The impact of evaporator fouling and filtration on the performance of packaged air conditioners

International Journal of Refrigeration 30 (2007) 506e514 www.elsevier.com/locate/ijrefrig

The impact of evaporator fouling and filtration on the performance of packaged air conditioners Li Yang, James E. Braun*, Eckhard A. Groll Ray W. Herrick Laboratories, School of Mechanical Engineering, Purdue University, West Lafayette, IN 47906, USA Received 15 September 2005; received in revised form 28 July 2006; accepted 1 August 2006 Available online 24 October 2006

Abstract The goal of the study presented in this paper was to evaluate the impact of different filter types on the performance of three typical packaged air conditioners under both clean and fouled conditions. In a companion paper, combinations of six different levels of filtration and four different coils were tested under clean and fouled conditions. From the tests, it was found that fouling has a relatively small impact on air-side effective heat transfer coefficient, but can have a large impact on coil pressure drop. Data from the experimental study were used in developing simulation models for the three packaged air conditioners. Simulations show that the equipment cooling capacity is reduced with fouling primarily because of a decrease in air flow due to the increased pressure drop. In most cases, EER (energy efficiency ratio) was reduced with fouling primarily due to increased fan power. However, the changes in EER were relatively small, in the range of 1e10%. Equipment having low efficiency filters had higher EER after fouling than equipment with high efficiency filters, because high efficiency filters result in significantly higher pressure drops than low efficiency filters. The impact of the evaporator side fan efficiency was found to be significant. The energy penalty associated with high efficiency filters was reduced greatly when fan efficiency increased. Although high efficiency filters cause higher energy penalties they provide considerably better air quality. The quantity of dust passing through the coil with an MERV14 filter was approximately 30 times less than the dust passing the coil with an MERV4 filter. This difference was doubled when the MERV14 filter was compared to a case with no filter in place. Ó 2006 Elsevier Ltd and IIR. All rights reserved. Keywords: Air conditioner; Air conditioning; Evaporator; Filter; Modelling; Fouling; Pressure drop

Syste`mes de conditionnement d’air monoblocs: impacts de l’encrassement de l’e´vaporateur et de la filtration Mots cle´s : Conditionnement d’air ; Conditionneur d’air ; E´vaporateur ; Filtre ; Mode´lisation ; Encrassement ; Chute de pression

1. Introduction

* Corresponding author. Tel.: þ1 765 494 9157; fax: þ1 765 494 0787. E-mail address: [email protected] (J.E. Braun). 0140-7007/$35.00 Ó 2006 Elsevier Ltd and IIR. All rights reserved. doi:10.1016/j.ijrefrig.2006.08.010

The buildup of dust on an evaporator leads to an increase in air-side pressure drop and eventual reduction of heat transfer coefficient. The impact of fouling on evaporator coils with different upstream filters was presented in a

L. Yang et al. / International Journal of Refrigeration 30 (2007) 506e514

507

Nomenclature a b COP c d e EER g K Qc Tci V V_

coil pressure drop factor ½Pa sbcðf Þ mbcðf Þ  coil pressure drop exponent coefficient of performance filter pressure drop factor ½Pa sdcðf Þ mdcðf Þ  filter pressure drop exponent coil effective heat transfer coefficient factor ½W sgc ðm2þgc KÞ1  energy efficiency ratio [Btu h1 W1] coil effective heat transfer coefficient exponent air duct pressure drop factor [Pa s2 m2] equipment cooling capacity [W] condenser side air inlet temperature [ C] air velocity [m/s] air volumetric flow rate [m3/s]

