- Email: [email protected]

Contents lists available at ScienceDirect

Energy and Buildings journal homepage: www.elsevier.com/locate/enbuild

Optimization of operating parameters of a ground coupled heat pump system by Taguchi method Hikmet Esen ∗ , Emre Turgut Department of Energy Systems Engineering, Faculty of Technology, Firat University, 23119 Elazig, Turkey

a r t i c l e

i n f o

Article history: Received 23 July 2015 Received in revised form 21 August 2015 Accepted 22 August 2015 Available online 24 August 2015 Keywords: Optimization Depth Ground coupled heat pump Vertical Taguchi method

a b s t r a c t This paper focuses on optimization of vertical ground coupled heat pump (VGCHP) system. This optimization is realized by examining effects of parameters on the system. Thus, experimental studies have been conducted under varying depth of boreholes and temperatures of condenser and evaporator. Then, the Taguchi method is applied to these parameters to optimize coefﬁcient of performance (COP) of VGCHP system. Also, analysis of variance (ANOVA) and signal/noise (S/N) ratio are used for evaluation of experiment results. It is found that depth of borehole is the most signiﬁcant parameter, affecting the COP by 67.77%. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Usage of ground coupled heat pumps (GCHPs) increases day by day. Temperature of soil at a certain depth remains constant and that is the most important property, making this system attractive. The GCHPs gather heat from the soil, using a ground heat exchanger (GHE) whose pipes are installed in soil by different geometries. Heat, drawn through the soil, is transferred into a water-antifreeze solution inside a loop of pipe. Refrigerant passes through a compressor, raising it to a higher temperature, then heat water or air for heating system. The cooled GHE ﬂuid returns to the soil where it gathers further energy from, in a continuous cycle since heating is required. Typically, for a domestic home the loop is laid ﬂat or coiled in trenches about 2 m deep but if there is not enough space in your area, you can install a vertical GHE down into the soil to a depth of up to 100 m. Vertical GHEs are also used where the soil is too shallow for trenching, and they minimize the disturbance to existing landscaping. In literature, countless numerical, experimental and economical studies related to GCHP system are made [1–24]. In this study, we focus on vertical GHEs. Experimental studies constitute a major part of scientiﬁc researches. To reach precise results, experimental design must be well planned [25]. Observation, derivation and interpretation of any variation on the response variable, while acting such changes needed with respect to the input variables of

∗ Corresponding author. E-mail addresses: [email protected], [email protected]ﬁrat.edu.tr (H. Esen). http://dx.doi.org/10.1016/j.enbuild.2015.08.042 0378-7788/© 2015 Elsevier B.V. All rights reserved.

any process, can be deﬁned as experimental design [26]. In engineering area, experimental studies have an important role in new product design and improvement. According to Salmon (1990) and Taguchi (1986), suggested numerous approaches to experimental designs that are sometimes called “Taguchi methods”. These methods apply two-, three-, and mixed-level fractional factorial designs. Large screening designs seem to be principally chosen by Taguchi adherents. Since that is a way for guaranteeing better performance in the design period of products or processes, this method refers to experimental design as “off-line quality control”. Some experimental designs can be used on-line while the process is running, such as when used in evolutionary operation [27]. With regard to optimization of GCHP operating parameters, few research works have been published. Ramniwas et al. [28] applied Taguchi technique to optimize the operating parameters of a GCHP system for space heating application. They observed that the condenser outlet temperature plays an important role in controlling the value of COP of a GCHP system when operated in heating mode. Verma and Murugesan [29] examined the performance of a solar assisted GCHP system for heating application using Taguchi technique and utility concept. The design parameters are optimized to obtain the optimum solar collector area and ground heat exchanger length for space heating application with optimum COP. Sivasakthivel et al. [30] proposed a methodology to optimize the performance of a GCHP system for space heating and cooling applications using Taguchi method and utility concept. In the present work, we attempt to determine the most signiﬁcant parameter, affecting COP of a vertical GCHP system for heating mode. Condenser inlet–outlet, evaporator inlet–outlet

