Applied Energy 86 (2009) 2451–2458
Contents lists available at ScienceDirect
Applied Energy journal homepage: www.elsevier.com/locate/apenergy
Determination of optimum working conditions R22 and R404A refrigerant mixtures in heat-pumps using Taguchi method K. Comakli *, F. Simsek, O. Comakli, B. Sahin Ataturk University, Mechanical Engineering Department, 25240 Erzurum, Turkey
a r t i c l e
i n f o
Article history: Received 3 August 2008 Received in revised form 25 February 2009 Accepted 1 March 2009 Available online 1 April 2009 Keywords: Heat-pump Gas mixture R404A R22 Taguchi method
a b s t r a c t In this study, refrigerants R22 and R404A five of their binary mixtures which contain about 0%, 25%, 50%, 75% and 100% mass fractions of R404A were tested. It is investigated experimentally the effects of gas mixture rate, evaporator air inlet temperature (from 24 to 32 °C), evaporator air mass flow rate (from 0.58 to 0.74 kg/s), condenser air inlet temperature (from 22 to 34 °C) and condenser air mass flow rate (from 0.57 to 0.73 kg/s) on the coefficient of performance (COP) and exergetic efficiency values of vapor compression heat-pump systems. To determine the effect of the chosen parameters on the system and optimum working conditions, an experimental design method suggested by Genichi Taguchi was used. In this study, it was observed that the most effective parameters are found to be the condenser air inlet temperature for COP and exergetic efficiency. Ó 2009 Elsevier Ltd. All rights reserved.
1. Introduction In recent years, ozone layer depletion, global warming and energy efficiency have become one of the most important global issues and scientist are proposed to resolve this issue. The refrigeration industry has gone through substantial changes due to the continuing debate on environmental issues such as depletion of the ozone layer, global warming. There is an urgent need to replace traditional refrigerants (CFCs and HCFCs) as the import of CFCs was banned in 1996 and HCFC imports will be progressively restricted, with complete phase-out early next century [1]. R22 has been used extensively as the refrigerant for residential heat-pump and air-conditioning systems for more than four decades due to its excellent safety, energy efficiency and operating characteristics. It is a partly halogenated refrigerant (HCFC) with a lifetime of approximately 20 years and ozone depletion potential (ODP) of 0.055 (Table 1). The Copenhagen Revision stated that substances of the HCFC type (controlled substances) ought to be frozen in 1996 with progressive reductions from 2004 to 2020 (99.5% cuts) and complete elimination by 2030. As R22 is gradually phased-out, non-ozone-depleting alternative refrigerants are being introduced. It has been proposed various substitutes to R22: R134a, R404A, R407C, R410A, R410B, R508, etc. (see Table 1). Among these alternatives, three directions seem to be gaining most favorable support depending on application and system design: the use of a look-alike zeotropic mixture such as 407C; the use of * Corresponding author. Tel.: +90 442 231 48 49; fax: +90 442 236 09 57. E-mail addresses:
[email protected],
[email protected] (K. Comakli). 0306-2619/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.apenergy.2009.03.003
higher pressure, nearly azeotropic mixtures R410A or R410B; and the use of the lower pressure refrigerant R134a [2]. Nowadays, it is well known that chlorine atoms liberated from chloro-fluorocarbons (CFCs) act as catalysts in ozone-depleting reaction and contribute to the greenhouse effect. Therefore, many actions have been performed to reduce the production and consumption of CFCs by different countries and international organizations. In 1987 the Montreal Protocol, an international environmental agreement, established requirements for the worldwide phase-out of ozone-depleting CFCs. The Montreal Protocol and further amendments to it led to the phase-out of CFCs in all developed countries in 1996, whereas the developing countries benefited from a more relaxed phase-out schedule [3–6]. The use of refrigerant mixtures in vapor compression heatpumps and other vapor compression cycles has been known for many years. There are two types of refrigerant mixtures: azeotropic and zeotropic mixtures. To improve performance, zeotropic refrigerant mixture is selected as a working fluid in the heat-pump. When using zeotropic refrigerant mixture, the temperature of the refrigerant in the evaporator and condenser does not constantly result from the difference in the boiling point of each refrigerant component. This phenomenon can be explained by the Lorenz cycle [7]. Many works have reported an increase in heat-pump performance when using zeotropic refrigerant mixture [8–22]. For example, Nuntaphan et al. [8] reported that the performance of the combined solar-water-heater and heat-pump was investigated using a simulation program. In this analysis, a model for a heatpump using the refrigerant mixture R22/R124/R152a was selected.
