Energy and Buildings 84 (2014) 258–267

Contents lists available at ScienceDirect

Energy and Buildings

journal homepage: www.elsevier.com/locate/enbuild

The artificial neural network model to estimate the photovoltaic

modul efficiency for all regions of the Turkey
İlhan Ceylan a , Engin Gedik a,∗ , Okan Erkaymaz b , Ali Etem Gürel c
a
Faculty of Technology, Karabük University, Karabük, Turkey
b
Engineering Faculty, BülentEcevit University, Zonguldak, Turkey
c
Düzce Vocational High School, Düzce University, Düzce, Turkey

a r t i c l e i n f o a b s t r a c t

Article history: Artificial neural network (ANN) is a useful tool that using estimates behavior of the most of engineering
Received 2 May 2014 applications. In the present study, ANN model has been used to estimate the temperature, efficiency and
Received in revised form 31 July 2014 power of the Photovoltaic module according to outlet air temperature and solar radiation. An experi-
Accepted 3 August 2014
mental system consisted photovoltaic module, heating and cooling sub systems, proportional integral
Available online 12 August 2014
derivative (PID) control unit was designed and built. Tests were realized at the outdoors for the constant
ambient air temperatures of photovoltaic module. To preserve ambient air temperature at the deter-
Keywords:
mined constant values as 10, 20, 30 and 40 ◦ C, cooling and heating subsystems which connected PID
Solar energy
Photovoltaic
control unit were used in the test apparatus. Ambient air temperature, solar radiation, back surface of
Artificial neural network the photovoltaic module temperature was measured in the experiments. Obtained data were used to
estimate the photovoltaic module temperature, efficiency and power with using ANN approach for all 7
region of the Turkey. The study dealing with this paper not only will beneficial for the limited region but
also in all region of Turkey which will be thought established of photovoltaic panels by the manufacturer,
researchers and etc.
© 2014 Elsevier B.V. All rights reserved.

1. Introduction the research program of State Planning Organization (DPT) and

The Scientific & Research Council (TUBITAK)” [4]. As it like all the
Demand for energy determining to economic, political and process PV systems efficiency are desired to be increased with
social events is increasing day by day in the world. Countries should done research and developments which efficiencies are between
have to take precautions for use their existing energy sources 15 and 20% now. For this purpose theoretical and experimental
efficiently or search new energy sources. As it known well the studies [5–9] have been carried out for increasing of PV systems
decreasing amount of fossil based energy sources and its haz- efficiency which depends on insolation, atmospheric and meteo-
ardous effect to the environment, the renewable energy sources rological conditions. “Module temperature is a parameter which
make themselves attractive for research by the man. Solar energy has great influence in the behavior of a PV system, as it modi-
is the important energy source placed among the renewable energy fies system efficiency and output energy” [10]. Hanlin and Stein
sources with its endless heat and light supply. Thermal and produc- [11] have modeled the module temperature of PV system using
tion of electricity are the basic applications of the solar energy. a transient heat-flow model. Single day of measured module tem-
Photovoltaic (PV) systems are commonly used all around world perature has used simultaneous non-linear least squares regression
as shown themselves to be one of the most promising applica- and optimized then tested for accuracy using a year’s worth of data
tions for dealing with electricity generation [1–3]. “In countries for one location. Environmental conditions on module tempera-
like Turkey, photovoltaic research and development activities are ture of selected PV system in Singapore have analyzed by the Ye
still mainly undertaken across a range of universities, government et al. [12]. Ceylan et al. [13] experimentally analyzed different PV/T
and industry facilities and the projects are mainly financed by systems for the cooling photovoltaic modules. A simple pipe was
placed on PV module as a spiral heat exchanger in order to provide
active cooling. Also, the system can easily be applied to large-scale
∗ Corresponding author. Tel.: +90 3704338200. systems.
E-mail addresses: ilhancey@gmail.com (İ. Ceylan), egedik@karabuk.edu.tr, Numerical or algebraic methods are the way to estimate the
gedik engin@hotmail.com (E. Gedik). behavior of PV module under natural sunlight [14]. An artificial

http://dx.doi.org/10.1016/j.enbuild.2014.08.003
0378-7788/© 2014 Elsevier B.V. All rights reserved.
 İ. Ceylan et al. / Energy and Buildings 84 (2014) 258–267 259

Fig. 1. The experimental system.

neural network (ANN) is a tool which is widely using to estimate to estimate, under real conditions, efficiency and the maximum
the efficiency and maximum power of PV systems [15,16]. Almaktar power of commercially available a multi crystalline silicon (mc-Si)
et al. [17] presented an ANN based approach for predicting photo- photovoltaic module by using the outside temperature and solar
voltaic module temperature using meteorological variables. Ravaee radiation. For this purpose an experimental setup was designed
et al. [18] presented a new application of ANN for modeling a Photo- and built. Tests were made outdoors and measured ambient air
voltaic Thermal collector (PV/T). Ambient temperature of collector, temperature, solar radiation and back surface of module temper-
cell and fluid temperature at duct inlet, fluid velocity in duct, solar ature data were used the training of ANN, estimate the efficiency
intensity and time were used in the input layer while the thermal and the module temperature of PV for the all 7 region of Turkey.
efficiency and electrical efficiency are outputs. Vasarevicius et al.
[19] presented analysis of IncCondMPPT algorithm and comparison 2. Methodology
of operation with and without ANN. Bahgat et al. [20] presented a
development and implementation of a PC-based maximum power “Turkey has a technical power generating capacity of 4–5 MW
point tracker (MPPT) for PV system using neural networks. Had- from PV applications. Currently, the most PV applications in Turkey
jab et al. [21] presented the results of the characterization and are used for stand-alone power systems. Depending on the devel-
modeling of the electrical current-voltage and power-voltage of opments about the price and efficiency of PV appliances, the Turkish
the photovoltaic panel BP 3160 W, using a new approach based PV market is expected to rapidly expand given the fact that there
on artificial intelligence. Reddy et al. [22] presented an application are more than 34,000 small residential areas including resort areas
of a neural network for the identification of the optimal operating along the coast lines in Turkey where solar-powered electricity
point of PV module maximum power tracking control. Kulaksız and would be more economical than grid supply [28]”. For this reason
Akkaya [23] used a genetic algorithm for improve the maximum the details given in this work will be able to much beneficial for the
power point tracking efficiency of a PV system with introduction PV market due to it composed all region of Turkey which has great
motor drive by optimizing the input dataset for an ANN model of solar energy potential.
PV modules. Karatepe et al. [24] presented a neural network based An experimental system was established as can be shown in
approach for improving the accuracy of the electrical equivalent Fig. 1. The PV module was to position indoor environment with
circuit of a photovoltaic module. Tajuddin et al. [25] presented a glass material of the test apparatus. Indoor environment temper-
maximum power point tracking (MPPT) technique for photovoltaic ature of the PV module was tried to fix as the values of 10, 20,
(PV) system using a modified differential evolution (DE) algorithm. 30 and 40 ◦ C. To obtain constant temperature values in the indoor
The standard DE is modified to deal with dynamic objective func- environment both heating and cooling systems were used. While
tion problem to suit with the nonlinear time-varying MPPT nature. electrical heater was using to increase the temperature, cooling
Mellit et al. [26] described a methodology to estimate the profile system based on basic vapor compression refrigeration cycle was
of the produced power of a 50 W p Si-polycrystalline photovoltaic used to decrease the temperature in the indoor environment. It
module. For this purpose, two ANNs have been developed for use has been possible to hold desired constant temperatures with
in cloudy and sunny days, respectively. In our previous work [27] using PID control equipment. Solar radiation intensity incoming
we have investigate the PV module efficiency experimentally and on the PV module numbered as 12 in Fig. 1 was measured with
we used the ANN module to estimate the module temperature for solar meter. Indoor environment temperature numbered as 13 was
the Aegean region of Turkey and the extended version of the study adjusted at desired temperature values with using number 2 and
is presented here which aimed to propose a model based on ANNs cooling equipment numbered as 4-5-6-7-8-9 in this figure. When
260 İ. Ceylan et al. / Energy and Buildings 84 (2014) 258–267

Fig. 2. Neural network structure.