330

H. Esen, E. Turgut / Energy and Buildings 107 (2015) 329–334

temperatures and depth of boreholes are considered as inﬂuencing parameters of the GCHP system. Detailed discussions, on the Taguchi method implementation for heating mode alone followed by the use of utility concept to determine the ﬁnal optimum COP of the GCHP system, are made in following sections. 2. Experimental devices and measurements In this study, vertically drilled boreholes were implemented for three different depths as 30, 60, and 90 m for a room in Elazig, Turkey. The GCHP system overview is given in Fig. 1. In the experimental study, thermal resistance and conductivity tests of soil were done in selected boreholes (60 m and 90 m). The thermal resistance/conductivity of 60 and 90 mboreholes were calculated as 0.05 kW−1 m−1 /1.70 Wm−1 K−1 and 0.03 kW−1 m−1 /1.70 Wm−1 K−1 , respectively [5]. The single U-borehole GHEs were connected to a heat pump unit, sited in room. The heating load of the room was 1.72 kW at design conditions. The main components used in experimental studies can be summarized as follows; pipe distance: 60 mm; pipe diameter: 40 mm; borehole diameter: 150 mm; polyethylene SDR 11; single borehole. Heat exchanger: (air cooled condenser for heating, evaporator for cooling); manufacturer: Friterm; HS 10; 3.77 kW; heat transfer surface: 10 m2 . Water-antifreeze solution heat exchanger; TTE 3; 6.97 kW (with Freon 22). Compressor: Tecumseh Europe; hermetic; FH 5524 F; 7.6 m3 /h; 2900 tr/mn; 2 HP; suction line: 5/8 in, discharge line: 3/8 in. Dryer: Carly; DCY 083; 11.1–18.8 kW. Fan circuit: Friterm; 350 mm; 2350 m3 /h, 145 W. Observe glass: Carly VCYL 13; green and yellow; 3/8 in. In heating mode, the water-antifreeze ﬂuid transfers its heat to refrigerant ﬂuid in the evaporator. R-22 refrigerant, which ﬂows through other closed GHE in the heat pump, evaporates by gathering heat from the brine solution, ﬂowing through the evaporator, then passes through compressor. Ozone Depleting Substances (ODS) are chemicals that can damage the earth’s ozone layer if they escape into the upper atmosphere. ODS include chloroﬂuorocarbons (CFCs) such as R12 and R502, hydro chloroﬂuorocarbons (HCFCs) such as R22 and drop in blends such as R408A, R123 and R142b. HFCSs are still in use as refrigerants in many buildings in Refrigeration and Air Conditioning (RAC) equipment. A ban on the use of all refrigerants ODS for the maintenance or servicing of existing RAC and heat pump equipment is place since 1st January 2015 [31]. A condenser fan carries warmed air into room. The solution in the GHEs gathers heat from the soil and transfers to room. The horizontal distances between the boreholes are 3.4 (VB3–VB2) and 2 (VB1–VB2) m, respectively (see Fig. 2). The measurement dots of the soil section are also shown in Fig. 2. In this study, U-pipe GHEs were vertically buried in the marn type ground ( = 1.70 Wm−1 K−1 ). For all boreholes, as a grout

Fig. 2. The sketch of ground section.

material, bentonite is used. A number of T-type thermocouples were installed to measure all the circulating water-antifreeze solution and soil temperature. T type thermocouples with an accuracy 0.018 ◦ C, a metal vane type anemometer with ±2% accuracy and pressure manometer with ±0.3% (at ±25 ◦ C) accuracy were used to accomplish these experiments. The error analysis of the measurements are calculated ±2.89% for the brine and R-22 temperatures, ±2.75% for pressures, ±4.35% for power inputs to the compressor, condenser fan and circulating pump, and ±3.00% for electric currents. Error value is taken ±0.20% in reading values of the table. 3. Analysis and veriﬁcation of experimental study Following equation gives heat (Q˙ eva ), extracted from the soil in heating mode: ˙ wa Cp,wa (Twa,o − Twa,i ) Q˙ eva = m

(1)

˙ c ), the water-antifreeze By Eqs. (2)–(4), power of compressor (W ˙ p ) and the condenser fan (W ˙ cf ) are calculated, circulating pump (W respectively. ˙ c = Ic Uc Cos ϕ W

(2)

˙ p = Ip Up Cos ϕ W

(3)

˙ cf = Icf Ucf Cos ϕ W

(4)

COP of GCHP system is determined by, COP =

Q˙ hl ˙ p+W ˙ cf ˙ Wc + W

(5)

where heating load (Q˙ hl ) is calculated by, Fig. 1. Vertical GCHP system overview.

