Beruflich Dokumente
Kultur Dokumente
discussions, stats, and author profiles for this publication at: http://www.researchgate.net/publication/257548208
CITATIONS
READS
22
152
4 AUTHORS, INCLUDING:
Ahmed Ouammi
Hanane Dagdougui
SEE PROFILE
SEE PROFILE
Rachid Benchrifa
University Mohammed V Agdal ; Facult de
21 PUBLICATIONS 115 CITATIONS
SEE PROFILE
Unite des Technologies et Economie des Energies Renouvelables (TEER), CNRST Rabat, Morocco
Department of Communication, Computer and System Sciences (DIST), Faculty of Engineering, University of Genoa, Genoa, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 18 October 2011
Received in revised form
25 March 2012
Accepted 31 March 2012
An articial neural network (ANN) model is used to forecast the annual and monthly solar irradiation in
Morocco. Solar irradiation data are taken from the new Satellite Application Facility on Climate
Monitoring (CM-SAF)-PVGIS database. The database represents a total of 12 years of data from 1998 to
2010. In this paper, the data are inferred using an ANN algorithm to establish a forward/reverse
correspondence between the longitude, latitude, elevation and solar irradiation. Specically, for the
ANN model, a three-layered, back-propagation standard ANN classier is considered consisting of three
layers: input, hidden and output layer. The learning set consists of the normalised longitude, latitude,
elevation and the normalised mean annual and monthly solar irradiation of 41 Moroccan sites. The
testing set consists of patterns just represented by the input component, while the output component is
left unknown and its value results from the ANN algorithm for that specic input. The results are given
in the form of the annual and monthly maps. They indicate that the method could be used by
researchers or engineers to provide helpful information for decision makers in terms of sites selection,
design and planning of new solar plants.
& 2012 Elsevier Ltd. All rights reserved.
Keywords:
Solar energy potential
Articial neural network
Geographic information system
Morocco
Contents
1.
2.
3.
4.
5.
6.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4876
Solar irradiation data sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4877
Renewable energy development in Morocco . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4878
Articial neural network modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4879
Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4880
5.1.
Solar data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4880
5.2.
ANN predictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4881
5.3.
Annual and monthly solar irradiation maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4881
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4885
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4888
1. Introduction
Solar energy is considered as one of the most valuable renewable
energy alternative. The exploitation of solar resources is well suited
to cope with the limitations of current patterns of energy generation
and consumption and to complement existing energy production
systems. Solar generation systems can be considered as an attractive
option to conventional power generation as well as to enhance
sustainable development especially in developing countries like
n
Correspondence to: Unite des Technologies et Economie des Energies
Renouvelables (TEER), CNRST B.P 8027 NU Rabat, Morocco. Tel.: 212 666892377.
E-mail addresses: ouammi@cnrst.ma,
ouammi.ahmed@gmail.com (A. Ouammi).
1364-0321/$ - see front matter & 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.rser.2012.03.071
4877
Table 1
Geographical coordinates of the 41 Moroccan sites.
Code
Site
Longitude
Latitude
Altitude (m)
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
28
29
30
31
32
33
34
35
36
37
38
39
40
41
Agadir
Ain bnimathar
Sa
Azilal
Bouarfa
Boujdour
Dakhla
El jadida
Errachidia
Esmara
Essaouira
Fes
Glyab
Imintanout
Kenitra
Khmisset
Khenifra
Khouribga
Laayoune
Lagouira
Larache
Marrakech
Msaysat
Nador
Ouarzazat
Oujda
Ousard
Ouazane
Chtoukane
Tanger
Tantan
Taounat
Taourirt
Tarfaya
Taroudant
Tata
Taza
Tetouan
Tinghir
Tiznit
Zagora
9.602
2.038
9.245
6.575
1.978
14.424
15.946
8.514
4.438
11.678
9.772
5.010
13.084
8.860
6.586
6.064
5.674
6.905
13.211
17.073
6.163
8.003
15.820
2.944
6.899
1.917
14.326
5.592
14.776
5.801
11.113
4.653
2.906
12.931
30.468
7.981
4.010
5.372
5.548
9.741
5.856
30.396
34.012
32.310
31.971
32.528
26.136
23.700
33.234
31.952
26.746
31.518
34.034
21.289
31.180
34.266
33.797
32.944
32.889
27.157
21.064
35.183
31.634
23.161
35.165
30.921
34.696
22.553
34.800
24.637
35.769
28.435
34.538
34.420
27.936
8.888
29.740
34.225
35.577
31.533
29.707
30.335
0
914
14
1331
1128
0
9
24
1050
178
0
410
0
907
4
460
854
805
0
0
40
455
135
29
1133
547
306
289
129
12
48
582
389
3
232
741
487
172
1314
231
730
4878
system itself, the sensor may be covered with dirt, frost, or snow, or
that the sensor is shadowed by nearby trees or buildings for some of
the time during the year. These problems can be removed by careful
sitting and maintenance, but it makes it more uncertain to use data
where you do not have direct experience with the measurements
[25]. In the case of lack or unavailability of direct measurements
data, it is possible to estimate the solar irradiation from measurements made nearby. It is also possible to combine data from
several different measurement locations to make an estimate for
the solar irradiation in a place somewhere between the measurement sites [25].