Fig. 3. Comparison of ANN and experimental data.

temperature of the indoor environment decreases under the set

value number 1 (automatic control equipment) activates the num- is taken, 0.0035 K−1 is taken for CIS, 0.0025 K−1 for CdTe and
ber 2 to get the desired temperature. Similar case has occurred for 0.002 K−1 for a-Si.
reverse namely, when temperature of the indoor environment rises In addition, the electrical efficiency of PV module is given as
above the set value, number 10 activates the cooling system to get follows:
the desired temperature. This process is work automatically in the
experimental setup. Digital thermostats were used to activate the m = c × g × ˛c × ıc (2.2)
heating and cooling systems. In the experiments solar radiation, where  g the transparency for the PV module glass is, ˛c is the
ambient air temperature of the PV module, back surface temper- absorptive of the solar cell and ıc is the packing factor; the values
ature of the PV module was measured and recorded. Experiments for these are taken as 0.90, 0.95 and 0.90, respectively.
were realized under the outdoors weather conditions during the Another expression of the module efficiency could be written as
day hours for 5 days at the Department of Energy Systems Engi- follows;
neering in Karabük University place at Black sea region of Turkey.
Obtained results from the experiments were used the training of P
m = (2.3)
ANN. Am × I (t)
A model based on ANN was examined to estimate the real test The output power P of the PV module is calculated using the
conditions, module temperature and efficiency of the PV module measured voltage and current values as follows.
for the all 7 region of Turkey. Artificial neural networks have been
widely used in many application areas of engineering. In the most of P =V ×I (2.4)
the applications, neural networks with the back-propagation (BP)
training algorithm are used. “Back propagation algorithm”, which The electrical energy gain obtained from the PV module can be
optimizes the weighted connections by allowing the error to spread calculated as follows:
.
from output layer toward the lower layers, was used as the train- E l,net = m × Am × I (t) (2.5)
electrical
ing system in training networks. Three layers feed-forward neural
network was used in the study and Levenberg–Marquardt back- where m is the module efficiency and Am is the module surface
propagation method was selected as learning algorithm. Proposed area [23].
three layers network structure was shown in Fig. 2. In this struc-
ture, measured solar radiation (I) and controlled module ambient 3. Results and discussion
air temperatures (Ta ) were used as two input variables. Photovoltaic
module back side temperatures (Tp ) were used as the output vari- In the present paper, measured solar radiation (I) and controlled
able. The number of neuron in hidden layer was defined as 12 with PV module ambient air temperatures (Ta ) were used as two input
trials. For details of the proposed ANN model of this study please variables while PV module back side temperatures (Tp ) was used
see Ref. [27]. as the output variable for ANN structure. The number of neuron in
Efficiency of the PV module has calculated with using measured hidden layer was defined as 12 with trials and 228 samples dataset
data for the region that realized test and with using estimated data were used in MATLAB neural networks tool in order to train and
from the ANN for the other regions. The panel temperature was esti- estimate. The estimated ANN model and experimental data were
mated with ANN in MATLAB software and then generating power compared each other as shown in Fig. 3.
and efficiency of PV module calculated using with equations given After test and training process of experimental values, photo-
below as (2.1)–(2.5). voltaic module temperatures were estimated with ANN depending
Electrical efficiency is calculated as follows: on outlet air temperature and solar radiation of the 7 region and 81
  cities in Turkey. The used outlet air temperature and solar radiation
c = 0 1 − ˇ (Tc − 25) (2.1)
for each region can be seen in Tables 1 and 2. Estimated photovoltaic
where 0 is the efficiency at standard test conditions module temperatures using ANN were also shown in Table 3. Using
(I(t) = 1000 W/m2 , Tc = 25 ◦ C), Tc is the solar cell temperature these estimated values, electrical efficiency change computed from
and ␤ is the electrical efficiency thermal coefficient. Eqs. (2.1) and (2.3) has shown in Fig. 4 for all region of Turkey. Pho-
Value of ˇ is dependent on the features of the materials from tovoltaic power for per square meter for each city was also shown
which PV module is produced. For crystal silicon almost 0.0045 K−1 in same figure.
 İ. Ceylan et al. / Energy and Buildings 84 (2014) 258–267 261

Table 1
Monthly average daily solar radiation values of the regions of Turkey (W/m2 ).

Mediterranean region of Turkey

Jan. Feb. March April May June July August Sep. Oct. Nov. Dec.