Q˙ hl = air V˙ air Cp,air (Tair,o − Tair,i )

(6)

H. Esen, E. Turgut / Energy and Buildings 107 (2015) 329–334

331

Fig. 3. Temperature variation of ambient and soil, at different depths, for heating period of six days.

Temperatures, obtained by experiments, are shown in Figs. 3 and 4 for heating mode. Fig. 3 shows the temperature variation of ambient and soil, at different depths, for heating period of six days. The numbers (1, 2, . . ., 8) in Fig. 3 refer dots in Fig. 2. At ﬁrst day of experimental study, the GCHP system operated when VB1 borehole was in usage. When the system started to operate, the Twa,o and Twa,i severely decreased. Second day, the system did not operate and third day worked when

VB2 borehole was in usage. On the fourth day, the system did not operate again. At last day of experimental study, the system operated when VB3 borehole was in usage. During the experiments minimum temperature of soil was 11 ◦ C (dot 1) in ﬁrst day. When VB2 borehole was in usage, two minimum temperature value of soil were measured 12 ◦ C (dot 2) and 13.43 ◦ C (dot 3), respectively. Finally, three minimum temperature value of soil were measured 13.15 ◦ C (dot 4), 14.22 ◦ C (dot 5) and

Fig. 4. Values of Tg-30m , Tg-60m , Tg-90m and COPs for ﬁve-month period.

332

H. Esen, E. Turgut / Energy and Buildings 107 (2015) 329–334

15 ◦ C (dot 6), respectively. Tg (dot 7 and dot 8) indicates surface temperature. Values of Tg-30m , Tg-60m , Tg-90m and COPs for ﬁve month period is given in Fig. 4 by monthly. In this period, minimum, average and maximum values of Tg-30m are 8.5, 13.35 and 20 ◦ C, respectively. For Tg-60m , these values are 12.2, 17.07 and 22 ◦ C, respectively. And ﬁnally for Tg-90m , these values are 17, 21.02 and 25 ◦ C, respectively. Again for this period, minimum, average and maximum values of COP-30 m are 0.98, 1.93 and 3.4, respectively. For COP-60 m, these values are 1.34, 2.37 and 3.69, respectively. And ﬁnally for COP90 m, these values are 1.77, 3.03 and 4.59, respectively. Depending on length of U-pipe, the maximum and minimum amount of heat, gathered from soil, is 81.9 and 45.6 Wm−1 . These values are reasonable according to Refs. [32,33]. 4. Optimization methodology 4.1. Taguchi technique Taguchi optimization is an experimental optimization technique that uses the standard orthogonal arrays for forming the matrix of experiments. By using this matrix it will help us to get maximum information from minimum number of experiments and also the best level of each parameter can be found. The main principle of the Taguchi method is given in Ref. [34]. In order to analyzing data, signal-to-noise (S/N) ratios are used. For analyzing S/N, three types of performance characteristics are used: lower the better (LB), higher the better (HB) and nominal the better (NB). To ﬁnd out the percentage contributions of each parameter, ANOVA is used. 4.2. Taguchi – design of experiments Experimental analysis is an important part of scientiﬁc research activities. In order to obtain more accurate results a well-organized experimental design is always needed. The earliest design of experiment methodology was developed by Sir Ronald Fisher (in about 1920s). Observation, derivation and interpretation of any variation on the target variable are the ways of the experimental design. All told ways need men, methods and machinery that can be controlled [25,26,35]. The major target of Taguchi method is to acquire data in a controlled way. Two important tools used in this method are S/N ratio, which evaluates quality giving point to variation, and orthogonal array, which accommodates several design factors at the same time [36–39]. In this approach, to calculate the deviation between the experimental and desired values, a loss function, can be transformed into an S/N ratio, is used. The signal factors are named as “controllable parameters” and considered by the researches. The noise factors are named as “uncontrollable factors” and can be deﬁned as exterior factors affecting outcome throughout the experiments but unable to attend the experimental design [40]. To obtain optimum design parameters, HB was selected for the COP. ij , can be calculated by the logarithmic value of S/N ratio of HB term, where ith performance characteristic examined at the jth experiment.