Numerous methods can be used to estimate the solar irradiation at ground level using data from satellites. Various types of
satellites can be used: Geostationary weather satellites take
pictures of the Earth at short intervals (every 15 or 30 min) so
have a very good time resolution. However, each pixel in the
picture typically represents a rectangle of a few km on each side,
so the estimate of solar irradiation for each pixel will be the
average of such an area. Polar-orbiting satellites y closer to
the Earth, so the space resolution is better. However, they do not
stay permanently above a particular area, so they are normally
able to take only a couple of pictures a day of a given area [25].
Generally, goodness and quality of satellite-based estimates must
be checked by comparison with high quality ground station
measurements [25].
Table 2
Monthly solar irradiation data.
Code
Jan
(Wh/m2/day)
Feb
(Wh/m2/day)
Mar
(Wh/m2/day)
Apr
(Wh/m2/day)
May
(Wh/m2/day)
Jun
(Wh/m2/day)
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
28
29
30
31
32
33
34
35
36
37
38
39
40
41
3720
3060
3430
3260
3370
4220
4560
3310
3310
4100
3650
3010
4620
3510
3100
3070
3070
3200
3810
4830
3020
3400
3980
3100
3450
2930
4730
2870
3830
2800
3790
2790
2980
4020
3430
3730
2940
2750
3280
3730
3600
4560
3920
4330
4050
4330
5020
5360
4240
4130
4920
4500
3790
5310
4310
3970
3840
3820
3950
4550
5660
3990
4140
4700
4040
4290
3780
5530
3620
4560
3740
4480
3400
3760
4760
4230
4580
3740
3510
4090
4520
4440
5750
5370
5700
5230
5700
5970
6280
5690
5510
5770
5830
5280
6140
5440
5600
5250
5080
5270
5430
6460
5770
5280
5450
5420
5490
5130
6310
5300
5360
5350
5370
4900
5140
5750
5350
5800
5150
5030
5430
5570
5770
6660
6510
6700
6140
6980
6870
7120
6700
6590
6830
6830
5920
6900
6560
6460
5970
5950
6150
6340
7180
6620
6160
6190
6770
6770
6290
7250
5860
6210
6380
6060
5740
6270
6480
6280
6920
5970
6030
6630
6450
6820
7250
7340
7460
6830
7490
7290
7340
7580
7070
7370
7670
6760
7130
7240
7430
6800
6720
6920
6770
7230
7560
6830
6430
7340
7380
6990
7450
6790
6510
7210
6270
6570
6950
6970
7000
7530
6830
6830
7250
6870
7350
7090
7800
7230
6550
7740
7430
7560
7140
7300
7850
7030
7630
7310
7410
6880
6640
6500
7130
7310
7530
6940
7410
7740
7620
7640
7690
7710
7090
7830
6800
6990
7260
7140
6880
7260
7720
7580
6570
7570
7050
7480
4879
yi
1
1 exi
Table 3
Monthly and annual solar irradiation data (continue).