Antalya 428.28 421.31 603.59 659.83 655.67 600 561.65 543.84 526.53 511.72 420.43 421.98
Burdur 451.48 439.86 608.88 675.03 659.73 601.75 568.78 541.33 523.52 515.44 431.82 439.72
Isparta 456.62 470.7 615.84 679.54 668.79 611.71 580.34 545.37 528 521.62 435.66 455.7
Mersin 422.85 438.74 580.95 625.3 631.79 613.6 581.66 551.22 502.99 485.46 401.63 411.64
Adana 423.98 428.32 591.1 635.2 624.49 591.67 548.85 526.74 482.76 485.86 397.61 429.93
Hatay 390.96 389.07 559.27 588.16 582.6 595.15 579.43 555.87 484.69 461.83 368.92 358.72
Osmaniye 428.88 415.19 605.03 628.97 611.39 585.6 543.68 525.47 478.82 488.43 395.27 426.89
K.maraş 472.68 471.66 630.86 648.41 657.26 592.69 560.89 524.93 499.51 500.66 431.65 468.91
East Anatolia region of Turkey
Erzincan 477.21 523.71 642.28 707.28 738.12 655 623.24 581.62 543.86 532.21 452.23 474
Elazığ 443.09 480.54 643.66 677.25 674.12 597.38 562.86 528.68 513.18 522.41 449.22 480.34
Tunceli 452.74 507.01 648 704.26 707.78 626.84 584.43 552.49 532.14 528.47 459.02 478
Bingöl 448.53 529.41 671.1 714.48 708.15 623.07 590.79 552.9 531.45 532.75 462.17 478.26
Erzurum 400 484.08 589.34 667.63 696.86 628.8 579.61 547.93 514.12 485.58 400.44 398.79
Muş 494.54 543.75 721.22 706.21 702.08 630.66 598.07 557.62 531.25 535.56 453.81 478.39
Bitlis 540.46 572.03 735.61 722.3 696.87 600 618.04 565.91 525.99 540.82 447.01 481.89
Kars 413.61 628.69 568.91 693.18 786.08 658.24 630.33 562.74 593.86 526.32 406.64 386.63
Ağrı 424.39 493.55 616 667.98 704.85 625.69 599.82 554.22 535.45 502.8 396.68 383.84
Ardahan 386.67 677.49 567.39 682.54 867.98 728.31 751.41 642.06 643.93 538.98 388.89 359.65
Van 362.43 443.75 560.22 603.53 655.79 605.19 614.59 551.5 509.21 475.82 375 350.91
Iğdır 270.41 326.97 430.6 506.22 563.02 529.98 517.08 494.42 464.39 405.76 291.54 238.86
Hakkâri 296.24 342.72 473.09 505.53 574.71 538.95 533.7 503.8 484.11 414.27 312.67 278.82
Malatya 444.44 475.47 628.22 652.67 675.88 598.42 564.1 525.35 511.04 512.36 442.97 478.02
Southeast Anatolia region of Turkey
Gaziantep 439.13 430.8 605.57 622.22 618.33 582.97 558.77 534.78 496.51 483.33 401.34 410.96
Kilis 442.55 430.74 602.94 614.53 600.98 588.85 567.8 542.36 502.55 470.74 390.88 395.2
Adıyaman 432.72 457.19 617.21 633.66 642.27 578.95 590.22 518.23 497.54 502.65 431.65 448.88
Şanlıurfa 414.53 441.28 591.04 624.08 620.48 558 529.79 509.43 496.54 491.57 412.27 409.09
Diyarbakır 453.08 505.11 608.77 657.42 684.51 621.94 571.43 450.72 532.28 503.09 416.49 438.81
Mardin 439.08 462.38 603.86 643.04 622.38 545.53 515.58 492.1 500.5 500.66 413.8 406.32
Batman 459.18 490.04 655.84 663.61 640.58 578.5 547.56 521.16 499.5 517.05 427 451.53
Şırnak 451.39 476.45 622.56 672.8 637.28 559.41 541.87 508.14 515.15 506.78 416.38 402.22
Siirt 492.19 508 677.15 691.06 650.41 588.54 569.61 535.68 507.51 527.12 428.57 445.27
Central Anatolia region of Turkey
Eskişehir 538.7 491.56 855.83 781.55 686.79 627.95 586.98 562.38 527.68 520 418.6 436.2
Konya 472.55 464.61 614.83 647.57 665.96 601.06 568.92 532.57 522.98 507.48 426.76 450.38
Ankara 453.08 505.11 608.77 657.42 686.07 621.94 571.43 546.41 532.28 503.09 416.49 438.81
Çankırı 433.6 560 572.11 644.07 719.47 630.3 585.98 549.06 551.64 494.21 407.16 423.68
Aksaray 462.29 459.26 607.35 631.12 664.53 591.84 557.76 520.38 512.74 497.28 420.18 458.44
Kırıkkale 443.27 476.95 566.21 627 687.07 611.43 559.53 535.28 529.41 496.19 409.66 444.11
Kırşehir 448.36 459.45 588.5 625.65 675.17 600.19 547.31 524.66 519.31 491.4 410.26 450.14
Yozgat 463.16 474.61 622.95 657.06 698.39 606.6 561.79 527.47 523.08 484.67 406.81 465.46
Niğde 460.32 483.58 627.6 650.95 669.82 600.53 569.08 523.77 506.96 494.09 423 464.1
Nevşehir 459.46 466.16 626.75 647.06 676.86 591.76 556.85 517.42 511.36 489.68 414.95 470.59
Kayseri 473.04 487.76 643.64 664.03 687.3 600.71 570.24 521.36 511.18 489.85 423 487.32
Karaman 484.3 459.83 613.44 633.89 655.49 606.43 574.87 536.18 516.32 368.08 423.53 443.15
Sivas 476.06 511.95 651.08 686.11 720.5 630.48 596.93 551.17 528.71 500 422.53 484.85
Black sea region of Turkey
Bolu 472.56 594.1 678.23 698.31 714.8 638.23 596.74 560.75 547.24 512.11 415.49 419.87
Düzce 435.33 541.18 636.54 689.17 691.84 617.92 568.27 546.02 527.64 510.71 414.39 404.61
Zonguldak 431.19 505.88 601.49 643.5 688.89 600.2 547.69 531.15 519.13 500 402.98 390.88
Karabük 436.95 536.36 603.2 651.85 726.47 622.31 573.72 545.27 547.19 505.36 409.52 410.6
Bartın 413.9 504.65 606.8 632.13 714.47 593.56 540.31 521.13 527.49 500.92 389.71 402.03
Kastamonu 427.73 542.79 614.78 654.13 737.91 622.2 573.17 541.03 555.41 594.62 392.02 400
Çorum 441.67 532.36 596.28 649.5 708.08 626.64 578.8 541.34 538.46 489.5 398.25 429.91
Sinop 410.4 513.57 606.34 638.97 723.08 636.65 589.29 561.16 545.57 487.41 377.31 383.18
Samsun 416.67 505.67 628.63 667.18 707.07 654.65 616.6 587.51 542.71 493 388.89 394.41
Amasya 420.17 539.78 629.56 678.79 708.33 640.92 598.62 558.76 540.07 486.09 394.68 426.33
Tokat 452.78 540.25 645.45 687.03 714.63 641.68 612.65 564.86 533.33 666.67 402.6 443.75
Ordu 447.67 547.72 708.59 717.95 727.87 675.5 678.25 630.23 554.4 510.34 405.53 424.44
Giresun 474.16 572.38 687.38 711.89 743.78 700 705.75 650 565.11 527.49 416.48 442.62
Gümüşhane 526.49 606.52 669.67 724.03 763.77 705.95 705.38 645.16 570.05 539.5 433.41 464.88
Trabzon 451.1 552.45 700.82 748.38 764.4 738.92 772.15 725.72 585.46 464.58 411.9 418.06
Bayburt 475.92 626.35 669.63 738.09 767.64 701.7 700.74 643.33 577.06 548.92 441.7 440.63
Rize 396.45 560.57 656.43 726.56 791.16 758.24 801.29 741.38 616.13 573.56 414.87 403.97
Artvin 409.74 598.06 604.32 701.97 827.24 762.96 816.99 731.67 653.06 574.35 414.52 382.17
Marmara region of Turkey
Çanakkale 358.67 396.36 546.18 575.29 605.87 535.47 498.73 497.73 471.11 433.19 363.08 334.2
Balıkesir 397.53 425.74 597.97 654.27 631.64 572.34 525.35 519.25 500 486.98 411.51 381.87
Edirne 366.41 419.96 535.35 553.97 588.05 538.32 491.9 500.47 465.09 452.74 350.88 350.15
Tekirdağ 406.5 451.61 581.23 614.32 619.89 532.56 486.58 492.15 497.6 498.23 407.23 383.39
Kırklareli 391.3 436.15 583.79 588.87 630.46 513.51 464.8 473.73 500 505.64 381.52 407.67
İstanbul 578.03 580.13 789.47 770.8 731.71 646.05 607.88 598.62 650.06 716.47 615.58 608.1
Bursa 415.09 505.54 608.16 666.67 648.31 613.25 553.8 544.09 514.72 518.84 420.69 388.23
262 İ. Ceylan et al. / Energy and Buildings 84 (2014) 258–267

Table 1 (Continued)

Mediterranean region of Turkey

Jan. Feb. March April May June July August Sep. Oct. Nov. Dec.

Yalova 428.13 524.71 598.89 649.12 641.23 607.43 543.92 543.59 515.34 524.41 419.75 401.32
Kocaeli 431.61 544.36 615.38 668.7 653.04 610.83 555.56 545.36 520.1 522.22 425.32 395.42
Bilecik 489.43 532.58 731.66 743.22 667.43 632.86 581.11 560 522.41 518.77 416.67 404.83
Sakarya 443.75 541.37 654.69 696.68 672.23 618.31 564.25 551.25 523.1 517.18 417.28 398.06
Aegean region of Turkey
İzmir 372.43 368.6 544.54 621.42 607.98 546.68 513.93 501.74 478.8 465.18 396.4 379.39
Denizli 411.88 426.09 599.12 665.82 645.23 592.43 559.59 536.19 512.85 510.2 426.02 423.17
Manisa 386.96 403.66 573.82 650.92 625.92 568.9 534.41 517.18 500 485.23 408.05 398.48
Kütahya 477.09 493.72 681.81 741.35 682.38 629.74 592.39 566.9 526.97 523.96 429.47 430.2
Aydın 385.66 384.62 572.86 640.3 617.83 560.65 525.23 517.03 493.4 491.53 413.19 402.25
Uşak 397.83 452.16 608.36 680.48 654.42 607.57 580.56 546.47 514.93 506.49 426.07 414.57
Muğla 411.31 390.32 595.51 660.15 627.65 580.56 543.7 534.92 509.07 504.46 425.96 402.57
Afyon 473.15 477.76 709.22 723.4 669.9 619.05 591.55 553.59 523.96 517.59 427.73 443.85

Fig. 4. Calculated PV module electrical efficiency and power values (a) for Mediterranean region (b) for East Anatolia region (c) for Southeast Anatolia region (d) for Central
Anatolia region (e) for Black sea Region (f) for Marmara region.

Power obtained any PV module cannot be an indicator for the that can be produced. Sunshine hours term of the region does not
energy that will be obtained without unknown sunshine hours of exist in the equations given as 1–5. In addition that this variable
the region which is planning to build of PV. For this reason sun- was not included while Fig. 4 creating. The map given in Fig. 5
shine hours of the region is an important criteria for the energy has been created whit multiplying of the calculated PV module

Fig. 5. Average calculated power of photovoltaic panels.

 İ. Ceylan et al. / Energy and Buildings 84 (2014) 258–267 263

Table 2
Monthly average outside air temperature of the regions of Turkey (◦ C).

Mediterranean region of Turkey

Jan. Feb. March April May June July August Sep. Oct. Nov. Dec.