ij = −10 log

1 1 n Y2 n

i=1

(7)

i

In Eq. (7), Y represents the resulting value and n shows experiment number [41]. Conventional experimental design methods are generally complex and not easy to use. Those methods require large number of with increasing number of process parameters. However, the Taguchi method uses a special design of orthogonal arrays to study the entire parameter space with only a small number of

Table 1 Control factors used in the experiments. Factors

Symbol

Level 1

Level 2

Level 3

Borehole depths Condenser outlet temperature Condenser inlet temperature Evaporator inlet temperature Evaporator outlet temperature

A B C D E

30 38 33 10 7

60 40 34 14 9

90 41 35 17 10

Table 2 Experimental design using L27 orthogonal array. 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 26 27

Control factors

S/N ratio ()

A

B

C

D

E

COP

1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3

1 1 1 2 2 2 3 3 3 1 1 1 2 2 2 3 3 3 1 1 1 2 2 2 3 3 3

1 1 1 2 2 2 3 3 3 2 2 2 3 3 3 1 1 1 3 3 3 1 1 1 2 2 2

1 1 1 2 2 2 3 3 3 3 3 3 1 1 1 2 2 2 2 2 2 3 3 3 1 1 1

1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3

6.6083 6.0639 5.8007 7.8187 7.4214 7.1205 8.5950 8.2324 7.8539 9.2480 8.7550 8.5627 8.1308 7.8539 7.6042 9.2180 8.6594 7.6763 9.5713 9.3964 8.9121 10.2377 9.8552 9.2480 11.4109 10.9061 10.6805

experiments [42]. By orthogonal arrays it is aimed to estimate the effects of several factors and interactions by minimizing the number of experiments. For each experiment S/N ratios are computed. The S/N ratio is the measurement of the functionality of the system, which exploits interaction between control and noise factors [33]. 4.3. Optimization results and discussions In this part, we focus on results of design method. Control factors, obtained by experimental data, are presented in Table 1. By converting those results into the S/N ratio, evaluation of experimental results is achieved. Experimental design method is based on orthogonal arrays to adjust experiment plan. The most suitable Taguchi orthogonal array and S/N ratios are considered to be L27 as shown in Table 2. The best optimization performance value is acquired at the highest S/N ratio regardless its type. For each experiment, average S/N ratios are presented in Table 3. Table 3 Average S/N ratios. Factors

Symbol

Level 1

Level 2

Level 3

Borehole depths Condenser outlet temperature Condenser inlet temperature Evaporator inlet temperature Evaporator outlet temperature

A B C D E

7.28 8.10 8.15 8.34 8.98a

8.41 8.37 9.10a 8.42 8.57

10.02a 9.25a 8.46 8.95a 8.16

a

Optimum level. Overall mean = 8.57 Db.

H. Esen, E. Turgut / Energy and Buildings 107 (2015) 329–334

333

Table 5 Results of the conﬁrmation experiment. Initial parameters

Level COPa S/N ratio (dB) a

Fig. 5. The effect of design parameters on COP.

As an outcome of such analyses, the optimum parameters are determined. Those S/N values are displayed in Fig. 5. As previously told, the largest S/N ratios at all levels of the parameters provide optimum performance. Optimum parameters, regarding this experimental design, are 10.02 (A – level 3), 9.25 (B – level 3), 9.10 (C – level 2), 8.95 (D – level 3), 8.98 (E – level 1). Observed variance in a particular variable is partitioned into components attributable to different sources of variation, in the ANOVA procedure. The governing equations of the ANOVA are as follows [42].

2 i

SSm =

SSfactor = SST =

(8)

n

2factor-i N

− SSm

(9)

2i − SSm

SSe = SST −

(10)

SSA

(11)

dftotal = n − 1

(12)

dffactor = k − 1

(13)

Vfactor = Ffactor

SSfactor dffactor

A1B1C2D3E1 2.27 7.12

Optimum parameters Prediction

Experiment

A3B3C2D3E1 3.98 12.02

A3B3C2D3E1 3.75 11.48

Improvement of S/N ratio for COP = 4.36 dB.