companion paper by Yang et al. [1] and correlations of coil pressure drops and air-side effective heat transfer coefficients were obtained. The current paper focuses on equipment modeling and presents the impact of fouling on packaged air conditioner performance with different filterecoil combinations. There are only a few published studies that have addressed the effects of filter and coil fouling on overall air conditioning system performance. Krafthefer and Bonne [2] studied the buildup rate on coil surfaces, the need for scheduled cleaning and fouling effects on air pressure drop and system energy consumption for heat pumps. They estimated a 10e13% decrease in coefficient of performance (COP) for typical indoor coilefilter fouling of a heat pump. Furthermore, they estimated operating cost savings of 10e25% through the use of a high efficiency air filter upstream of the indoor coil. Breuker and Braun [3] conducted an experiment with a 3-ton rooftop unit under 96 conditions (combinations of 4 loads and 24 fault levels, including fouling faults). For the case of fouling, uniform condenser fouling was simulated in the test by blocking the condenser coil with strips of paper. The level of condenser fouling was expressed as a total percent reduction in the surface area. Evaporator fouling was simulated by reducing the air flow rate and it was expressed as a percent reduction from the nominal air flow rate. A 12% reduction in both cooling capacity and COP occurred for a 25% loss of evaporator air flow caused by fouling, while only around a 5% loss in capacity and an 8% loss in COP occurred when about 25% of the condenser coil was block due to fouling. The current study goes well beyond what has been done in studying the impact of fouling on air conditioner performance. It utilizes extensive testing that was performed in the laboratory for different filterecoil combinations and that is described in a companion paper by Yang et al. [1]. The data from this study were integrated with simulation models for three prototypical packaged air conditioners of different

W Wc DP h

power [W] compressor power [W] pressure drop [Pa] fan efficiency

Subscripts c clean coil coil dist air distribution system f fouled fe evaporator side fan fc condenser side fan filter filter sys system

sizes: a 123 kW (35 ton) unit (medium to large commercial), an 18 kW (5 ton) unit (small commercial) and an 11 kW (3 ton) unit (small commercial or residential). The models were used to evaluate the impact of fouling on cooling capacity and EER. More detailed descriptions of both the experimental and modeling studies are given by Yang et al. [4]. 2. Equipment energy analysis Fouling affects coil air-side heat transfer coefficients which influences the cooling capacity and efficiency of an air conditioner. Fouling also increases coil and filter pressure drop which reduces air flow rate and increases fan power, which in turn reduces cooling capacity and increases compressor power. As demonstrated by Yang et al. [1], the impact of fouling on pressure drop is more significant than the impact on heat transfer coefficient. In order to study the overall impact of fouling on equipment performance, a computer model was employed. Fouling factors from laboratory tests were used along with baseline simulations for small, medium, and large direct expansion packaged air conditioners. The fouled conditions were obtained after injection of a quantity of dust upstream of the filterecoil combination meant to simulate one year of operation in the field (see Ref. [1]). 2.1. Equipment descriptions Yang et al. [1] tested four coils and five types of upstream filters of different efficiencies. The four different coils included an eight-row wavy-fin coil (HX8W), an eight-row lanced-fin coil (HX8L), a four-row lanced-fin coil (HX4L), and a two-row lanced-fin coil (HX2L). Wavy-fin coils with more rows and lower fin densities are generally used in large equipment, while lanced-fin coils with higher fin densities and fewer rows are used in smaller systems. The MERV

L. Yang et al. / International Journal of Refrigeration 30 (2007) 506e514

values for the selected filters ranged from 4 to 14, which represent the efficiency range used in HVAC applications. Tests with no upstream filter were also conducted for each of the coils. Specific performance characteristics (pressure drop and heat transfer) for the different coil geometries were integrated into simulation models for equipment representative of residential, small commercial, and medium commercial applications. A public-domain program developed at Purdue, ACMODEL, was employed for the simulations. Table 1 gives information on the units considered. The 11 kW and 18 kW units are representative of residential and small commercial applications, respectively, and were simulated in previous studies where model predictions were validated with detailed experimental results (see Refs. [5,6]). The original models were modified to incorporate heat exchanger geometries consistent with those specified in Table 1 with pressure drop correlations and heat transfer fouling factors determined by Yang et al. [1] for coilefilter combinations under clean and fouled conditions. In addition, it was necessary to incorporate fan models for these units since they were not originally considered by LeRoy [5]. The 123 kW unit is typical of a larger commercial rooftop unit. In this case, the physical data necessary for simulating the unit were obtained from the manufacturer and included compressor map equations, condenser and evaporator configurations and fan curves. The original coil was replaced with the HX8W and HX8L coil geometries with pressure drop correlations and heat transfer fouling factors determined by Yang et al. [1] under clean and fouled conditions for different filters. Simulated performance for cooling capacity and EER of this unit were within 5% of published data at the rating conditions. Table 1 Description of the simulated air conditioning equipment 11 kW rooftop