Code
Jul
(Wh/m2/day)
Aug
(Wh/m2/day)
Sep
(Wh/m2/day)
Oct
(Wh/m2/day)
Nov
(Wh/m2/day)
Dec
(Wh/m2/day)
Annual
(Wh/m2/day)
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
28
29
30
31
32
33
34
35
36
37
38
39
40
41
6620
7660
7410
6250
6690
7130
7170
7080
7030
7880
7190
7650
7990
7460
7280
7030
6260
7170
7510
7210
6870
7310
7480
7130
7530
7650
7600
6630
7640
6650
6590
6640
6540
6910
7020
7410
7310
6520
7230
6820
7110
6180
6730
6850
6490
5790
6870
6870
7030
6280
7000
6870
6840
6620
6940
7120
6860
6660
6740
6470
6850
7250
6550
6090
7140
6470
6870
6990
7080
6230
7290
5730
6850
6750
6530
6280
6600
6690
7090
6380
6320
6410
5460
5720
5920
5550
5790
6160
6240
6100
5440
6010
5920
5600
6040
5830
5960
5600
5560
5520
5670
6420
6140
5350
5440
6030
5650
5670
6230
5700
5490
5920
5420
5450
5570
5970
5210
5760
5480
5670
5740
5470
5690
4160
5020
4360
3950
4400
5260
5620
4130
4470
5470
4030
4100
5510
4490
4380
4140
4150
4120
5400
5710
3930
4550
5310
4470
4980
4230
5400
4040
5310
4020
5200
3860
4030
5150
4280
4890
4050
3980
4480
4580
4580
3800
4020
3580
3120
3740
4440
4840
3310
3520
4660
3600
3150
4850
3650
3310
3200
3220
3280
4410
5160
3230
3780
4510
3300
3870
3120
4790
3020
4070
3040
3940
2840
3090
4220
3480
3750
3040
2940
3570
3730
3700
3340
2810
3120
2950
3010
3890
4230
2990
2970
3730
3290
2730
4220
3150
2800
2770
2760
2840
3520
4600
2690
3010
3700
2770
3160
2670
4330
2600
3570
2490
3440
2460
2720
3660
3090
3310
2650
2460
3070
3730
3190
5440
5370
5560
5290
5540
5880
6090
5600
5280
5840
5680
5210
5950
5570
5500
5240
5160
5270
5430
6270
5580
5250
5350
5540
5480
5250
6190
5220
5330
5390
5070
5000
5210
5570
5260
5650
5140
5190
5400
5370
5520
4880
where Op,j is the observed output on unit j for learning pattern p and
Dp,j is the desired output.
The ANN learning procedure is performed on learning set of
patterns, where, in our model, each learning pattern p is represented by three parameters (input layer) and by one output
parameter (output layer).
Table 4
Predictions of the annual mean solar irradiation obtained by the ANN kriging method learning on 41 sites and testing on the
same data.
Code
Site
Error
(Wh/m2/day)
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
28
29
30
31
32
33
34
35
36
37
38
39
40
41
Agadir
Ain bnimathar
Sa
Azilal
Bouarfa
Boujdour
Dakhla
El jadida
Errachidia
Esmara
Essaouira
Fes
Glyab
Imintanout
Kenitra
Khmisset
Khenifra
Khouribga
Laayoune
Lagouira
Larache
Marrakech
Msaysat
Nador
Ouarzazat
Oujda
Ousard
Ouazane
Chtoukane
Tanger
Tantan
Taounat
Taourirt
Tarfaya
Taroudant
Tata
Taza
Tetouan
Tinghir
Tiznit
Zagora
5440
5370
5560
5290
5540
5880
6090
5600
5280
5840
5680
5210
5950
5570
5500
5240
5160
5270
5430
6230
5580
5250
5350
5540
5480
5250
6190
5220
5330
5390
5070
5030
5210
5570
5260
5650
5140
5190
5400
5370
5520
5451
5401
5457
5283
5496
5674
5852
5451
5306
5610
5581
5168
5916
5452
5428
5288
5210
5282
5402
6218
5463
5222
5277
5478
5432
5218
5941
5263
5279
5433
5161
5162
5238
5608
5282
5525
5173
5165
5422
5432
5560
11
31
103
7
44
206
238
149
26
230
99
42
34
118
72
48
50
12
28
12
117
28
73
62
48
32
249
43
51
43
91
132
28
38
22
125
33
25
22
62
40
0.20
0.59
1.84
0.14
0.79
3.51
3.91
2.66
0.49
3.94
1.75
0.80
0.57
2.13
1.30
0.92
0.97
0.23
0.52
0.19
2.10
0.53
1.37
1.12
0.88
0.60
4.03
0.83
0.96
0.80
1.80
2.62
0.53
0.68
0.41
2.21
0.65
0.48
0.40
1.16
0.72
Average
71
1.3
4881
4882
Fig. 2. Annual mean solar irradiation map of Morocco as obtained by the ANN kriging technique.