Antalya 9.8 10.3 12.7 16.1 20.5 25.4 28.4 28.2 24.7 20 14.9 11.4
Burdur 2.6 3.6 7 11.5 16.5 21.2 24.7 24.4 19.9 14.3 8.6 4.3
Isparta 1.9 2.8 6.1 10.7 15.6 20.2 23.6 23.2 18.6 13 7.4 3.5
Mersin 10.2 10.9 13.7 17.5 21.4 25.2 27.9 28.3 25.2 21.2 15.7 11.7
Adana 9.6 10.5 13.5 17.5 21.8 25.7 28.1 28.5 26 21.5 15.5 11.1
Hatay 8.2 9.8 13.2 17.2 21.2 24.8 27.2 27.7 25.6 20.8 14.1 9.6
Osmaniye 8.4 9.6 12.5 16.8 21 25.2 27.9 28.4 25.3 20.5 13.7 9.6
K.Maraş 4.8 6.3 10.5 15.3 20.3 25.1 28.3 28.4 25.1 19 11.7 6.7
East Anatolia region of Turkey
Erzincan −2.9 −1.2 4.4 10.7 15.6 20 24 23.7 18.9 12.1 5.2 0.1
Elazığ −0.8 0.5 5.8 11.9 17.2 22.9 27.3 26.8 21.6 14.6 7.1 1.9
Tunceli −2 −0.4 5.6 11.9 17.1 22.7 27.3 26.9 21.6 14.7 6.9 1
Bingöl −2.5 −1.5 3.8 10.6 16.3 22.1 26.7 26.4 21.1 14 6.6 0.5
Erzurum −9.4 −8.1 −2.3 5.4 10.6 14.9 19.3 19.3 14.5 8 0.6 −6
Muş −7.4 −6 0.7 9 14.9 20.3 25.3 25.2 20.1 12.6 4.5 −2.8
Bitlis −2.9 −2.1 1.7 7.6 13.2 18.5 22.8 22.3 17.6 11.3 4.7 −0.8
Kars −10.4 −8 −2.4 5.3 10.2 13.9 17.5 17.6 13.4 7.3 0.3 −6.5
Ağrı −10.8 −9.4 −3.1 6.1 12 16.6 21.2 21.2 16.2 9.2 1.4 −6.4
Ardahan −11.4 −10.1 −3.6 4.4 9.4 12.9 16.3 16.2 12.2 6.6 −0.1 −7.7
Van −3.5 −2.9 1.5 7.7 13.1 18.2 22.3 21.9 17.2 10.7 4.3 −0.7
Iğdır −3.3 −0.5 6.4 13.1 17.7 22.1 25.8 25.1 19.9 12.7 5.7 −0.1
Hakkâri −4.7 −3.4 1.9 8 14.2 20.3 25 24.7 20.1 13 5.1 −1.5
Malatya 0.1 1.5 6.9 13 18.1 23.3 27.4 26.9 22.3 15.4 7.7 2.4
Southeast Anatolia region of Turkey
Gaziantep 3 4.2 8.2 13.2 18.6 24.1 27.8 27.4 22.8 16.2 9.4 4.9
Kilis 5.6 6.9 10.5 15.3 20.6 25.3 28.1 27.9 24.8 19.5 12.5 7.4
Adıyaman 4.5 5.7 9.9 15 20.6 26.8 31 30.5 25.7 18.9 11.6 6.5
Şanlıurfa 5.6 6.9 10.9 16.1 22.2 28.2 31.9 31.2 26.8 20.2 12.7 7.5
Diyarbakır 1.8 3.5 8.5 13.8 19.3 26.3 31.2 30.3 24.8 17.2 9.2 4
Mardin 3 4 8 13.4 19.6 25.6 29.9 29.5 25.1 18.3 10.7 5.3
Batman 2.7 4.9 9.6 14.7 19.8 26.6 31.1 30.1 24.8 17.7 9.9 4.6
Şırnak 1.8 2.9 6.7 11.6 17.1 23 27.4 27.7 23.1 16.4 8.8 3.7
Siirt 2.7 4.2 8.5 13.7 19.4 26 30.5 29.9 25 18 10.3 4.9
Central Anatolia region of Turkey
Eskişehir −0.1 1.3 5.1 10.2 15.1 19.1 21.7 21.4 17.2 12 6.2 2.1
Konya −0.2 1.2 5.7 11 15.7 20.2 23.6 23 18.6 12.5 6.1 1.8
Ankara 0.3 1.8 6.1 11.3 16.1 20.2 23.5 23.3 18.7 13.1 7.1 2.7
Çankırı −0.6 0.9 5.6 11 15.7 19.8 23 22.4 17.6 11.9 5.6 1.6
Aksaray 0.4 1.8 6.4 11.5 16.1 20.3 23.7 23 18.5 12.9 6.9 2.5
Kırıkkale 0.4 2.1 6.8 12.2 16.9 21.2 24.6 24.1 19.5 13.6 6.9 2.5
Kırşehir −0.2 1.1 5.4 10.6 15.3 19.6 23.1 22.8 18.2 12.4 6.2 2
Yozgat −1.9 −1 2.9 8.3 13 16.8 19.7 19.6 15.5 10.3 4.6 0.5
Niğde −0.4 0.8 5.2 10.5 15.1 19.3 22.6 22.2 17.7 12.1 6.1 1.8
Nevşehir −0.4 0.6 4.7 9.9 14.5 18.5 21.7 21.3 17 11.8 6.2 1.9
Kayseri −1.8 0 0.5 10.6 15 19.1 22.6 22 17.1 11.6 5 0.5
Karaman 0.4 1.6 6 11.3 16.1 24 23.5 22.9 18.5 12.8 6.7 2.6
Sivas −3.3 −2.1 3 9.1 13.6 17.2 20.2 20.1 16.2 10.9 4.6 −0.4
Black sea region of Turkey
Bolu 0.7 2 5 9.8 14 17.4 19.9 19.7 16.1 11.8 6.9 3.1
Düzce 3.7 5.1 7.7 12.3 16.6 20.5 22.6 22.3 18.6 14.3 9.5 5.9
Zonguldak 6 6 7.5 11.4 15.5 19.7 21.9 21.8 18.6 15.2 11.7 8.4
Karabük 2.9 4.5 7.9 12.7 17.3 20.9 23.9 23.5 19.5 14.3 8.2 4.5
Bartın 4.1 4.6 7 11.2 15.6 19.8 22.1 21.6 17.7 13.7 9.1 6
Kastamonu −1 0.6 4.4 9.5 14 17.5 20.3 19.8 15.6 10.6 5 1
Çorum −0.4 0.9 5.1 10.5 14.8 18.5 21.1 21 17 11.8 5.9 1.9
Sinop 6.9 6.5 7.5 10.7 14.9 19.8 22.7 23 19.9 16.2 12.4 9.3
Samsun 7 6.9 8 11.3 15.6 20.3 23.3 23.4 20 16.1 12.4 9.3
Amasya 2.6 4.4 8.3 13.5 17.8 21.6 24.1 23.9 20 14.6 8.6 4.7
Tokat 1.7 3.3 7.4 12.5 16.4 19.8 22.3 22.3 18.7 13.7 7.9 3.9
Ordu 6.7 6.6 7.9 11.4 15.6 20.3 22.9 23.1 19.8 15.9 11.8 8.8
Giresun 7.2 7.1 8.1 11.4 15.5 20.1 22.8 23 20 16.3 12.5 9.6
Gümüşhane −1.8 −0.7 3.6 9.4 13.7 17.2 20.2 20.1 16.6 11.4 5 0.5
Trabzon 7.4 7.3 8.5 11.8 15.9 20.4 23.2 23.3 20.3 16.5 12.7 9.6
Bayburt −6.5 −5.3 0 7 11.7 15.4 19 18.8 14.7 9.2 2.6 −3.2
Rize 6.5 6.4 7.9 11.6 16 20.3 22.8 23 19.9 16 11.7 8.5
Artvin 2.6 3.7 6.9 11.8 15.7 18.6 20.6 20.7 17.9 14 8.9 4.4
Marmara region of Turkey
Çanakkale 6.2 6.5 8.3 12.5 17.5 22.4 25 24.8 20.8 16 11.9 8.4
Balıkesir 4.7 5.7 8.3 13.1 17.8 22.6 24.8 24.5 20.6 15.8 10.4 6.7
Edirne 2.6 4.3 7.7 12.9 18.1 22.4 24.7 24.2 19.8 14.2 9.1 4.5
Tekirdağ 4.8 5.1 7.3 11.9 16.8 21.4 23.8 23.6 19.9 15.4 11 7.2
Kırklareli 2.8 3.9 6.8 12 17.2 21.6 23.9 23.2 19.2 13.9 8.9 5
İstanbul 6.5 6.5 8.3 12.7 17.5 22.1 24.4 24.2 20.9 16.4 12.2 8.7
Bursa 5.2 6.1 8.4 12.8 17.6 22.2 24.5 24.1 20.1 15.3 10.6 7.4
264 İ. Ceylan et al. / Energy and Buildings 84 (2014) 258–267

Table 2 (Continued)

Mediterranean region of Turkey

Jan. Feb. March April May June July August Sep. Oct. Nov. Dec.