18 kW rooftop

Refrigerant Compressor Exp. device Condenser rows Evaporator rows Evaporator coil type

R22 Reciprocating Scroll Thermostatic expansion valve 1 2 2 4 HX2L HX4L

Evaporator tube diameter (mm) Evaporator tube thickness (mm) Evaporator fin density (fin/cm) Evaporator fin thickness (mm) Evaporator face area (m2) Evaporator Nominal air flow (m3/s)

9.53

12.7

123 kW rooftop

2.2. Modeling process ACMODEL can predict the performance of unitary air conditioners and heat pumps and has been validated by Rossi [7], LeRoy [5], and Shen et al. [8]. In order to characterize performance, it is necessary to determine cooling capacity (Qc), compressor power (Wc), evaporator-side fan power (Wfe), condenser side fan power (Wfc), and EER. ACMODEL predicts both Qc and Wc, whereas Wfe and Wfc were determined using fan characteristics and system pressure drop estimates. In North America, EER is the net cooling capacity (difference between cooling capacity and evaporator-side fan power) in British thermal units per hour divided by the power input in watts at rating conditions or EER ¼

Qc  Wfe  3:412 Wc þ Wfe þ Wfc

For the units considered in this study, evaporator fan power (Wfe) had a very significant effect on EER. At a typical fan efficiency of 30e40%, Wfe was around 13e37% of the total power. Fouling reduces cooling capacity and EER due to two effects. First of all, fouling can reduce the heat transfer coefficient for the same air velocity. However, this is a relatively small effect as described by Yang et al. [1]. More importantly, fouling increases the flow resistance, which can reduce air flow and increase fan power. Fig. 1 depicts the system pressure drop and fan curve characteristics for both clean and fouled conditions. The fan operates at the intersection of the fan and system characteristics. As the coil and filter foul, the pressure drop increases and the air flow is reduced. In order to model fouling effects, the impacts of fouling on both heat transfer and pressure drop characteristics were considered using experimental data from Yang et al. [1]. The effect of increased pressure drop on air flow and fan power

Scroll 4 8 HX8L, HX8W 12.7

0.65

0.65

0.55

5.51

4.72

3.15

0.19

0.114

0.15

0.372

0.372

1.858

0.57

0.94

4.72

ð1Þ

ΔPsys,f Fan Curve fouled operating point

ΔPsys,c clean operating point

System Pressure Drop

508

V (reduced)

V (designed)

V Fig. 1. System pressure drop and fan curve.

L. Yang et al. / International Journal of Refrigeration 30 (2007) 506e514

509

was determined using manufacturers’ fan curves obtained for each air conditioning unit. For each combination of air conditioner and filter under clean conditions, it was assumed that the evaporator fan speed setting was adjusted to give the nominal air flow rates given in Table 1. The six fouling cases (five filter cases and one no-filter case) were simulated separately with lower air flow rates determined from the fan curve and system pressure drop characteristics specific to each case. The models used to determine the evaporator-side air pressure drop, fan power, and heat transfer degradation due to fouling are described in the following subsections.

and fouled conditions. For the fouling tests, different combinations of coils and filters were tested with dust loading representative of a year’s worth of operation with different upstream filters in place. For a given air velocity, the pressure drop model derived from the experimental results predicts significant increases in coil and filter pressure drops with fouling and decreasing filter efficiency. The system pressure drop includes frictional losses that occur along the entire air duct length and within fittings. This distribution system pressure loss would depend not only on the installation but also on the air velocity which changes with fouling. The distribution losses were assumed to depend on the square of the air velocity, so that

2.2.1. System pressure drop The total system static pressure drop (DPsys) included three parts: air-side pressure drop across the filter (DPfilter), coil (DPcoil) and air distribution system (DPdist). Mathematically, this is expressed as