2452
K. Comakli et al. / Applied Energy 86 (2009) 2451–2458
Table 1 Characteristic feature of R22 and possible substitutes to R22 [8]. Refrigerant
R22
R-134a
R404A
R410A
R407C
R508A
R508B
Composition (wt%) Molar mass (kg/kmol) Boiling point at 1 atm (°C) Critical temperature (°C) Critical pressure (bar) Ozone depletion potential Global warming potential
– 86.48 40.80 96 49.9 0.055 1700
– 102.03 26.1 101.06 40.64 0 1300
R125/143a/134a (44/52/4) 97.6 46.5 72.1 37.32 0 3700
R32/125 (50/50) 72.59 51.81 70.17 47.70 0 1900
R32/125/134a (23/25/52) 86.2 43.7 86.05 46.34 0 1600
R23/116 (39/61) 100 87.4 11 37 0 12,000
R23/116 (46/54) 95.39 87.4 14 39.3 0 12,000
From the simulation program, the results show that the highest performance occurred at a mass fraction of R22 = 20%, R124 = 57% and R152a = 23%, a compressor speed of 20 rps, and mass flow rate of refrigerant at 0.01 kg/s. The coefficient of performance (COP) of this system is between 2.5 and 5.0. Ge and Cropper [9] developed a detailed simulation model for air-cooled finnedtube condensers using the distributed method. The results showed that useful energy savings are expected when this new control strategy is employed. The model does, however, predict that somewhat lower heat-load performances are expected when R404A is used to replace R22 in the condenser. Park et al. [10] investigated thermodynamic performance of R433A and HCFC22 is measured in a heat-pump bench tester under air-conditioning and heat-pumping conditions. Test results showed that the coefficient of performance of R433A is 4.9–7.6% higher than that of HCFC22 while the capacity of R433A is 1.0– 5.5% lower than that of HCFC22 for both conditions. Park et al. [11] studied performances of two pure hydrocarbons and seven mixtures composed of propylene, propane, HFC152a, and dimethylether were measured to substitute for HCFC22 in residential air-conditioners and heat-pumps. They showed that the coefficient of performance of these mixtures is up to 5.7% higher than that of HCFC22. Nanxi et al. [12] developed a new triple mixture named HTR01 for a moderately high temperature watersource heat-pump. Tests with 2.92 and 300 kW heat-pump systems showed that when the difference between the water temperatures of the condenser outlet and the evaporator inlet was less than 30 °C, the COPh was always larger than 3. He et al. [13] indicated that the COP of a heat-pump using a R22/R142b refrigerant mixture was higher than those of R22 and R142b by approximately 3.5%. Kiatsiriroat and Na Thalang [14] claim that optimum mass fraction of R22 in the refrigerant blend R22/R124/R152a is approximately 20–40% for an air-conditioning system. R404A, a zeotropic blend of R125/143a/134a (44/52/4 wt.%, respectively) is the only known substitute that might be regarded as a drop-in. Its compatibility with plant materials is good, except for the lubricating oils. Indeed, with R404A polyester oils must be adopted. Although replacement of R12 by R134a for refrigeration has been widely accepted and this is in commercial use, R134a R404A and R407C are not yet available for residential applications in the US, but are commonly found in residential air-conditioning systems and heat-pumps in Europe [23,24]. R22 is the most widely used refrigerant in air-conditioning applications in Turkey. The main purpose of this study was to investigate to the possibilities of using R404A as a working fluid to replace R22 for vapor compression heat-pumps. Considering that existing heat-pump units using R22 can continue to be serviced with R22 until 2020, this study also included the use of various mass mixtures of R22/R404A in the heat-pump. It is proposed that the new substitute refrigerants be used with making some changes to system components. Therefore, this study includes the preliminary results obtained with making no changes to system components. Our investigation continues in the redesigned heat-pump, which will include the comparison of new re-
sult with those of this investigation. The effect of mixture ratio gas mixture ratio, evaporator air inlet temperature, evaporator air mass flow rate, condenser air inlet temperature and condenser air mass flow rat on the COP and exergetic efficiency have not been investigated in detail, because it requires a vast number of experiments, which enormously increases the experimental cost and period. However, quantitative estimations of the various parameters affecting the performance of the heat-pump, and the main factors for optimum design can be determined by an optimization criterion. Therefore, the purpose of this study is to perform an optimization of design working conditions. The optimum working conditions of the parameters which affect the COP and exergetic efficiency of heat-pumps using R22/R404A refrigerant mixtures are investigated experimentally using the Taguchi method. Taguchi method consists of a plan of experiments with the objective of acquiring data in a controlled way, executing these experiments and analyzing data, in order to obtain information about the behavior of a given process. In other words, Taguchi method is an optimal parameter design of experimental tool, which first chooses several important parameters from a governing equation or relative characteristics of engineering, such as weight, length, or configuration and inputs them into