Yalova 6.4 6.7 8.2 12.3 16.9 21.4 23.6 23.5 20 15.8 11.7 8.7
Kocaeli 6.1 6.6 8.5 13 17.5 21.8 23.7 23.6 20.3 16 11.8 8.4
Bilecik 2.4 3.5 6.6 11.5 16.1 19.9 22.1 21.9 18.3 13.9 8.9 4.7
Sakarya 5.9 6.4 8.4 12.8 17.2 21.4 23.3 23 19.5 15.5 11.4 8.2
Aegean region of Turkey
İzmir 8.8 9.4 11.7 15.9 20.9 25.7 28 27.6 23.6 18.9 14.1 10.6
Denizli 5.8 6.9 10 14.6 19.7 24.7 27.4 26.9 22.4 16.8 11.4 7.6
Manisa 6.7 7.9 10.7 15.2 20.5 25.5 28.1 27.7 23.4 18 12.2 8.5
Kütahya 0.4 1.7 5.2 10 14.6 18.4 20.9 20.6 16.6 11.8 6.7 2.6
Aydın 8.1 9.2 11.8 15.8 20.9 25.9 28.4 27.4 23.3 18.4 13.3 9.6
Uşak 2.3 3 6.3 10.8 15.8 20.3 23.6 23.4 19 13.4 8 4.2
Muğla 5.5 6 8.6 12.5 17.6 22.9 26.3 26 21.7 16 10.5 7
Afyon 0.2 1.5 5.4 10.3 15 19.1 22.3 22 17.8 12.3 6.8 2.5

Table 3
Estimated back side temperature of the photovoltaic module (◦ C).

Mediterranean region of Turkey

Jan. Feb. March April May June July August Sep. Oct. Nov. Dec.

Antalya 22.6 22.9 24.9 29.6 37.9 46.4 46.8 46.3 44.6 33.7 24.8 23.4
Burdur 13.8 15.7 21.8 28.3 29.9 39.5 45.9 44.9 33.7 24.6 21.7 16.8
Isparta 12.4 13.9 21.9 28.4 29.6 36.4 44.8 43.5 30.2 23.6 20.7 15.3
Mersin 22.8 23.1 24.7 30.0 40.3 46.1 47.0 46.5 43.8 36.6 25.0 23.6
Adana 22.5 23.0 24.9 30.3 41.4 46.6 46.5 45.6 43.0 37.5 24.9 23.3
Hatay 21.7 22.7 23.9 28.5 39.3 46.1 47.0 46.7 43.0 34.6 24.3 22.5
Osmaniye 21.6 22.5 24.8 29.2 39.0 46.4 46.3 45.6 42.6 34.5 24.3 22.5
K.Maraş 17.0 18.8 25.1 28.6 37.3 46.3 46.8 45.5 43.6 30.6 23.5 19.3
East Anatolia region of Turkey
Erzincan 3.8 8.2 23.8 29.9 31.4 36.4 45.0 44.9 31.2 22.8 17.9 8.8
Elazığ 6.9 9.6 24.3 28.5 31.0 43.8 46.8 45.6 38.7 24.8 20.2 12.2
Tunceli 4.7 8.6 24.6 29.9 31.7 43.3 47.0 46.5 39.47 24.9 19.8 10.5
Bingöl 3.9 8.2 27.0 30.3 31.1 42.1 46.9 46.4 37.8 24.3 19.4 9.6
Erzurum 1.1 0.6 13.1 26.6 29.3 27.5 33.0 32.4 24.7 20.2 9.9 0.9
Muş 1.1 4.3 33.0 29.8 30.4 37.0 46.4 46.0 34.5 23.2 16.9 4.0
Bitlis 7.3 11.3 34.6 31.1 29.8 31.3 43.5 42.1 28.2 22.2 17.3 7.2
Kars 1.0 8.6 10.6 29.2 33.3 28.3 30.1 28.7 24.9 19.0 9.3 0.9
Ağrı 0.7 0.4 15.3 26.7 29.9 28.8 39.5 38.8 26.4 20.9 11.6 0.9
Ardahan 1.9 10.2 9.0 28.2 36.5 30.9 31.7 29.0 26.7 18.4 8.3 1.2
Van 2.6 3.3 14.2 21.8 27.8 30.7 42.6 40.8 27.3 22.5 16.9 6.7
Iğdır 1.7 6.5 19.8 23.7 28.9 40.7 44.8 43.3 32.0 24.0 17.3 4.8
Hakkâri 1.1 2.6 12.2 19.8 25.0 35.4 45.1 43.5 33.2 24.1 16.9 3.8
Malatya 8.8 11.5 23.3 27.6 32.3 44.4 46.8 45.4 40.5 25.5 20.9 13.1
Southeast Anatolia region of Turkey
Gaziantep 14.6 16.8 22.3 26.1 31.9 45.4 46.8 45.9 40.9 26.0 22.5 18.0
Kilis 18.6 20.3 23.4 27.2 37.6 46.4 46.9 46.3 43.5 31.1 23.9 21.0
Adıyaman 17.2 18.5 23.9 27.7 38.1 46.9 46.1 45.3 43.7 30.4 23.5 19.6
Şanlıurfa 19.0 20.1 23.1 28.2 42.3 46.7 45.9 45.1 43.9 33.7 24.0 21.0
Diyarbakır 12.2 14.7 22.6 28.2 34.9 46.1 46.3 42.2 44.9 27.3 22.3 16.4
Mardin 14.6 16.0 22.0 27.3 34.7 45.9 45.1 43.9 43.6 29.0 23.1 18.6
Batman 13.8 16.8 26.5 28.9 35.6 46.9 46.2 45.4 43.4 28.2 22.6 17.1
Şırnak 12.2 14.0 22.7 28.2 29.8 43.5 46.2 44.6 42.2 26.4 22.0 16.1
Siirt 13.5 15.7 27.7 29.7 34.6 46.7 46.5 45.9 43.9 28.9 22.9 17.6
Central Anatolia region of Turkey
Eskişehir 10.7 11.1 40.7 33.1 30.0 33.4 40.9 39.6 27.6 22.8 19.7 13.0
Konya 8.2 10.9 21.6 26.4 29.5 36.3 44.6 42.7 30.1 23.3 19.5 12.2
Ankara 9.1 12.1 21.3 27.2 30.5 36.6 44.5 43.7 30.5 23.8 20.6 14.1
Çankırı 7.4 13.5 18.6 26.2 31.1 35.4 43.9 42.0 28.5 23.0 19.0 12.0
Aksaray 9.3 12.1 21.4 25.6 29.8 36.5 44.6 42.2 29.7 23.7 20.4 13.5
Kırıkkale 9.4 12.6 19.2 25.8 31.1 39.6 45.6 44.4 32.6 24.1 20.5 13.7
Kırşehir 8.1 10.7 19.4 24.8 29.7 34.3 43.4 42.0 29.2 23.4 19.7 12.6
Yozgat 5.0 6.7 21.2 26.1 29.8 28.4 33.9 32.9 25.6 22.1 17.6 9.5
Niğde 7.7 10.2 22.5 26.5 29.4 33.4 42.9 40.7 28.1 23.2 19.5 12.1
Nevşehir 7.7 9.7 22.3 26.0 29.4 31.1 40.4 38.0 27.1 23.0 23.1 12.2
Kayseri 5.3 8.7 22.9 27.4 30.0 32.8 42.9 40.1 27.2 22.9 18.1 9.6
Karaman 9.4 11.7 21.6 25.7 29.4 45.3 44.6 42.6 29.7 23.9 20.2 13.9
Sivas 3.3 6.3 24.6 28.5 30.7 29.7 36.2 35.0 26.3 22.3 17.5 8.0
Black sea region of Turkey
Bolu 9.9 17.6 27.7 29.3 30.6 30.2 35.2 33.9 26.4 22.8 20.4 15.0
Düzce 15.9 17.0 24.3 29.3 31.0 37.5 42.2 41.7 30.2 24.6 22.5 19.4
Zonguldak 19.3 17.8 21.6 26.4 30.2 34.6 40.7 39.9 30.0 25.3 23.6 21.8
Karabük 14.5 16.3 21.9 27.4 32.1 38.8 45.1 43.9 33.0 24.6 21.6 17.4
Bartın 16.8 16.2 21.7 25.6 31.0 34.8 41.0 39.0 28.4 24.2 22.3 19.5
Kastamonu 6.6 11.9 21.0 26.3 31.2 29.9 36.1 33.8 25.9 23.1 18.1 10.7
 İ. Ceylan et al. / Energy and Buildings 84 (2014) 258–267 265

Table 3 (Continued)

Mediterranean region of Turkey

Jan. Feb. March April May June July August Sep. Oct. Nov. Dec.