DPdist;c ¼ KVc2

ð5aÞ

KVf2

ð5bÞ

DPsys;c ¼ DPfilter;c þ DPcoil;c þ DPdist;c

ð2aÞ

DPsys;f ¼ DPfilter;f þ DPcoil;f þ DPdist;f

ð2bÞ

DPdist;f ¼

The constant K was assumed not to be impacted by fouling. For the modeling results presented in this paper, the value of K was determined that would force each fan to run at its peak efficiency for the design air velocity with the highest efficiency filter tested (MERV14 for the two- and eight-row coils and MERV8 for the four-row coil) under clean conditions. Table 4 gives values for K and pressure drop at the design air velocity for the different air conditioners considered. For other filters and/or fouled conditions, the fan did not operate at its peak efficiency.

where the subscripts c and f refer to clean or fouled conditions, respectively. The pressure drops across the filter and coil increase with fouling due to increased flow resistance. To some extent, this is compensated for by a reduction in air velocity. Experimental data from Yang et al. [1] were used to correlate coil and filter pressure drop using the following expressions DPcoil;c ¼ ac Vcbc

ð3aÞ

af Vfbf

ð3bÞ

DPcoil;f ¼

cc Vcdc

ð4aÞ

DPfilter;f ¼ cf Vfdf

ð4bÞ

DPfilter;c ¼

2.2.2. Fan power The evaporator fan motor power for clean or fouled conditions depends on the system pressure drop, volumetric air flow rate, and combined fan/motor efficiency according to

where V is air velocity, the coefficients a, b, c, and d are empirical coefficients determined from the experimental data of Yang et al. [1], and the subscripts c and f refer to the clean and fouled conditions, respectively. Tables 2 and 3 give values for the pressure drop empirical coefficients for clean

Wfe;c ¼

V_ c DPsys;c hc

ð6aÞ

Wfe;f ¼

V_ f DPsys;f hf

ð6bÞ

where Wfe is evaporator fan power requirement, V_ is air volumetric flow rate (m3 s1), h is a combined fan/motor efficiency, and the subscripts c and f refer to clean and fouled

Table 2 Coil air-side pressure drop factors (unit of a: Pa sb mb) MERV14

MERV11

MERV8

MERV6

MERV4

No-filter

HX8L

ac bc

30.135 2.160

af bf

28.540 2.279

29.550 2.247

29.301 2.329

33.056 2.290

35.489 2.250

41.149 2.210

HX8W

ac bc

44.823 1.781

af bf

45.154 1.841

46.727 1.830

50.310 1.896

53.269 1.884

56.058 1.843

60.970 1.843

HX4L

ac bc

33.990 1.636

af bf

e e

e e

36.290 1.747

e e

41.733 1.708

60.271 1.811

HX2L

ac bc

27.477 1.229

af bf

25.149 1.402

29.503 1.272

28.467 1.380

32.338 1.320

34.107 1.319

79.999 1.265

L. Yang et al. / International Journal of Refrigeration 30 (2007) 506e514

510

Table 3 Filter pressure drop factors (unit of c: Pa sd md)

cc dc cf df

MERV14

MERV11

MERV8

MERV6

MERV4

40.942 1.716 95.236 1.430

38.581 1.453 79.53 1.585

19.496 1.542 51.431 1.646

35.051 1.280 38.942 1.978

19.459 1.375 30.052 1.543

conditions, respectively. The fan motor power could either increase or decrease with fouling depending on the fan characteristic and operating point. Fouling leads to a decrease in volumetric flow rate but an increase in system pressure drop as depicted in Fig. 1. The fan efficiency also changes with fouling because of a change in air flow rate and can either increase or decrease depending on the initial operating point for no fouling. It may increase with fouling if the flow rate is greater than the flow rate associated with the peak efficiency and decreases otherwise. Correlations for fan pressure rise and combined fan/motor efficiency as a function of air flow rate were developed from manufacturers’ data for each air conditioning unit and were incorporated in the model. 2.2.3. Heat transfer degradation Heat transfer correlations were determined from the experimental data of Yang et al. [1] for the different coil geometries considered in this study and then incorporated within ACMODEL. The effects of fouling on heat transfer coefficient were incorporated by including a fouling factor. The heat transfer coefficient for fouled and clean heat exchangers were then determined according to the following equations: hf ¼ fhc