one appropriate plan table designed by Taguchi with plural levels for each parameter. By comparing the calculated results for each parameter for each level from a response table, a set of optimal parameters with corresponding weight can be found. One of the advantages of the Taguchi method over the conventional experiment design methods, in addition to keeping the experimental cost at the minimum level, is that it minimizes the variability around the target when bringing the performance value to the target value. Another advantage is that optimum working conditions determined from the laboratory work can also be reproduced in the real production environment [25–39]. Taguchi method has been applied for various engineering systems, but the application of the Taguchi method for the energy based system has been scarce [32–39]. A methodology to work on geometrically complex heat transfer systems was investigated by Nakayama [32] using the Taguchi method and through a genetic algorithm-type reasoning. The methodology was demonstrated on the cases of heat conduction through composite slabs. Optimum design of natural-circulation solar-water-heater by the Taguchi method was presented by Lu et al. [33]. Bilen et al. [34], Sahin [35], Sahin and Demir [36,37] applied Taguchi method to enhancement heat transfer. They showed that the Taguchi method can successfully be applied to heat transfer studies. Yun and Lee [38] analyzed the effect of various design parameters on the heat transfer and pressure drop characteristics of the heat exchanger with a slit using the Taguchi method. The optimum design value of each parameter was presented and the reproducibility of the results was discussed. To the authors’ knowledge there has been little work on the application of the Taguchi method on heat-pumps. Comakli et al. [39] determined optimum working conditions in heat-pumps using nonazeotropic refrigerant mixtures.
2453
K. Comakli et al. / Applied Energy 86 (2009) 2451–2458
denser and compressor. Six Bourdon type manometers were installed at the inlets and outlets of the condenser, compressor and evaporator to measure the pressures. Compressor input current was measured using an ampermeter. The overall composition of the refrigerant mixtures was determined by measuring the charging quantity of individual refrigerant using a precision digital balance. First, the system was charged with R22. Temperature and pressure values in the key-points of the plant, as shown in Fig. 1, were continuously monitored in order to check the achievement of steady-state conditions. Usually, the start-up time required about 1 h. After each experimental run, the raw data which consisted of temperatures from the thermocouples, pressures from the manometers, air flow rate from the flow meter, compressor input current and volt from the ammeter were recorded. On completion of the measurements to obtain the baseline performance results, the
2. Experimental set-up and technique A schematic diagram of the experimental apparatus is shown in Fig. 1. The main elements of the system are a scroll compressor powered by a 1.5 HP, a plate-fin air-cooled condenser, thermostatic expansion valve, a plate-fin air-cooled evaporator, electrical air heater in order to keep the air temperature passing into the evaporator and condenser at the desired level. Sight glass is installed at the outlets from the evaporator to provide visual confirmation of the refrigerant state and other auxiliary equipment. Temperature and pressure of the working fluid were measured at several locations of interest, as shown in Fig. 1. K-type chromenickel thermocouples were used to measure the temperatures of the working fluid, and the thermocouples were calibrated with a digital temperature controller. The working fluid temperatures were measured at the inlets and outlets of the evaporator, con-
Heater T T
Condenser T
T
Air
Fan (4750 m3/h)
P
T
Oil Separator
BY-PAS VALF
P
Pressure
T
Temperature
Compressor P T DRAYER
P
S. VALF
T DRYER
RESIVER
Sight Glass Heater
T
EXP. VALF
Evaporator T T
Air
Condense
Fig. 1. Schematic view of the experimental set-up.
Fan (4750 m3/h)
2454
K. Comakli et al. / Applied Energy 86 (2009) 2451–2458
Table 2 The parameters and the values corresponding to their levels studied in experiments. Parameters
A B C D E
Levels
Gas mixture ratio, R22/R404A (%) Evaporator air mass flow rate (kg/s) Evaporator air inlet temperature (°C) Condenser air mass flow rate (kg/s) Condenser air inlet temperature (°C)
1
2
3
4
5
0 0.58 24 0.57 34
25 0.62 26 0.61 32
50 0.66 28 0.65 28
75 0.70 30 0.69 24
100 0.74 32 0.73 22
where U, I and Cos u are voltage (V), current (A) and power factor, respectively. The power factor is measured from ampermeter. The coefficient of performance (COP) for a heat-pump cycle indicates the overall power consumption for a desired output and was evaluated using the following equation:
COP ¼
Q_ c _C W
The exergetic efficiency of the system is calculated by
_ h m
compressor and the system were drained and evacuated. The system then charged with the refrigerant to be tested. This procedure was repeated for R404A and each mixture under investigation. The tested mixtures were mixtures of R22/404A (mass fractions of R404A was 0%, 25%, 50%, 75% and 100%). The experimental conditions used in the work are given in Table 2 collectively.