Çorum 7.75 11.7 19.8 26.4 30.6 31.9 38.9 37.8 27.4 23.0 19.4 12.6
Sinop 20.5 18.3 21.9 25.8 31.0 35.5 43.3 43.5 34.2 26.0 23.8 22.4
Samsun 20.5 18.8 23.8 27.8 30.8 37.3 44.3 44.6 34.5 26.0 23.8 22.4
Amasya 14.0 16.3 24.0 29.1 32.4 40.9 45.4 44.8 34.4 24.8 22.0 17.6
Tokat 12.0 15.1 24.9 29.2 31.3 35.6 42.6 42.1 30.5 28.6 21.4 16.1
Ordu 19.8 18.5 30.0 30.5 31.2 37.5 42.8 43.9 34.1 25.9 23.6 22.0
Giresun 19.7 19.7 28.5 30.2 31.4 37.0 42.1 43.5 35.0 26.4 23.9 22.3
Gümüşhane 7.6 17.0 26.8 30.8 31.7 31.8 37.26 36.6 27.3 22.3 18.0 9.5
Trabzon 20.4 19.2 29.4 31.6 31.6 37.4 40.9 42.0 36.3 26.1 24.0 22.5
Bayburt 1.0 12.9 26.2 32.4 32.2 30.6 34.4 33.0 25.5 20.7 13.9 2.9
Rize 20.1 18.7 26.0 30.8 31.4 36.9 40.2 41.4 35.5 26.6 23.6 21.9
Artvin 14.1 19.1 21.4 29.8 31.0 33.4 36.5 38.2 31.4 24.8 22.1 17.1
Marmara region of Turkey
Çanakkale 19.4 20.1 19.9 23.7 29.4 41.6 43.5 43.3 34.9 25.4 23.6 21.4
Balıkesir 17.7 19.0 21.8 27.7 30.7 42.9 44.6 44.1 35.2 25.7 23.0 20.2
Edirne 13.6 17.1 19.3 23.6 30.1 41.7 42.9 43.0 31.7 24.6 22.1 16.8
Tekirdağ 17.9 17.8 20.3 24.9 28.9 38.8 41.9 41.9 33.0 25.42 23.3 20.8
Kırklareli 14.4 16.2 20.1 23.8 29.7 38.8 41.1 40.6 31.0 24.3 22.1 18.2
İstanbul 19.5 19.7 34.6 32.0 32.3 42.0 45.6 45.5 39.0 31.4 25.1 22.7
Bursa 18.4 17.9 22.5 28.3 30.8 42.3 45.4 44.7 32.5 25.4 23.1 21.0
Yalova 19.8 18.4 21.8 27.1 29.7 40.2 44.0 43.9 33.8 25.9 23.6 22.0
Kocaeli 19.4 18.5 23.1 28.4 30.8 41.3 44.5 44.1 34.9 26.17 23.6 21.8
Bilecik 13.0 15.1 32.1 31.5 29.9 35.8 41.9 41.0 29.4 24.3 22.1 17.7
Sakarya 19.0 18.2 26.0 29.7 31.0 40.3 44.1 43.3 32.5 25.6 23.4 21.7
Aegean region of Turkey
İzmir 22 22.4 22.6 28 38.7 46 45 44.3 41.4 29.5 24.4 23.1
Denizli 19.2 20.3 22.9 28.9 35.4 46 46.8 45.9 40.8 26.8 23.4 21
Manisa 20.3 21.4 22.3 28.6 37.6 46.4 45.9 45.1 42 28.2 23.8 21.9
Kütahya 9.4 11.8 28 31.6 29.6 31.8 38.5 37 26.8 22.6 20.1 14
Aydın 21.6 22.3 23.1 28.6 38.8 46.4 45.5 45.1 41.6 29 24.2 22.6
Uşak 13.4 14.5 21.4 28.4 29.2 36.7 44.8 43.8 30.9 23.9 21.3 16.9
Muğla 18.8 19.5 21.9 27.8 30.2 43.7 46.1 45.6 38.9 26 23 20.6
Afyon 9 11.5 30.7 30.7 29.3 33.2 42.5 41.1 28.5 23.1 20.2 13.7

Table 4
Monthly average daily hours of bright sunshine for the regions of Turkey (h).

Mediterranean region of Turkey

Jan. Feb. March April May June July August Sep. Oct. Nov. Dec.

Antalya 4.95 6.1 7.24 8.29 9.7 11.55 11.84 11.29 9.8 7.68 5.97 4.55
Burdur 4.74 5.82 6.98 7.97 9.61 11.4 11.85 11.25 9.78 7.45 5.72 4.23
Isparta 4.38 5.46 6.82 7.77 9.42 11.1 11.7 11.13 9.64 7.17 5.44 3.95
Mersin 4.99 6.04 7.35 8.38 9.94 11.18 11.45 11.03 10.02 7.91 6.15 4.64
Adana 4.67 5.65 6.97 7.84 9.72 11.29 11.77 11.22 10.15 7.78 5.86 4.21
Hatay 5.09 6.22 7.17 8.28 10.23 11.14 10.89 10.47 9.8 7.86 6.37 4.99
Osmaniye 4.57 5.66 6.76 7.87 9.83 11.39 11.79 11.19 10.15 7.78 5.92 4.24
K.Maraş 4.21 5.47 6.61 7.85 9.57 11.49 12.07 11.43 10.13 7.55 5.56 3.86
East Anatolia region of Turkey
Erzincan 3.73 4.85 6.15 7.14 8.63 10.29 10.67 10.23 9.12 6.52 4.71 3.27
Elazığ 4.13 5.14 6.37 7.56 9.39 11.45 12.01 11.33 9.86 7.14 5.12 3.56
Tunceli 4.02 4.99 6.25 7.27 9 10.88 11.43 10.86 9.49 6.85 4.88 3.41
Bingöl 4.08 4.93 6.02 7.18 9.08 11.01 11.51 10.87 9.54 6.87 4.89 3.45
Erzurum 3.85 4.71 5.82 6.95 8.28 9.86 10.3 9.91 8.5 6.24 4.57 3.31
Muş 3.66 4.8 5.56 7.25 9.13 10.83 11.42 10.76 9.6 6.89 4.98 3.47
Bitlis 3.46 4.72 5.56 7.13 9.27 10.95 11.31 10.62 9.81 6.86 5.19 3.59
Kars 3.82 4.74 6.24 7.04 7.9 9.89 10.55 10.36 8.15 6.46 4.82 3.44
Ağrı 4.1 5.43 6.25 7.62 9.08 10.82 11.32 10.79 9.45 7.14 5.42 3.96
Ardahan 3.75 4.31 6.01 6.93 7.12 8.76 8.85 8.94 7.33 5.9 4.68 3.42
Van 5.27 6.4 7.39 8.5 10.11 11.55 11.65 10.97 10.31 7.65 6.16 4.93
Iğdır 5.88 7.34 8.43 9.64 10.87 12.34 12.59 11.65 10.25 8.33 6.86 5.61
Hakkari 6.65 8.17 8.92 9.95 11.31 12.71 12.61 11.85 10.7 8.69 7.42 6.42
Malatya 4.23 5.3 6.59 7.86 9.41 11.43 12.09 11.44 9.96 7.28 5.26 3.64
Southeast Anatolia region of Turkey
Gaziantep 4.6 5.78 6.82 8.1 9.93 11.63 11.74 11.07 10.03 7.8 5.98 4.38
Kilis 4.7 5.92 6.8 8.12 10.15 11.48 11.43 10.86 9.81 7.86 6.14 4.58
Adıyaman 4.51 5.49 6.74 8.08 9.7 11.78 12.25 11.52 10.17 7.56 5.56 4.01
Şanlıurfa 4.68 5.62 6.92 8.14 9.96 12.24 12.42 11.66 10.11 7.71 5.87 4.4
Diyarbakır 3.73 4.89 6.16 7.21 8.76 10.21 11.06 10.45 8.83 6.48 4.73 3.36
Mardin 4.36 5.45 6.74 7.9 10.01 12.52 12.84 12.03 10.07 7.59 5.83 4.43
Batman 3.92 5.02 6.16 7.64 9.71 11.72 12.09 11.34 10.05 7.33 5.48 3.92
Şırnak 4.32 5.52 6.65 7.61 9.87 12.12 12.42 11.67 9.9 7.38 5.86 4.5
Siirt 3.84 5 6.04 7.38 9.64 11.52 11.78 11.07 9.99 7.19 5.53 4.02
266 İ. Ceylan et al. / Energy and Buildings 84 (2014) 258–267

Table 4 (Continued)

Mediterranean region of Turkey

Jan. Feb. March April May June July August Sep. Oct. Nov. Dec.