ð7aÞ

hc ¼ ec V gc

ð7bÞ

where h is air-side heat transfer coefficients, f is a heat transfer fouling factor, the subscripts c and f refer to clean and fouled conditions, respectively, V is air velocity (for either clean or fouled conditions), and the coefficients e and g are empirical coefficients determined from the experimental data of Yang et al. [1]. Table 5 gives values for the heat transfer empirical coefficients for clean conditions and Table 6 gives fouling factors for all combinations of coils and filter cases. The fouling factors given in Table 6 are between about 0.82 and 1.04 and were incorporated in the model. The lowest value of f was associated with the two-row coil and no Table 4 Pressure drop associated with air distribution system at design air velocity 123 kW Rooftop 18 kW Rooftop 11 kW Rooftop DPdist (Pa) 672 K (Pa s2 m2) 104.2

150 23.2

125 53.9

Table 5 Coil air-side effective heat transfer coefficient factors (unit of ec: W sgc ðm2þgc KÞ1 )

ec gc

HX2L

HX4L

HX8W

HX8L

35.047 0.320

37.095 0.166

36.479 0.182

35.638 0.259

upstream filter. Values of f greater than 1 indicate a heat transfer enhancement and were obtained for the eight-row, wavy-fin coil with moderate efficiency filters (MERV8 and MERV11). For these cases, the relatively small amount of dust on the coil provided a heat transfer enhancement as described by Yang et al. [1]. 2.3. Baseline unit performance Fig. 2 presents baseline EER for the different units that were simulated as a function of ambient temperature. At rating conditions, the rated EER ranges from 8 and 10 for these units. Over the range of temperatures considered, EER decreases by 30e40% with outdoor temperature. 3. Fouling impact on equipment capacity and EER This section presents results for the impact of fouling on simulated cooling capacity and EER. All results are presented for a single rating condition having a condenser air Table 6 Effective heat transfer coefficient fouling factor of all test cases Air MERV14 MERV11 MERV8 MERV6 MERV4 Novelocity filter (m/s) HX8W 1.5 2.0 2.5 3.0

0.973 0.990 1.004 1.015

1.009 1.014 1.018 1.021

1.024 1.031 1.036 1.040

0.990 0.996 1.001 1.004

0.960 0.986 1.006 1.023

0.958 0.947 0.921 0.917

HX8L 1.5 2.0 2.5 3.0

0.962 0.942 0.927 0.915

0.952 0.951 0.950 0.950

0.926 0.935 0.942 0.948

0.974 0.976 0.978 0.980

1.032 1.024 1.017 1.012

0.962 0.943 0.931 0.925

HX4L 1.5 2.0 2.5 3.0

e e e e

e e e e

0.935 0.949 0.960 0.969

e e e e

0.934 0.950 0.963 0.974

0.933 0.935 0.937 0.939

HX2L 1.5 2.0 2.5 3.0

0.974 0.982 0.989 0.994

0.975 0.967 0.961 0.956

0.962 0.956 0.952 0.948

0.959 0.968 0.980 0.990

0.954 0.942 0.933 0.926

0.943 0.901 0.856 0.821

L. Yang et al. / International Journal of Refrigeration 30 (2007) 506e514

capacity degradation was about 4e5% for the 11 kW (HX2L) unit. The capacity degradation tends to be somewhat greater for smaller systems that employ shallower coils with higher fin densities. The decrease in cooling capacity was between 2% and 4% for the 123 kW unit (HX8L or HX8W) and between 5% and 7% for the 18 kW (HX4L) unit.