3. Data reduction 3.1. Determination of heat-pump characteristics After each experimental run, raw data which consist of temperatures from the thermocouples, pressures from the manometers, air flow rate from the flow meter, and compressor input power from the watt meter and refrigerant mass from the precision balance are recorded. From the measured parameters: The heat delivered by the condenser is calculated by
_ r ðhC;out hC;in Þ Q_ C ¼ m
ð1Þ
The heat extracted by the evaporator is calculated by
_ r ðhE;in hE;out Þ Q_ E ¼ m
ð2Þ
The power input to the compressor is calculated by
_C¼ W
pffiffiffi 3 Cos u U I W
ð3Þ
ð4Þ
h
T ðs
s
Þ
0 C;out c;in C;in gex ¼ _ r c;out _ r hE;in hE;out T 0 ðsE;in sE;out Þ WC þ m
ð5Þ
3.2. Experimental plan for optimization When analyzing experimental results, the variability in the performance of the product is reduced by using control parameters, to the target value by using the adjustment parameter(s). Taguchi, during this process, suggested the use of a new experimental design method different from conventional ones such as orthogonal array (OA), performance statistics (signal to noise ratio), loss function, and etc. Optimum working conditions determined by using the experimental data should always be able to provide the same or very close performance values (i.e., COP and exergetic efficiency) at different times and different working environments. The optimization criteria used to meet this task should be able to control both whether the target value has been reached, and also whether the variability around the target is kept at the minimum. According to Taguchi, performance statistics can meet the above requirement. Kackar [25] reported that Taguchi has developed more than 60 performance statistics which can be used depending on the problem being investigated. The Taguchi method consists of planned experiments, with the objective of acquiring data in a controlled way, executing these experiments and analyzing data in order to obtain information about the behavior of the process. One of the advantages of the Taguchi method over the conventional experimental methods, in addition to reduction in the experimental cost, is that
Table 3 Chosen L25(56) experimental plan. Experiment no.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Parameters and their levels
COP
Exergy efficiency (%)
A
B
C
D
E
F (empty)
Repetition 1
Repetition 2
Repetition 1
Repetition 2
1 1 1 1 1 2 2 2 2 2 3 3 3 3 3 4 4 4 4 4 5 5 5 5 5
1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5
1 2 3 4 5 2 3 4 5 1 3 4 5 1 2 4 5 1 2 3 5 1 2 3 4
1 2 3 4 5 3 4 5 1 2 5 1 2 3 4 2 3 4 5 1 4 5 1 2 3
1 2 3 4 5 4 5 1 2 3 2 3 4 5 1 5 1 2 3 4 3 4 5 1 2
1 2 3 4 5 5 1 2 3 4 4 5 1 2 3 3 4 5 1 2 2 3 4 5 1
0.46 0.90 1.69 2.34 3.04 1.82 2.49 0.06 0.09 0.74 0.38 1.40 2.29 2.78 0.36 2.36 0.37 0.33 1.30 2.21 1.43 2.05 2.36 0.21 0.82
0.60 1.01 1.50 2.50 3.07 1.56 2.33 0.07 0.10 1.10 0.14 1.36 2.47 2.39 0.14 2.80 0.27 0.68 1.54 2.38 1.46 2.11 2.42 0.23 0.87
2.68 4.33 6.69 7.28 7.90 4.49 5.34 0.31 0.37 2.60 1.60 4.77 6.24 6.82 1.70 5.41 1.89 1.58 5.03 6.76 5.02 6.27 5.67 1.01 3.47
2.75 4.40 6.45 7.31 7.50 4.06 5.13 0.35 0.42 3.77 0.66 4.96 6.48 5.57 0.63 7.48 1.31 3.20 6.10 7.03 5.27 6.97 6.53 1.15 3.88
2455
K. Comakli et al. / Applied Energy 86 (2009) 2451–2458 Table 4 Uncertainties in the values of the relevant variables. Variables
Uncertainty (%)
Temperature (T) Pressure (P) Voltage (V) Current (I) Power factor (Cos u) Mass flow rate (mr)
2.5 1.6 1.7 1.7 1.7 1.3
1.2 1 0.8 0
25
50 Mixture ratio (%)
75
100 404A
ð6Þ
6,2 5,2 4,2 3,2 2,2 0
25
50
R22
1.5 1.4 1.3 1.2 0.58
0.62
0.66
0.7
0.74
Performance statistics
1.6
75
100
Mixture ratio (%)
404A
5,2 4,7 4,2 3,7 3,2 0,58
0,62
0,66
0,7
0,74
28 Ev. inlet temperature ( oC)
30
32
0,65
0,69
0,73
Ev. Air flow rate (kg/s)
1.5 1.5 1.4 1.4 1.3 24
26
28 o Ev. inlet temperature( C)
30
32
Performance statistics
Performance statistics
!
where ZL is the performance statistics, n the number of repetitions done for an experimental combinations, and Yi the performance value of the ith experiment.