Central Anatolia region of Turkey

Eskişehir 3.23 4.74 4.37 6.18 8.78 10.16 10.75 10.1 8.85 6.25 4.73 3.37
Konya 4.19 5.51 6.88 8.03 9.46 11.28 11.97 11.36 9.79 7.35 5.53 3.93
Ankara 3.73 4.89 6.16 7.21 8.76 10.21 11.06 10.45 8.83 6.48 4.73 3.36
Çankırı 3.69 4.75 6.24 7.08 8.27 9.9 10.7 10.09 8.23 6.05 4.47 3.21
Aksaray 4.11 5.4 6.8 7.97 9.36 11.27 12.12 11.53 9.81 7.36 5.45 3.73
Kırıkkale 3.79 4.99 6.57 7.48 8.66 10.32 11.17 10.63 8.84 6.57 4.76 3.31
Kırşehir 3.97 5.18 6.61 7.72 9.02 10.78 11.73 11.15 9.32 6.98 5.07 3.51
Yozgat 3.8 5.12 6.1 7.29 8.72 10.6 11.41 10.92 9.1 6.85 4.99 3.33
Niğde 4.41 5.48 6.74 7.85 9.51 11.39 12.16 11.57 10.06 7.61 5.65 3.9
Nevşehir 4.07 5.32 6.43 7.65 9.16 11.17 12.05 11.48 9.68 7.27 5.36 3.57
Kayseri 4.08 5.31 6.37 7.59 9.21 11.22 12.03 11.47 9.84 7.39 5.39 3.55
Karaman 4.46 5.85 7.14 8.44 9.84 11.51 12.02 11.47 10.11 7.77 5.95 4.31
Sivas 3.76 5.02 5.99 7.2 8.73 10.5 11.09 10.65 9.23 6.8 4.97 3.3
Black sea region of Turkey
Bolu 3.28 4.41 5.19 6.53 8.31 9.73 10.44 9.74 8.15 5.78 4.26 3.12
Düzce 3.17 4.25 5.2 6.37 8.21 9.71 10.4 9.67 7.96 5.6 4.03 3.04
Zonguldak 3.27 4.25 5.37 6.62 8.1 9.83 10.59 9.79 7.84 5.54 4.02 3.07
Karabük 3.41 4.4 5.62 6.75 7.97 9.77 10.58 9.83 7.84 5.6 4.2 3.02
Bartın 3.31 4.3 5.29 6.66 7.81 9.94 10.79 9.94 7.64 5.41 4.08 2.96
Kastamonu 3.39 4.44 5.45 6.65 7.86 9.82 10.66 9.87 7.67 5.58 4.26 3.1
Çorum 3.6 4.79 5.92 6.99 8.29 9.91 10.66 10.16 8.32 6.19 4.57 3.21
Sinop 3.46 4.42 5.36 6.62 7.8 9.44 10.08 9.32 7.57 5.56 4.32 3.21
Samsun 3.6 4.41 5.17 6.43 7.92 9.15 9.52 8.97 7.61 5.72 4.32 3.22
Amasya 3.57 4.65 5.48 6.6 8.16 9.58 10.14 9.7 8.1 6.11 4.51 3.19
Tokat 3.6 4.72 5.5 6.71 8.27 9.74 10.12 9.79 8.4 6.33 4.62 3.2
Ordu 3.44 4.4 4.77 6.24 7.96 9.06 8.92 8.6 7.72 5.8 4.34 3.11
Giresun 3.29 4.49 5.15 6.56 8.04 9 8.87 8.6 7.91 5.82 4.37 3.05
Gümüşhane 3.02 4.6 5.54 6.74 8.17 9.25 9.3 8.99 8.28 5.95 4.43 2.99
Trabzon 3.17 4.29 4.88 6.16 7.64 8.35 7.9 7.62 7.43 5.42 4.2 2.99
Bayburt 3.53 4.63 5.63 6.72 8.22 9.42 9.49 9.14 8.37 6.03 4.46 3.2
Rize 3.38 4.21 5.21 6.4 7.47 8.19 7.75 7.54 7.19 5.37 4.17 3.02
Artvin 3.49 4.13 5.56 6.61 7.12 8.1 7.65 7.64 6.86 5.38 4.27 3.14
Marmara region of Turkey
Çanakkale 4.21 5.5 6.28 7.77 9.54 11.56 11.85 11.01 9 6.81 4.93 3.83
Balıkesir 4.05 5.05 5.92 7.26 9.23 10.99 11.44 10.65 8.84 6.53 4.69 3.64
Edirne 3.93 5.31 5.94 7.69 9.54 10.96 11.73 10.71 8.88 6.03 4.56 3.37
Tekirdağ 3.69 4.96 5.54 7.26 9.05 11.21 11.92 10.83 8.32 5.66 4.15 3.13
Kırklareli 3.68 5.09 5.43 7.37 8.93 11.47 12.5 11.23 8.3 5.32 4.22 2.87
İstanbul 3.46 4.43 5.32 6.85 8.61 10.51 11.17 10.14 7.83 5.22 3.85 2.96
Bursa 3.71 4.51 5.64 6.96 8.9 10.11 10.78 9.98 8.49 5.84 4.36 3.4
Yalova 3.27 4.25 5.41 6.84 8.78 9.96 10.7 9.75 8.15 5.53 4.05 3.04
Kocaeli 3.29 4.17 5.2 6.55 8.56 9.79 10.44 9.59 7.96 5.4 3.95 3.06
Bilecik 3.31 4.45 4.77 6.27 8.75 9.86 10.48 9.75 8.48 5.86 4.44 3.31
Sakarya 3.2 4.23 5.01 6.33 8.39 9.72 10.35 9.56 8.01 5.53 4.05 3.09
Aegean region of Turkey
İzmir 4.86 5.86 6.96 8.03 9.77 11.89 12.2 11.48 9.67 7.61 5.55 4.27
Denizli 4.88 5.75 6.86 7.9 9.64 11.36 11.83 11.19 9.73 7.35 5.61 4.23
Manisa 4.6 5.45 6.57 7.62 9.49 11.32 11.77 11.06 9.26 7.11 5.22 3.94
Kütahya 3.75 4.78 5.5 6.65 8.91 10.29 10.77 10.09 8.9 6.26 4.75 3.51
Aydın 5.16 5.98 7 8.09 9.76 11.79 12.09 11.45 9.85 7.67 5.76 4.45
Uşak 4.6 5.33 6.46 7.48 9.37 10.83 11.42 10.76 9.38 6.93 5.14 3.98
Muğla 5.13 6.2 7.12 8.18 9.91 11.73 11.9 11.31 9.92 7.85 6.01 7.67
Afyon 3.91 5.17 5.64 7.05 9.27 10.75 11.36 10.73 9.39 6.82 5.12 3.74

power values and sunshine hours illustrated in Fig. 4 and Table 4, outside air temperature. The most efficiency cities are shown as
respectively. bold in Fig. 4 also produced energy.
Estimated module temperature with ANN depending on solar Outside air temperature and solar radiation were taken con-
radiation and outlet air temperature has shown in Table 1. Gener- sidered for the predicting PV module temperature by using ANN
ating power and efficiency of panel calculated from Eqs. (2.1)–(2.5) also outside air relative humidity and air velocity are neglected.
has shown in Fig. 4. As the solar radiation increased, the efficiency of Air velocity is very important factor for the cooling PV module.
PV module decreased but generated power of PV module increased PV module will be affected as positive from the outside air veloc-
with the increasing of solar radiation. This situation can be seen ity in terms of this study. Although this study has been made for
in Fig. 4. Solar radiation can be high in low outside air tempera- Turkey, it will be pioneering work as a reference for the other
ture or solar radiation can be low in high outside air temperature countries.
especially in winter season. So outside air temperature the most Calculated average daily power of PV module was shown in Fig. 5
important factor for panel temperature also generated power of PV created by using Flash color coding program. Outside air temper-
module. Although solar radiation increasing, the generated power ature, back surface temperature of PV module, solar radiation and
of PV module cannot be increased depending upon outside air tem- the sunshine hour’s values are the variables which were used to cre-
perature and panel temperature. The most appropriate situation for ating Fig. 5. The power which will be produced from the PV panels
high generated power is maximum solar radiation and the lowest was calculated depend on the PV module temperature computed
 İ. Ceylan et al. / Energy and Buildings 84 (2014) 258–267 267