ﬁtted to a data set, in order to identify the model that best ﬁts the population from which the data were sampled. The calculated F values are compared to suitable standard conﬁdence tables. If the calculated F values are higher from the table values, the parameters have important impact on COP characteristics. Contribution rate is another value used in method assessment. This value indicates the rate of impact of any factor on experiment result, between others. As a result of ANOVA, the forceful degree on the COP characteristics of parameters is obtained by calculating the percent contribution of design parameters. Results of the ANOVA are presented in Table 4. Table 4 has shown that the depth is the most signiﬁcant parameter affects the COP (67.77%). The contributions of the parameters such as: condenser outlet–inlet temperature, evaporator inlet–outlet temperature are found to be; 12.74%, 8.28%, 3.86%, and 5.89%, respectively. According to the related literature results were highly consistent. Some researches applied Taguchi technique to optimize the operating parameters of GCHP system for space heating and cooling applications [28,30]. Sivasakthivel et al. [30] have presented the percentage contribution of each parameter on the performance of GCHP, in that condenser outlet/inlet temperatures, evaporator outlet/inlet temperatures are observed to play a major role. Ramniwas et al. [28] applied Taguchi technique to optimize the operating parameters of a GCHP system for space heating applications. They observed that the condenser outlet temperature plays an important role in controlling the value of COP of a GCHP when operated in heating mode. S/N ratio is calculated as:

(14)

V = factor Verror

(15)

ˆ = m +

k

(¯ i − m )

(16)

i=1

where SST is total square sums, SSm is mean value of square sums, SSfactor is factor value of square sums, SSe is error value of square sums, factor-i the sum of ith level of factor, N is number of factor levels, df is degrees of freedom, n the number of the experiments, k is factor’s level number, Vfactor is the variance of factor and Ffactor is the F-test value of the factor. To state parameters which have signiﬁcant effect on COP, an Ftest is used. In case, the F-test is any statistical test in which the test statistic has an F-distribution under the null hypothesis. It is most often used when comparing statistical models that have been

where m is total mean S/N ratio, ¯ i is mean S/N ratio at the optimal level, and k main design parameter number. The results of conﬁrmation experiment conducted with the optimum design parameters are presented in Table 5. As a result of comparison of initial and optimum parameters obtained from the Taguchi approach, COP increases up to 1.65 times and S/N ratio improvement for COP is 4.36 dB. It can be concluded that conﬁrmation experiments data approved the validity of the Taguchi approach.

Table 4 Results of ANOVA. Factor A B C D E Error Total a

Degree of freedom (df) 2 2 2 2 2 16

At least 99% conﬁdence level.

Sum of square (SS) 34.2482 6.4847 4.2331 2.0031 3.0255 0.4517 50.4463

Variance (V) 17.1241 3.2424 2.1165 1.0016 1.5128 0.02823

F-test a

606.59 114.86a 74.97a 35.48a 53.59a

F

Contribution (%)