Performance statistics
1.6 1.4
Ev. Air flow rate (kg/s)
Performance statistics
n 1X 1 n i¼1 Y 2i
1.8
R22
Performance statistics
Z L ¼ 10Log
4,6 4,5 4,4 4,3 4,2 4,1 4,0 24
26
1.5 1.4 1.3 1.2 0.57
0.61
0.65
0.69
0.73
Performance statistics
Performance statistics
it minimizes the variability around the target when bringing the performance value to the target value. Another advantage is that optimal working conditions, determined from the laboratory study, can be reproduced in real applications. The Taguchi method will not be explained here. However, readers who are interested in the method are referred to Refs. [25–30]. The orthogonal array (OA) experimental design method was chosen to determine the experimental plan, five parameters with five levels (values) each, L25(56) (Table 3), since it is the most suit-
able for the conditions being investigated [26]. L25(56) means that six parameters and five levels was studied. The parameters affecting gas mixture ratio, namely evaporator air inlet temperature, evaporator air mass flow rate, condenser water inlet temperature and condenser water mass flow rate are inserted in columns A, B, C, D and E. Columns F that is empty is used for calculation of error. In order to observe the effect of noise sources on the COP and exergetic efficiency, each experiment was repeated twice under the same conditions at the different times. The aim is to obtain the maximum COP and exergetic efficiency. The performance statistics was chosen as the optimization criterion and was used for ‘the larger the better’, and was evaluated by using the following equations [25]:
Cond. air flow rate (kg/s)
4,7 4,5 4,3 4,1 3,9 3,7 3,5 0,57
0,61
3.0 2.5 2.0 1.5 1.0 0.5 0.0 22
24
26
28
30
32
34
Cond. inlet temperature ( oC)
Fig. 2. The effect of each parameter on coefficient of performance (COP).
Performance statistics
Performance statistics
Cond. Air flow rate (kg/s) 8,0 6,0 4,0 2,0 0,0 22
24
26 28 30 Cond. inlet temperature ( oC)
32
Fig. 3. The effect of each parameter on exergetic efficiencies.
34
2456
K. Comakli et al. / Applied Energy 86 (2009) 2451–2458
In the Taguchi method, the experiment corresponding to the optimum working conditions might not have been undertaken during the whole period of experimentation. In such cases, the performance value corresponding to optimum working conditions can be predicted by utilizing the balanced characteristic of the OA. For this aim, the additive model may be used [28]:
Y i ¼ l þ X i þ ei
ð7Þ
where l is the overall mean performance value, Xi the fixed effect of the parameter level combination used in the ith experiment, and ei the random error in the ith experiment. The optimum performance is calculated using Eq. (7). Because Eq. (7) provides a point estimation calculated by using experimental data to determine whether the results of the confirmation experiments are meaningful or not, the confidence interval must be evaluated. The confidence interval at a chosen error level may be calculated by [27–29]
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1þm 1 Y i F a;1;DF MSe MSe þ N nr
ð8Þ
where F is the value of F table, a the error level, DFMSe the degree of freedom of mean square error, m the degree of freedom used in the prediction of Yi, N the number of total experiments, and nr the number of repetitions in the confirmation experiments. If the experimental results are in percentages, before evaluating Eqs. (7) and (8), first the omega transformation of the percentage values should be applied using the following equation and then the interested values are determined by the reverse transformation using the same equation [27];
XðdbÞ ¼ 10Log
1 1 P
ð9Þ
The order of the experiments was obtained by inserting parameters into the columns of OA, L25(56), which is chosen according to the experimental plan given in Table 3. However, the order of experiments was chosen random in order to avoid noise sources which had not been considered initially and which could take place during an experiment and affect the results in a negative way. The interactive effects of the parameters were not taken into account in the theoretical analysis because some preliminary tests showed