12.0 10.0

EER

8.0 6.0 HX8L

4.0

3.2. EER impact

HX8W HX4L

2.0

HX2L

0.0 25

30

35

40

45

Tci (°C) Fig. 2. Baseline EER vs. condenser air inlet temperature for different units described in Table 1.

inlet temperature of 35  C and an evaporator air inlet temperature and relative humidity of 26.7  C and 61%, respectively. 3.1. Capacity impact Fouling can affect equipment cooling capacity through reductions in both heat transfer coefficient and air flow. However, Yang et al. [1] found that the degradation in heat transfer coefficient for the same air flow was relatively small, even with significant fouling. Although the impact of fouling on pressure drop and air flow is much more significant, the simulated degradation in cooling capacity was found to be small after fouling with dust loading typical of a year’s worth of operation. Fig. 3 shows the cooling capacity degradation for all filter cases for the 11 kW (HX2L) unit as an example. The capacity ratio (capacity divided by the capacity of the no-filter case at clean conditions) was used as the vertical axis. All clean cases had the same cooling capacity because, as previously described, the fan settings were adjusted to give the design air flow rate (1.52 m/s for 11 kW unit). Overall the

EER was estimated for the different coilefilter combinations with clean and fouled conditions and normalized relative to the EER for no-filter clean cases. Figs. 4e7 present results for all the units considered. Degradations in EER with fouling or use of a high efficiency filter are mostly due to increased fan power. With this in mind, the following conclusions can be made from these results: (1) The clean equipment without an upstream filter had the highest EER for all coil cases because of the lowest system pressure drop. (2) The equipment with the highest MERV filter had the lowest EER for all clean coil cases because of a higher system pressure drop. It is interesting to note that the EER for use of the MERV8 filter was greater than that with the MERV6 filter. The MERV6 and MERV8 filters have different manufacturers and the MERV8 had a lower pressure drop under clean conditions than the MERV6 filter (see Ref. [1]). (3) The EER decreased significantly after fouling. The degradations ranged from 2% to 10% (2% for no-filter case) for the 123 kW units (HX8L and HX8W), 3% to 8% (8% for no-filter case) for the 18 kW unit (HX4L), and 6% to 10% (10% for no-filter case) for the 11 kW unit (HX2L). For the larger equipment with deeper coils, the impact of fouling was greater for the high efficiency filter, whereas the opposite was true for the smaller units. This is due to that fact that fouling has a greater impact on shallower coil designs as described by Yang et al. [1]. For the no filter fouling tests of Yang

1

1

0.9

0.9

EER Ratio

Capacity Ratio

511

0.8 0.7 0.6

clean fouled

0.5

0.8 0.7 0.6

clean fouled

0.5 4 V1

R ME

R ME

1 V1

ME

8 RV

ME

6 RV

ME

4 RV

r

lte ofi

N

Fig. 3. Capacity ratio of all coilefilter cases for 11 kW unit with HX2L.

1

4

V1

R ME

V1

R ME

RV

ME

8

V6

R ME

RV

ME

4

r

lte

fi No

Fig. 4. EER ratio of all coilefilter cases for 123 kW unit with HX8L.

L. Yang et al. / International Journal of Refrigeration 30 (2007) 506e514 1

1

0.9

0.9

EER Ratio

EER Ratio

512

0.8 0.7 0.6

clean fouled

0.7 0.6

clean fouled

0.5

0.5 4

1 RV

ME

8 RV

1

1 RV

ME

ME

6 RV

ME

ME

4 RV

er

V1

et al. [1], the coil pressure drop penalty was nearly four times greater and the heat transfer penalty was nearly twice as great for the two-row coils as compared with the eight-row coils. (4) The greatest impact of filter choice on EER occurred for the larger capacity equipment. This was caused by different fan performance: the 123 kW unit had a steep fan efficiency curve so that EER varied greatly with system pressure drop, which was influenced by both filter choice and fouling. The 11 kW unit had a more flat fan efficiency curve. The differences would be eliminated if these units had similar fan performance characteristics. The influence of fan curves will be discussed in the following subsection.

3.3. Influence of fan efficiency on EER For the previous results, the evaporator side fan power Wfe was approximately 13e37% of the total power with fan efficiencies ranging from about 30% to 40%. The results are sensitive to the fan curve and fan efficiency. To study the influence of fan efficiency, two cases were considered:

V1

R ME

R ME

N

8 RV

1

4

lt ofi

Fig. 5. EER ratio of all coilefilter cases for 123 kW unit with HX8W.

