question is whether compressor efficiency is affected by the change

from HFC134a to HFO1234ze(E). Compressor performance is controlled by many

factors. At identical impeller tip Mach number (u2/a0) we should expect identical

performance only to be corrected for differences in frictional losses. The 5.8% drop

in actual impeller speed required for head and flow factor similarity means that all

fluid velocities will be 5.8% lower when the compressor is running with

HFO1234ze(E) compared to HFC134a. Given the identical vapor kinematic

viscosities of HFC134a and HFO1234ze(E), it is to be expected that the frictional

and mixing losses which are proportional to the square of the fluid velocities will be

reduced by 12.0%. Assuming an 80% fluid efficiency this would mean a reduction

of the 20% fluid loss by 12.0% resulting in a 2.4 point increase in aero efficiency.

That anticipated compressor aero efficiency improvement was confirmed during

back-to-back testing of TT300 and TT350 compressors with HFC134a and

HFO1234ze(E). The test results for the TT300 compressor are summarized in

Figures 4 and 5. The 2 – 2.5 point higher aero efficiencies shown in Figure 4 will

result in a correspondingly higher head factor for equal tip Mach number lines on the

compressor map, as shown in Figure 5.

u2/a0=0.87 u2/a0=1.02 u2/a0=1.16

Figure 4. Back-to-back test results of a TT300 compressor with HFC134a and

HFO1234ze(E) at three identical tip Mach numbers showing a 2-2.5 point aero

efficiency improvement for HFO1234ze(E).

Figure 4. Back-to-back test results of the DTC TT300D compressor with

HFC134a and HFO1234ze(E) at three identical tip Mach numbers showing a

head factor increase corresponding to the measured aero efficiency improvement

for HFO1234ze(E).

CONCLUSIONS

• DTC compressor capacity can be maintained when replacing HFC134a with

HFO1234ze(E) by switching the existing fluid module to the next frame size

fluid module.

• Testing at DTC has indicated the potential of a 4.0 to 4.5 % efficiency

improvement for commercial chillers when switching from HFC134a to

HFO1234ze(E). This improvement consists of a 2% benefit obtained by

selecting the next frame size compressor combined with a 2-2.5% benefit as a

result of the apparent lower viscous losses of HFO1234ze(E) versus HFC134a

thanks to its lower impeller speed.

• This improved compressor efficiency has a bigger impact on reducing the carbon

footprint of the chiller than the extremely low GWP value of HFO1234ze(E) of 6

versus 1300 for HFC134a.

• For air-cooled chiller systems using HFO1234ze(E) a condenser redesign might

be is required to prevent excessive pressure drop as a result of the 50% higher

condenser volumetric flow rate at equal capacity that could negate the potential

performance benefit of HFO1234ze(E).

REFERENCES

1. Minor, B, Spatz, M, HFO-1234yf low GWP refrigerant update, Paper 2349

of the International Refrigeration and Air Conditioning Conference at

Purdue, July 14-17, 2008.

2. Yana Motta, S.F., Vera Becerra, E.D., Spatz, M.W, Analysis of LGWP

Alternatives for Small Refrigeration (Plugin) Applications, Paper 2499 of

the International Refrigeration and Air Conditioning Conference at Purdue,

July 12-15, 2010.

3. ARI 550/590 (I-P)-2011, Performance Rating of Water Chilling Packages

Using the Vapor Compression Cycle, Air-Conditioning, Heating, and

Refrigeration Institute (formerly ARI), 2011.

4. Sand, J.R., Fisher, S.K., Baxter, V.D., TEWI Analysis: Its Utility, Its

Shortcomings, and Its Results, International Conference on Atmospheric

Protection, Taipei, Taiwan, September 13-14, 1999.

5. NIST Reference Fluid Thermodynamic and Transport Properties Database

(REFPROP): Version 9.0 http://www.nist.gov/srd/nist23.cfm

6. Brasz, J.J., Centrifugal Compressor Behavior with Alternate Refrigerants

paper 96-WA/PID-2 presented at the 1996 ASME International Mechanical

Engineering Congress and Exhibition, Atlanta, Ga. November 17-22, 1996

7. Brasz, J.J., Variable-Speed Centrifugal Compressor Behavior with Low

GWP Refrigerants, 2009 IMechE conference on Compressors and their

Systems, September 7-9, 2009.