that they could be neglected. The validity of this assumption was checked by confirming experiments conducted at the optimum conditions. 3.3. Experimental uncertainties By using the estimation method of Kline and McClintock [40], maximum uncertainties of the COP and exergy efficiency are found as follows: COP, 3.53%; exergy efficiency, 3.53%. The individual contributions to the uncertainties of the COP and exergy efficiency for each of the measured physical properties are summarized in Table 4. 4. Results and discussion The collected data were analyzed by using ANOVA-TM computer software package for evaluation of the effect of each parameter on the optimization criteria. The results obtained are given in Figs. 2 and 3. The order of graphs in Figs. 2 and 3 are according to the degree of the influence of parameters on the performance statistics. At the first sight, it is difficult and complicated to deduct experimental conditions for the graphs given in these figures. Let us take Fig. 2, which shows the variation of the performance statistics with parameters on COP. Also, let us try to determine experimental conditions for the first point. The evaporator air inlet temperature for this point is 24 °C that is level 1 for which column C is 1. The performance statistics value of the first data point is, thus, the average of those obtained from experimental nos. 1, 10, 14, 18 and 22. The experimental conditions for the second point are the conditions of the experiments for which column C is 2 (i.e., experiments nos. 2, 6, 15, 19 and 23), and so on. The numerical value of the maximum point in each graph shows the best value of that particular parameter, given in Table 5 for each parameter, and they indicate the optimum condition in the range of the experimental conditions. The most effective parameters on COP and the exergetic efficiency are found as follows; condenser air inlet temperature, gas mixture ratio, evaporator air mass flow rate, and evaporator air inlet temperature and condenser air mass low rate, respectively. Generally, COP and the exergetic efficiency increases by increasing evaporator air in-
Table 5 Optimum working conditions and performance values obtained this condition for heat-pump system. Parameters
COP
Exergy efficiency (%)
General
a b c d e
COP
Exergy efficiency (%)
A Gas mixture ratio
B Evaporator air mass flow rate (kg/s)
C Evaporator air inlet temperature (°C)
D Condenser air mass flow rate (kg/s)
E Condenser air inlet temperature (°C)
Prediction Confidence Real Prediction Confidence Real interval interval
Optimum level Optimum value
1b
5c
5d
2e
5a
3.10
2.78–3.42
3.21 7.98
6.75–9.23
8.67
0
0.74
32
0.61
22
Optimum level
1b
5c
4d
3e
5a
3.08
2.84–3.32
3.15 7.95
7.03–8.87
8.28
Optimum value
0
0.74
30
0.65
22
Optimum level Optimum value
1
5
5
2
5
3.10
2.78–3.42
3.21 7.98
6.75–9.23
8.67
0
0.74
32
0.61
22
1st Degree effective parameter. 2nd Degree effective parameter. 3rd Degree effective parameter. 4th Degree effective parameter. 5th Degree effective parameter.
K. Comakli et al. / Applied Energy 86 (2009) 2451–2458
let temperature and evaporator air mass flow rate. COP and the exergetic efficiency decreases by increasing gas mixture ratio, condenser air inlet temperature (see Fig. 2). The condenser air mass flow rate has low effect on the COP and the exergetic efficiency. By using Figs. 2 and 3, optimum working conditions were determined and given in Table 5. First of all, each goal was optimised, separately. Then, all these goals were optimised together, considering the priority of the goals, and the results of the optimization were presented in Table 5 entitled ‘‘General”. If the experimental plan given in Table 3 is examined carefully together with Table 2, it can be seen that the experiment corresponding to optimum conditions (i.e., for COP, A1B5C5D2E5, see Table 5) was not performed during experimental plan. For instance, the COP predicted by using Eq. (7) and the result of the prediction were presented in Table 5 entitled ‘‘Prediction”. Also, a 95% significance level confidence interval of prediction is calculated by using Eq. (8) and the results were presented in Table 5 entitled ‘‘Confidence interval”. To test predicted results, confirmation experiments were carried out twice at the optimum conditions and the average of results were presented in Table 5 entitled ‘‘Real”. In order to test the predicted results, confirmation experiments were conducted twice at the optimum conditions. From the fact that the behavior of heat-pump systems obtained from confirmation experiments (i.e., ‘‘Real” COP of 3.15, 3.27 and 3.22, 3.21 being the average) are within the calculated confidence intervals (2.78– 3.42), it can be said that experimental results are within ±5% error (see Table 5). These results show that the interactive effects of the parameters are, indeed, negligible and also prove that the Taguchi method can successfully be applied heat-pump systems, with a very limited number of experiments and shorter time to obtain the optimum values of parameters.