via ANN. Obtained data was multiplied with sunshine hour’s and [5] R. Eke, H. Demircan, Performance analysis of a multi crystalline Si photovoltaic
the map was created. module under Mugla climatic conditions in Turkey, Energy Conversion and
Management 65 (2013) 580–586.
It can be said that from the map the metropolitan city Istan- [6] B.J. Huang, Y.C. Huang, G.Y. Chen, P.C. Hsu, K. Li, Improving solar PV system
bul placed at Marmara Region of Turkey has high power values as efficiency using one-axis 3-position sun tracking, Energy Procedia 33 (2013)
the average energy is 475–500 W h/m2 -day. Power that can be pro- 280–287.
[7] S.Y. Wu, Q.L. Zhang, L. Xiao, F.H. Guo, A heat pipe photovoltaic/thermal (PV/T)
duced from the PV modules was found higher of the cities which hybrid system and its performance evaluation, Energy and Buildings 43 (2011)
solar radiation and sunshine hours high than others according to 3558–3567.
Fig. 5. This can be seen from the figure colored as red and orange. [8] H. Bahaidarah, A. Subhan, P. Gandhidasan, S. Rehman, Performance evaluation
of a PV (photovoltaic) module by back surface water cooling for hot climatic
Likewise, the power that can be produced from the PV modules
conditions, Energy 59 (2013) 445–453.
was found lower which cities has lower solar radiation and sun- [9] C.D. Corbin, Z.J. Zhai, Experimental and numerical investigation on thermal and
shine hours and they showed as green in the map. According to the electrical performance of a building integrated photovoltaic–thermal collector
system, Energy and Buildings 42 (2010) 76–82.
map 450–500 W h/m2 -day power can be produced from the impor-
[10] M.C.A. Garcia, J.L. Balenzategui, Estimation of photovoltaic module yearly
tant part of Turkey in Fig. 5 and also these parts are shown in red temperature and performance based on nominal operation cell temperature
and orange. calculations, Renewable Energy 29 (2004) 1997–2010.
[11] A.L. Hanlin, J.S. Stein, Improvement and validation of a transient model to pre-
dict photovoltaic module temperature, in: SAND2012-4307, Sandia National
4. Conclusion Laboratories, New Mexico, USA, 2012.
[12] Z. Ye, A. Nobre, T. Reindl, J. Luther, C. Reise, On PV module temperatures in
Experimental and ANN approach of a PV system is stud- tropical regions, Solar Energy 88 (2013) 80–87.
[13] İ. Ceylan, A.E. Gürel, H. Demircan, B. Aksu, Cooling of a photovoltaic module
ied regarding its temperature, efficiency and power. The system with temperature controlled solar collector, Energy and Buildings 42 (2014)
is tested under the climatic conditions of Karabük, Turkey and 96–101.
obtained data used for the all regions of the Turkey. Based on the [14] F. Almonacid, C. Rus, L. Hontoria, M. Fuentes, G. Nofuentes, Characterization of
Si-crystalline PV modules by artificial neural networks, Renewable Energy 34
results obtained, the following conclusions are drawn:The outside (2009) 941–949.
air temperature is very important factor in term of photovoltaic [15] M.S. Ashhab, Optimization and modeling of a photovoltaic solar integrated
module temperature. As can be seen in Table 1, Outside air tem- system by neural networks, Energy Conversion and Management 49 (2008)
3349–3355.
perature can be compared with estimated module temperature. [16] S.K.H. Chow, E.W.M. Lee, D.H.W. Li, Short-term prediction of photovoltaic
For this study, the external air velocity and relative humidity values energy generation by intelligent approach, Energy and Buildings 55 (2012)
were neglected. Also research can be improved by adding of these 660–667.
[17] M. Almaktar, H.A. Rahman, M.Y. Hassan, I. Saeh, Artificial neural network-based
two values.As the solar radiation increased, photovoltaic module
photovoltaic module temperature estimation for tropical climate of Malaysia
electrical efficiency decreased according to Fig. 4. But power of Pho- and its impact on photovoltaic system energy yield, Progress in Photovoltaics:
tovoltaic module increased as the solar radiation.Istanbul is the Research and Applications (2013), http://dx.doi.org/10.1002/pip.2424/.
[18] H. Ravaee, S. Farahat, F. Sarhaddi, Artificial neural network based model of pho-
largest city in Turkey and is also high solar radiation values for
tovoltaic thermal (PV/T) collector, The Journal of Mathematics and Computer
investments that can be made to this city are very important.The Science 4 (2012) 411–417.
map created as a result of this study will be an important source to [19] D. Vasarevicius, R. Martavicius, M. Pikutis, Application of artificial neural
the investors and individual users for the estimate power that can networks for maximum power point tracking of photovoltaic panels, Elektron-
ika ir Elektrotechnika 18 (2012) 65–68.
be produced from the PV panels.Similar maps can be created for [20] A.B.G. Bahgat, N.H. Helwa, G.E. Ahmad, E.T. El Shenawy, Maximum power point
the other countries and for the various types of PV modules using tracking controller for PV systems using neural networks, Renewable Energy
the method dealing with in this study. 30 (2005) 1257–1268.
[21] M. Hadjab, S. Berrah, H. Abid, Neural network for modeling solar panel, Inter-
national Journal of Energy 1 (2012) 9–16.
Acknowledgement [22] J.N. Reddy, B.M. Manjunatha, M. Matam, Improving of photovoltaic system
with neural network based MPPTc to DC shunt motor, International Journal
of Modern Engineering Research 3 (2013) 2901–2907.
The authors would like to thank the Karabük University Sci- [23] A.A. Kulaksız, R. Akkaya, Training data optimization for ANNs using genetic
entific Research Projects Unit, Karabük/TURKEY for providing the algorithms to enhance MPPT efficiency of a stand-alone PV system, Turk-
financial supports for this study under the KBÜ-BAP-13/2-YL-037 ish Journal of Electrical Engineering & Computer Sciences 20 (2012)
241–254.
project. [24] E. Karatepe, M. Boztepe, M. Çolak, Neural network based solar cell model,
Energy Conversion and Management 47 (2006) 1159–1178.
References [25] M.F.N. Tajuddin, S.M. Ayob, Z. Salam, M.S. Saad, Evolutionary based maximum
power point tracking technique using differential evolution algorithm, Energy
[1] F. Almonacid, C. Rus, P.J. Perez, L. Hontoria, Estimation of the energy of a PV gen- and Buildings 67 (2013) 245–252.
erator using artificial neural network, Renewable Energy 34 (2009) 2743–2750. [26] A. Mellit, S. Sağlam, S.A. Kalogirou, Artificial neural network-based model for
[2] F. Dinçer, Overview of the photovoltaic technology status and perspective in estimating the produced power of a photovoltaic module, Renewable Energy
Turkey, Renewable and Sustainable Energy Reviews 15 (2011) 3768–3779. 60 (2013) 71–78.
[3] G.B. Weiss, C. Wray, W. Delp, P. Ly, H. Akbari, R. Levinson, Electricity production [27] İ. Ceylan, O. Erkaymaz, E. Gedik, A.E. Gürel, The prediction of photovoltaic
and cooling energy savings from installation of a building-integrated photo- module temperature with artificial neural networks, Case Studies in Thermal
voltaic roof on an office building, Energy and Buildings 56 (2013) 210–220. Engineering 3 (2014) 11–20.
[4] A. Mellit, S. Saglam, S.A. Kalogirou, Artificial neural network-based model for [28] N. Caglayan, C. Ertekin, F. Evrendilek, Spatial viability analysis of grid-connected
estimating the produced power of a photovoltaic module, Renewable Energy photovoltaic power systems for Turkey, Electrical Power and Energy Systems
60 (2013) 71–78. 56 (2014) 270–278.