ME

4 RV

6

RV

ME

ME

er

filt

No

Fig. 7. EER ratio of all coilefilter cases for 11 kW unit HX2L.

fan efficiency were 38% for 123 kW (HX8L and HX8W) units, 28% for 18 kW (HX4L) unit and 29% for 11 kW (HX2L) unit. These efficiencies were the peak values for the actual fan curves. (2) The fan efficiency was assumed to be 100% and did not change with air flow rate. Figs. 8e11 show EER ratios for the equipment having the HX8L and HX2L coils for these two cases. In comparing these results with those of Figs. 4 and 7, it is apparent that the impact of fouling on EER is sensitive to the fan performance characteristics. Given a perfect fan curve (h ¼ 100%), the differences among all filter cases were much smaller than for the original analysis using the actual fan curves. In general, the penalty associated with using high efficiency filters was reduced for the case of 100% efficient fans. This was particularly true for the higher capacity unit. It is also interesting to note that for the case of 100% efficient fans, the absolute value of EER increased approximately 11% and 27% for the 11 kW and 123 kW units, respectively. Although it is not possible to design a 100% efficient fan, these results do indicate the importance of improving fan efficiency especially when considering the widespread use of high efficiency filters.

(1) The fan efficiency was assumed to not change with air flow rate and pressure drop. The assigned values of the

1 0.9

EER Ratio

1 0.9

EER Ratio

0.8

0.8

0.8 0.7 0.6

0.7 0.6

clean fouled

0.5 MERV14

MERV11

clean fouled

0.5 1

4

V1

R ME

V1

R ME

6

8

RV

ME

RV

ME

4 RV

ME

er

filt

No

MERV8

Fig. 6. EER ratio of all coilefilter cases for 18 kW unit with HX4L.

Fig. 8. EER ratio of all coilefilter cases for 123 kW unit with HX8L and constant fan efficiencies of 38%.

1

1

0.9

0.9

EER Ratio

EER Ratio

L. Yang et al. / International Journal of Refrigeration 30 (2007) 506e514

0.8 0.7 0.6

0.8 0.7 clean fouled

0.6

clean fouled

0.5

0.5 ME

1 RV

ME

8 RV

ME

6 RV

ME

4 RV

ME

r

4. Dust transmission The primary purpose of the filtration system is to capture dust and other particles. Data collected by Yang et al. [1] were used to determine dust quantities passing through the upstream filter and coil that would be supplied to the indoor area for different filterecoil combinations. Fig. 12 shows the percentage of dust passing the filter and coil relative to the injected dust for all cases. Without a filter, the coil and duct work capture a significant amount of dust. However, as previously demonstrated this dust buildup has an adverse effect on system performance, which would only get worse over years of dust accumulation. A moderate amount of filtering has a significant effect in reducing dust transmitted to the indoor space. However, the differences between MERV11 and MERV14 are small. It would seem that the MERV6 provides a good compromise between dust transmission and energy performance. 5. Conclusions and recommendations The following significant conclusions were obtained from the equipment modeling and test results: 1 0.9 0.8 0.7 0.6

clean fouled

0.5 11

14

RV

ME

RV

ME

ME

R

V8

ME

R

V6

ME

4 RV

V1

ter

l

fi No

6

8

RV

RV

ME

ME

4

RV

ME

er

filt

No

Fig. 11. EER ratio of all coilefilter cases for 11 kW unit with HX2L and constant fan efficiencies of 100%.

(1) Large equipment seems to be affected less by fouling than small equipment due to the use of deeper coils with lower fin densities. (2) Fouling decreases equipment cooling capacity because of reduced air flow rate. For a given air velocity, the heat transfer coefficient could decrease with fouling. However the decrease in air flow is a more significant effect. An average decrease of 8% in air flow rate after fouling was determined from simulations and test results. (3) The impact of filter choice on cooling capacity is relatively small. However, using high efficiency filters results in significant EER penalties for fouling especially for large equipment. For a 123 kW unit, the EER values decreased by 8e10% after fouling for MERV14 cases, but by 1e2% for MERV4 and no-filter cases. This was due to an increase in pressure drop and the influence of fan efficiency. (4) Primarily, fouling leads to reduced EER because of lower cooling capacity and higher fan power. Given actual fan curves with an efficiency of approximately 30%, the fan power was between 13% and 37% of the total power for the cases considered. The impact of fouling on fan power ranged from a 7% reduction to a 40% increase. The effect of fan efficiency was 35% HX8W

30%

HX8L HX4L

25%

HX2L

20% 15% 10% 5% 0%

11

14

RV

ME

Fig. 10. EER ratio of all coilefilter cases for 123 kW unit with HX8L and constant fan efficiencies of 100%.