5. Conclusions Our research demonstrates that R22/R404A mixtures can be used in replacement for R22 or R407A in vapor compression heat-pump systems. The most effective parameters and the optimum conditions for the heat-pump systems were determined. The following conclusions can be derived from the above results and discussion. It is shown from the results that optimum combination of gas mixture is 100% R22 and 0% R404A. The reason for this, thermodynamic properties of the R22 are better than R404A. In addition, as mentioned in the introduction section the main purpose of this study was to investigate to the possibilities of using R404A as a working fluid to replace R22 for vapor compression heat-pumps. Therefore, the experimental system was built according to use R22. So, it is expected result that 100% R22 is the most effective working fluid in this system. For R22/R404A gas mixture, it is seen that the performance of heat-pump is the best in gas mixture of 50%R22–50%R404A and 25%R22–75%R404A. As a result, as the usage of R22 is limited, R404A or R22/R404A gas mixture can be used instead of R22 in transition period. However, R404A is more expensive than R22. Thus, 50%R22–R404A gas mixture should be chosen in heat-pump in which R22 is used. The most effective parameter on the COP is found to be the condenser air mass flow rate. The maximum COP was observed at 22 °C condenser air inlet temperature, for pure R22 refrigerant, for 32 °C evaporator air inlet temperature, for 0.74 kg/s evaporator air mass flow rate and for 0.61 kg/s condenser air mass flow rate. The most effective parameter on the exergetic efficiency is found to be condenser air inlet temperature. The maximum exergetic efficiency was observed at 22 °C condenser air inlet temper-
2457
ature, pure R22 refrigerant, for 30 °C evaporator air inlet temperature, for 0.74 kg/s evaporator air mass flow rate and for 0.65 kg/s condenser air mass flow rate. Optimum working conditions were obtained as A1B5C5D2E5 for COP and A1B5C4D3E5 for exergetic efficiency. The source temperature (condenser and evaporator temperatures) is an effective factor for exergetic efficiency. Thus, optimum conditions of exergetic efficiency is different from optimum conditions of COP. When all the goals were take into account together, the trade-off among goals was considered and the optimum results were obtained for exergetic efficiency at A1B5C5D2E5 conditions. Since COP is more general criteria than exergetic efficiency in literature evaluating heat-pump characteristic, general optimum levels obtained COP was chosen as optimum conditions for this study. Because the optimum conditions determined by the Taguchi method in a laboratory environment are also reproducible in real production environments, and the findings of the present laboratory scale study may be very useful for heat-pump applications in industrial scale. Acknowledgements This work was supported by The Turkish Scientific and Technological Research Council of Turkey (TUBITAK Project No. 105M030) and Ataturk University, Research Project Foundation (Project No. BAP-2005/16). The Authors wish to thank to TUBITAK and Ataturk University. References [1] Purkayastha B, Bansal PK. An experimental study on HC290 and a commercial liquefied petroleum gas (LPG) mix as suitable replacements for HCFC22. Int J Refrig 1998;21:3–17. [2] Karagoz S, Yilmaz M, Comakli O, Ozyurt O. R134a and various mixtures of R22/ R134a as an alternative to R22 in vapour compression heat pumps. Energy Convers Manage 2004;45:181–96. [3] Cavallini A. Working fluids for mechanical refrigeration, review paper. Int J Refrig 1996;19(8):485–96. [4] Mattarolo L. Refrigerants and environment protection. In: I. Ulusal Sog˘utma ve _ _ Iklimlendirme Sempozyumu, Istanbul Turkey, 16–18 Mayıs; 1990. p. 51–70. [5] McMullan JT. Refrigeration and the environment – issues and strategies for the future. Int J Refrig 2002;25:89–99. [6] Rachidi T, Bernatchou A, Charia M, Loutfi H. New fluids as substitute refrigerants for R12. Sol Energy Mater Solar Cells 1997;46:333–47. [7] Euakit T. Performance study of automobile air conditioning system with R22/ R152a/R124 refrigerant blend. D.Eng Thesis, King Mongkut’s Institute of Technology Thonburi, Thailand; 1996. [8] Nuntaphan A, Chansena C, Kiatsiriroat T. Performance analysis of solar water heater combined with heat pump using refrigerant mixture. Appl Energy 2009;86(5):748–56. [9] Ge Y, Cropper R. Air-cooled condensers in retail systems