6.22 6.22 6.22 6.22 6.22

67.77 12.74 8.28 3.86 5.89 1.46 100

334

H. Esen, E. Turgut / Energy and Buildings 107 (2015) 329–334

5. Conclusions This study can be ﬁnalized in two parts. First one is experimental analysis and second one is optimization with Taguchi method. In the experimental part of study, the average value of COP-30 m, COP60 m and COP-90 m is calculated 1.93, 2.37 and 3.03, respectively over the ﬁve-month period. In the optimization part of study, the effects of design parameters of a GCHP on the COP are investigated. Importance of design parameters on the COP is determined by using ANOVA. The optimum parameter combination for the maximum COP is obtained by using the analysis of S/N ratio. Finally, the validity of Taguchi experimental design method, used in this study, is checked by conﬁrmation experiments. Achieved results might be summarized as follows: • Using the Taguchi optimization analysis it is found that depth of vertical boreholes are the most inﬂuencing parameters for achieving the maximum COP during heating period. The most signiﬁcant parameter on the COP is found to be borehole depth with the level of 67.77%. The contributions of the parameters such as: condenser outlet–inlet temperature, evaporator inlet–outlet temperature are found to be 12.74%, 8.28%, 3.86%, and 5.89%, respectively. • Optimum parameters levels for maximum COP determined as A3B3C2D3E1. • All design parameters are found to be 99% conﬁdence level. • The ﬁnal step of the study called experimental veriﬁcation test methods are used. Table 5 indicates the comparative test results for COP. From the conﬁrmation test results, average 5.778% error was obtained for COP between predicted values by the regression model and conﬁrmation test results. Acknowledgements The authors gratefully acknowledge the ﬁnancial support from the Scientiﬁc Research Projects Management Council of the Fırat University for this study performed under project with grant no. 2005/1153 and the Scientiﬁc and Technological Research Council of Turkey (TUBITAK) for its ﬁnancial support through contract no. 106Y188 (2006–2007). References [1] Y. Shang, M. Dong, S. Li, Intermittent experimental study of a vertical ground source heat pump system, Appl. Energy 136 (2014) 628–635. [2] Y. Bi, X. Wang, Y. Liu, H. Zhang, L. Chen, Comprehensive exergy analysis of a ground-source heat pump system for both building heating and cooling modes, Appl. Energy 86 (2009) 2560–2565. [3] M. Inalli, H. Esen, Experimental thermal performance evaluation of a horizontal ground-source heat pump system, Appl. Therm. Eng. 24 (2004) 2219–2232. [4] A. Hepbasli, O. Akdemir, Energy and exergy analysis of a ground source (geothermal) heat pump system, Energy Convers. Manag. 45 (2004) 737–753. [5] H. Esen, M. Inallı, Thermal response of ground for different depths on a vertical ground source heat pump system in Elazı˘g, Turkey, J. Energy Inst. 82 (2009) 95–101. [6] A. Koyun, H. Demir, Z. Torun, Experimental study of heat transfer of buried ﬁnned pipe for ground source heat pump applications, Int. Commun. Heat Mass Transf. 36 (2009) 739–743. [7] S.P. Lohani, D. Schmidt, Comparison of energy and exergy analysis of fossil plant, ground and air source heat pump building heating system, Renew. Energy 35 (2010) 1275–1282. [8] O. Ozyurt, D.A. Ekinci, Experimental study of vertical ground-source heat pump performance evaluation for cold climate in Turkey, Appl. Energy 88 (2011) 1257–1265. [9] W. Yang, Experimental performance analysis of a direct-expansion ground source heat pump in Xiangtan, China, Energy 59 (2013) 334–339. [10] H. Wang, Q. Zhao, J. Wu, B. Yang, Z. Chen, Experimental investigation on the operation performance of a direct expansion ground source heat pump system for space heating, Energy Build. 61 (2013) 349–355.