Fig. 3. Solar irradiation map of Morocco (October) as obtained by the ANN kriging technique.
4883
Fig. 4. Solar irradiation map of Morocco (November) as obtained by the ANN kriging technique.
Fig. 5. Solar irradiation map of Morocco (December) as obtained by the ANN kriging technique.
4884
Fig. 6. Solar irradiation map of Morocco (January) as obtained by the ANN kriging technique.
Fig. 7. Solar irradiation map of Morocco (February) as obtained by the ANN kriging technique.
4885
Fig. 8. Solar irradiation map of Morocco (March) as obtained by the ANN kriging technique.
Fig. 9. Solar irradiation map of Morocco (April) as obtained by the ANN kriging technique.
This shows that ANN models are more accurate and versatile to
predict solar irradiation.
6. Conclusion
Solar power generation systems can be considered as an
attractive option to conventional power generation as well to
enhance sustainable development especially in developing countries like Morocco which traditionally lacks fossil fuels, but has an
important environmental wealth, land availability and great solar
potential. Data have been inferred using an ANN algorithm to
establish a forward/reverse correspondence between the longitude, latitude, elevation and the mean annual and monthly solar
irradiation. The learning set consists of the normalised longitude,
4886
Fig. 10. Solar irradiation map of Morocco (May) as obtained by the ANN kriging technique.
Fig. 11. Solar irradiation map of Morocco (June) as obtained by the ANN kriging technique.
algorithm for that specic input. In our approach, the testing set
consists of the same 41 sites. The obtained results indicate that
the method could be used by researchers or engineers to provide
helpful information for decision makers in terms of sites selection, design and planning of new solar plants. Future research will
4887
Fig. 12. Solar irradiation map of Morocco (July) as obtained by the ANN kriging technique.
Fig. 13. Solar irradiation map of Morocco (August) as obtained by the ANN kriging technique.
4888
Fig. 14. Solar irradiation map of Morocco (September) as obtained by the ANN kriging technique.
Table 5
Predictions of the monthly mean solar irradiation.
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
Highest solar
irradiation data
(Wh/m2/day)
Highest solar
Error
irradiation predictions (Wh/m2/day)
(Wh/m2/day)
4830
5660
6460
7250
7670
7850
7990
7290
6420
5710
5160
4600
4799
5600
6430
7200
7600
7824
7710
7228
6370
5668
5100
4536
31
60
30
50
70
26
280
62
50
42
60
64
0.6
1.1
0.5
0.7
0.9
0.3
3.5
0.9
0.8
0.7
1.2
1.4
2750
3400
4900
5740
6270
6500
6250
5730
5210
3860
2840
2460
40
40
70
52
10
149
287
38
100
107
30
56
1.5
1.2
1.4
0.9
0.2
2.3
4.6
0.7
1.9
2.8
1.1
2.3
69
Average
82
1.7
Average
References
[1] Fadare DA. Modelling of solar energy potential in Nigeria using an articial
neural network model. Applied Energy 2009;86:141022.
[2] Alam S, Kaushik SC, Garg SN. Assessment of diffuse solar energy under general
sky condition using articial neural network. Applied Energy 2009;86:55464.
[3] Soteris AK. Articial neural networks in renewable energy systems applications: a review. Renewable Sustainable Energy Reviews 2001;5:373401.
[4] Ermis K, Midilli A, Dincer I, Rosen MA. Articial neural network analysis of
world green energy use. Energy Policy 2007;35:173143.
[5] Mabel MC, Fernandez E. Analysis of wind power generation and prediction
using ANN: a case study. Renewable Energy 2008;33:98692.