using R22 and R404A refrigerants. Appl Energy 2004;78(1):95–110. [10] Park K, Shim Y, Jung Dongsoo. Performance of R433A for replacing HCFC22 used in residential air-conditioners and heat pumps. Appl Energy 2008;85(9):896–900. [11] Park K, Seo T, Jung D. Performance of alternative refrigerants for residential air-conditioning applications. Appl Energy 2007;84(10):985–91. [12] Nanxi L, Shi L, Lizhong H, Mingshan Z. Moderately high temperature water source heat-pumps using a near-azeotropic refrigerant mixture. Appl Energy 2005;80(4):435–47. [13] He X, Spindler UC, Jung DS. Investigation of R22/R142b mixture as a substitute for R12 in single evaporator domestic refrigerators. ASHRAE Trans 1992;98:150–9. [14] Kiatsiriroat T, Na Thalang K. Performance analysis of vapor compression refrigeration with R22/R124/R152a refrigerant. Int J Energy Res 1997;21:221–32. [15] Kiatsiriroat T, Euakit T. Performance analysis of an automobile airconditioning system with R22/R124/R152a refrigerant. Appl Therm Eng 1997;17:1085–97. [16] Ge YT, Cropper R. Performance simulation of refrigerated display cabinets operating with refrigerants R22 and R404A. Appl Energy 2008;85(8):694–707. [17] Arcaklıoglu E, Çavusoglu A, Erisen A. Thermodynamic analyses of refrigerant mixtures using artificial neural networks. Appl Energy 2004;78:219–30. [18] Perez-Blanco H, Kim K, Ream S. Evaporatively-cooled compression using a high-pressure refrigerant. Appl Energy 2007;84:1028–43. [19] Smith SJ, Shao L, Riffat SB. Pressure drop of HFC refrigerants inside evaporator and condenser coils as determined by CFD. Appl Energy 2001;70:169–78.
2458
K. Comakli et al. / Applied Energy 86 (2009) 2451–2458
[20] Medrano M, Bourouis M, Coronas A. Double-lift absorption refrigeration cycles driven by low-temperature heat sources using organic fluid mixtures as working pairs. Appl Energy 2001;68:173–85. [21] Wu S, Eames IW. Innovations in vapour-absorption cycles. Appl Energy 2000;66:251–66. [22] Cheong KW, Riffat SB. Monitoring hydrofluorocarbon refrigerant leakage from air-conditioning systems in buildings. Appl Energy 1996;53:341–7. [23] Aprea C, Greco A. An experimental evaluation of the greenhouse effect in R22 substitution. Energy Convers Manage 1998;39:877–87. [24] US Environmental Protecton Agency.
. [25] Kackar RN. Off-line quality control, parameter design and Taguchi method. J Quality Eng 1985;17:176–209. [26] Phadke MS. Quality engineering using robust design. New Jersey: Prentice Hall; 1986. [27] Taguchi G. Taguchi techniques for quality engineering. New York: Quality Resources; 1987. [28] Phadke MS, Kackar RN, Speeney DV, Grieco MJ. Off-Line quality control in integrated circuit fabrication using experimental design. Bell Syst Tech J 1983;62:1273–309. [29] Ross PJ. Taguchi techniques for quality engineering. New York: McGraw; 1987. [30] Roy RK. Design of experiments using the Taguchi approach. John Wiley & Sons Inc.; 2001. [31] Çelik C, Burnak N. A systematic approach to the solution of design optimisation problem. Total Quality Manage 1998;9:101–8.
[32] Nakayama W. A methodology to work on geometrically complex heat transfer systems: the cases of heat conduction through composite slabs. Int J Heat Mass Transfer 2003;46:3397–409. [33] Lu SM, Li YCM, Tang JC. Optimum design of natural-circulation solar-waterheater by the Taguchi method. Energy 2003;28:741–50. [34] Bilen K, Yapici S, Celik C. A Taguchi approach for investigation of heat transfer from a surface equipped with rectangular blocks. Energy Convers Manage 2001;42:951–61. [35] Sahin B. A Taguchi approach for determination of optimum design parameters for a heat exchanger having circular-cross sectional pin fins. Heat Mass Transfer 2007;45(3):493–502. [36] Sahin B, Demir A. Performance analysis of a heat exchanger having perforated square fins. Appl Therm Eng 2008;28(5-6):621–32. [37] Sß ahin B, Demir A. Thermal performance analysis and optimum design parameters of heat exchanger having perforated pin fins. Energy Convers Mange 2008;49(6):1684–95. [38] Yun JY, Lee KS. Influence of design parameters on the heat transfer and flow friction characteristics of the heat exchanger with slit fins. Int J Heat Mass Transfer 2000;43:2529–39. [39] Comakli O, Celik C, Erdogan S. Determination of optimum working conditions in heat-pumps using nonazeotropic refrigerant mixtures. Energy Convers Mange 1999;40:193–203. [40] Kline SJ, McClinctock FA. Describing uncertainties in single-sample experiments. Mech Eng 1953;75:3–8.