[11] M.R. Ally, J.D. Munk, V.D. Baxter, A.C. Gehl, Energy and exergy analysis of a ground-source heat pump for domestic water heating under simulated occupancy conditions, Int. J. Refrig. 36 (2013) 1417–1430. [12] C. Montagud, J.M. Corberan, F. Ruiz-Calvo, Experimental and modelling analysis of a ground source heat pump system, Appl. Energy 109 (2013) 328–336. [13] I. Sarbu, C. Sebarchievici, General review of a ground-source heat pump systems for heating and cooling of buildings, Energy Build. 70 (2014) 441–454. [14] D.D. Col, M. Azzolin, G. Benassi, M. Mantovan, Experimental analysis of optimal operation mode of a ground source heat pump system, Energy Procedia 45 (2014) 1354–1363. [15] H. Esen, M. Inalli, A. Sengur, M. Esen, Artiﬁcial neural networks and adaptive neuro-fuzzy assessments for ground-coupled heat pump system, Energy Build. 40 (2008) 1074–1083. [16] H. Esen, M. Inalli, A. Sengur, M. Esen, Modelling a ground-coupled heat pump system by a support vector machines, Renew. Energy 33 (2008) 1814–1823. [17] D.P. Jenkins, R. Tucker, R. Rawlings, Modelling the carbon-saving performance of domestic ground-source heat pumps, Energy Build. 41 (2009) 587–595. [18] W. Gang, J. Wang, Predictive ANN models of ground heat exchanger for the control of hybrid ground source heat pump systems, Appl. Energy 112 (2013) 1146–1153. [19] B. Cai, Y. Liu, Q. Fan, Y. Zhang, Z. Liu, S. Yu, R. Ji, Multi-source information fusion based fault diagnosis of ground-source heat pump using Bayesian network, Appl. Energy 114 (2014) 1–9. ˜ [20] J.M. Belman-Flores, S.E. Ledesma, M.G. Garcia, J. Ruiz, J.L. Rodríguez-Munoz, Analysis of a variable speed vapor compression system using artiﬁcial neural networks, Exp. Sys. Appl. 40 (2013) 4362–4369. [21] H. Sayyadi, M. Nejatolahi, Thermodynamic and thermoeconomic optimization of a cooling tower-assisted ground source heat pump, Geothermics 40 (2011) 221–232. [22] A. Balbay, M. Esen, Experimental investigation of using ground source heat pump system for snow melting on pavements and bridge decks, Sci. Res. Essays 5 (24) (2010) 3955–3966. [23] M. Esen, T. Yuksel, Experimental evaluation of using various renewable energy sources for heating a greenhouse, Energy Build. 65 (2013) 340–351. [24] A. Balbay, M. Esen, Temperature distributions in pavement and bridge slabs heated by using vertical ground-source heat pump systems, Acta Sci. Technol. 35 (4) (2013) 677–685. [25] J. Antony, Design of Experiments for Engineers and Scientists, Elsevier Science and Technology Books, 2003, October. [26] D.H. Besterﬁeld, C. Besterﬁeld, G.H. Besterﬁeld, M. Besterﬁeld, Total Quality Management, Prentice Hall Inc., New Jersey, 1995. [27] W. Jankrajang, Design of Experiment Approach for Improving Rice Milling Quality (Master thesis), Kasetsart University, 2003. [28] K. Ramniwas, K. Murugesan, P.K. Sahoo, Optimization of operating parameters of ground source heat pump using Taguchi method, in: 23rd IIR Conference, Prague, Czech Republic, August 21–26, 2011. [29] V. Verma, K. Murugesan, Optimization of solar assisted ground source heat pump system for space heating application by Taguchi method and utility concept, Energy Build. 82 (2014) 296–309. [30] T. Sivasakthivel, K. Murugesan, H.R. Thomas, Optimization of operating parameters of ground source heat pump system for space heating and cooling by Taguchi method and utility concept, Appl. Energy 116 (2014) 76–85. [31] Environmental Protection Agency, Complying with Regulations Controlling Fluorinated Greenhouse Gases and Ozone Depleting Substances, A Guidance Note for Contractors of Equipment Containing F-gas and ODS, July 2015, Available from: http://www.epa.ie/pubs/advice/air/ods/3%20IRL% 20Guidance%20Note%20Operators%20ODS%20fgas%20Complying%20with %20the%20Regulations%20v1.0.pdf (cited 2015, 13 August 2015). [32] S. Kavanaugh, Field tests for ground thermal properties – methods and impact on ground-source heat pump design, ASHRAE Trans. 106 (2000) 851–855. [33] S. Kavanaugh, C. Martin, Ground thermal conductivity testing controlled site analysis, ASHRAE Trans. 108 (2002) 945–952. [34] G. Taguchi, R. Jugulum, The Mahalanobis-Taguchi Strategy, A Pattern Technology System, John Wiley & Sons, New York, 2002. [35] R.K. Roy, Design of Experiments Using the Taguchi Approach: 16 Steps to Product and Process Improvement, John Wiley & Sons, New York, 2001. [36] D.C. Montgomery, Design and Analysis of Experiments, ﬁfth ed., John Wiley & Sons, New York, 1997. [37] P.M. George, N. Pillai, N. Shah, Optimization of shot peening parameters using Taguchi technique, J. Mater. Process Technol. 153–154 (2004) 925–930. [38] N. Celik, E. Turgut, Design analysis of an experimental jet impingement study by using Taguchi method, Heat Mass Transf. 48 (2012) 1407–1413. [39] M.S. Phadke, Quality Engineering Using Robust Design, Prentice Hall, New Jersey, 1989. [40] E. Turgut, G. C¸akmak, C. Yıldız, Optimization of the concentric heat exchanger with injector turbulators by Taguchi method, Energy Convers. Manag. 53 (2012) 268–275. [41] R.K. Roy, A Primer on the Taguchi Method, Competitive Manufacturing Series, New York, 1990. [42] P.J. Ross, Taguchi Techniques for Quality Engineering, second ed., McGraw-Hill, New York, 1996.