V1

R ME

R ME

N

Fig. 9. EER ratio of all coilefilter cases for 11 kW unit with HX2L and constant fan efficiencies of 29%.

1

4

lte ofi

Amount of dust passing coil (%)

1

4

1 RV

EER Ratio

513

RV

ME

4 RV

6

8

RV

ME

RV

ME

ME

r

lte

fi No

Fig. 12. Dust quantities passing coil and filter for all test cases.

514

L. Yang et al. / International Journal of Refrigeration 30 (2007) 506e514

considered through simulation. The energy penalty associated with high efficiency filters was reduced considerably with higher efficiency fans. (5) Equipment with high efficiency upstream filters has lower EERs than equipment with low efficiency filters. This is because of increased pressure drop. (6) Equipment with high MERV upstream filters will provide significantly better air quality. For a 123 kW unit having an eight-row evaporator coil, the quantity of dust passing through the coil with an MERV4 filter was approximately 30 times the dust passing through the coil with an MERV14 filter. Without an upstream filter, the quantity of dust passing through the coil was approximately 60 times the value for an MERV14 filter. Only one year’s dust loading was considered in this work. At this level, the impacts of fouling on coil performance were relatively small. Further study with greater dust loadings is recommended. In addition, smaller particle sizes should be considered in future work. The ASHRAE standard dust used in this study primarily consists of coarse particles with sizes greater than about 6 mm. Acknowledgements This study was sponsored by the Air-conditioning and Refrigeration Technology Institute (ARTI) under ARTI 21CR Project Contract Number 611-40050. We appreciate the comments and support from people associated with the sponsoring organizations and all members of the ARTI 21-CR committee: Bill Freeborne, Steven R. Szymurski,

Barney Burroughs, Dieter Grether, Chuck Rose, and Souhel Stanbouly.

References [1] L. Yang, J.E. Braun, E.A. Groll, The impact of fouling on the performance of filtereevaporator combinations, International Journal of Refrigeration 30 (3) (2007) 489e498. [2] B.C. Krafthefer, U. Bonne, Energy use implications of methods to maintain heat exchanger coil cleanliness, ASHRAE Transactions 92 (Pt. 1B) (1986) 420e431. [3] M.S. Breuker, J.E. Braun, Common faults and their impacts for rooftop air conditioners, International Journal of HVAC&R Research 4 (3) (1998) 303e317. [4] L. Yang, J.E. Braun, E.A. Groll, The Role of Filtration in Maintaining Clean Heat Exchanger Coils, Final Report ARTI-21CR/611-40050-01, Air-Conditioning and Refrigeration Technology Institute (ARTI), 2004. [5] J.T. LeRoy, Capacity and power demand of unitary air conditioners and heat pumps under extreme temperature and humidity conditions, Master’s thesis, Herrick Labs, Purdue University, 1997. [6] J.T. LeRoy, E.A. Groll, J.E. Braun, Computer model predictions of dehumidification performance of unitary air conditioners and heat pumps under extreme operating conditions, ASHRAE Transactions 104 (Pt. 2) (1998) 773e788. [7] T.M. Rossi, Detection, diagnosis, and evaluation of faults in vapor compression cycle equipment, PhD thesis, Herrick Labs, Purdue University, Ind. Report No. 1796-3 HL 95-13, 1995. [8] B. Shen, E.A. Groll, J.E. Braun, Improvement and validation of unitary air conditioner and heat pump simulation models for R-22 and HFC alternatives at off-design conditions, Final Report for 1173-RP, American Society of Heating, Refrigerating, and Air-Conditioning Engineers (ASHRAE), 2006.