[6] Bilgili M, Sahin B, Yasar A. Application of articial neural networks for the
wind speed prediction of target station using reference stations data. Renewable Energy 2007;32:235060.
[7] Fatih O, Hocaoglu, Mehmet Fidan, Omer N. Gerek Mycielski approach for
wind speed prediction. Energy Conversion and Management 2009;50:
143643.
[8] Tarek HMEl-Fouly, Ehab FEl-Saadany, Magdy MASalama. One day ahead
prediction of wind speed and direction. IEEE Transactions on Energy Conversion 2008;1:23.
2790
3440
4970
5792
6280
6649
6537
5768
5310
3967
2870
2516
[9] Riahy GH, Abedi M. Short term wind speed forecasting for wind turbine
applications using linear prediction method. Renewable Energy 2008;33:3541.
[10] Pourmousavi Kani SA, Ardehali MM. Very short-term wind speed prediction:
a new articial neural network-Markov chain model. Energy Conversion and
Management 2011;52:73845.
[11] Kusiak A, Li W. Virtual models for prediction of wind turbine parameters.
IEEE Transactions on Energy Conversion 2010;1:25.
[12] Ouammi A, Sacile R, Zejli D, Mimet A, Benchrifa R. Sustainability of a wind
power plant: application to different Moroccan sites. Energy 2010;35:
422636.
[13] Dagdougui H, Ouammi A, Sacile R. A regional decision support system for
onsite renewable hydrogen production from solar and wind energy sources.
International Journal of Hydrogen Energy 2011;36(22):1432434.
[14] Koca A, Oztop HF, Varol Y, Koca GO. Estimation of solar irradiation using
articial neural networks with different input parameters for Mediterranean
region of Anatolia in Turkey. Expert Systems with Applications 2011;38:
875662.
[15] Almonacid F, Rus C, Perez PJ, Hontoria L. Estimation of the energy of a PV
generator using articial neural network. Renewable Energy 2009;34:
274350.
[16] Senkal Ozan. Modeling of solar irradiation using remote sensing and articial
neural network in Turkey. Energy 2010;35:4795801.
[17] Mubiru J, Banda EJKB. Estimation of monthly average daily global solar
irradiation using articial neural networks. Solar Energy 2008;82:1817.
[18] Reikard G. Predicting solar irradiation at high resolutions: a comparison of
time series forecasts. Solar Energy 2009;83:3429.
[19] Rosiek S, Batlles FJ. Modelling a solar-assisted air-conditioning system
installed in CIESOL building using an articial neural network. Renewable
Energy 2010;35:2894901.
[20] Soares J, Oliveira AP, Boznar MZ, Mlakar P, Escobedo JF, Machado AJ.
Modeling hourly diffuse solar-irradiation in the city of Sao Paulo using a
neural-network technique. Applied Energy 2004;79:20114.
[21] Miguel AF, Silva A. Solar irradiation in diffusely enclosures with partitions.
Applied Energy 2010;87:83642.
[22] Dorvloa ASS, Jervase JA, Al-Lawati A. Solar irradiation estimation using
articial neural networks. Applied Energy 2002;71:30719.
[23] Sozen
A, Arcaklioglu E, ozalp
ME, Kanit G. Use of articial neural networks for
mapping of solar potential in Turkey. Applied Energy 2004;77:27386.
[24] Sen Z. Solar energy in progress and future research trends. Progress in Energy
and Combustion Science 2004;30(4):367416.
4889
[25] /http://re.jrc.ec.europa.eu/pvgis/S.
[26] Benkhadra A. Does Morocco provide a new model for bridging old and new
energy systems? Moroccos Annual Investment Conference. London, 9 November,
2009.
/http://www.mem.gov.ma/Actualites/2009/Novembre/Pdf/London_
speech.pdfS.
[27] Zejli D, Bennouna A. Wind energy in Morocco: which strategy for which
development? In: Mason M, Mor A, editors. Renewable energy in the Middle
East. Dordrecht: Springer; 2009. p. 15173.
[28] Benkhadra A. Strategie energetique: Bilan detape. 2e mes Assises Energetiques. Ministe re de lEnergie, des Mines, de lEau et de lEnvironnement. May
2011.