Académique Documents
Professionnel Documents
Culture Documents
10166-10172
© Research India Publications. http://www.ripublication.com
Hussein A Kazem
Associate Professor, Faculty of Engineering, Sohar University PO Box 44, Sohar, PCI 311 Oman.
Jabar H. Yousif
Assistant Professor, Faculty of Computer & Information Technology, Sohar University PO Box 44, Sohar, PCI 311 Oman.
Miqdam T Chaichan
Assistant Professor, Department of Mechanical Engineering University of Technology Baghdad, Iraq.
Abstract the generation rates cause instability in the power grid. Thus,
Solar energy is the most important renewable energy source, to overcome this problem, the energy is generated at the time
which could be utilized through photovoltaic or thermal availability of natural resources is and the excess energy is
systems. This paper aims to design and implement a machine
stored on the need for later use. At this point, the problem of
learning technique called Support Vector Machine (SVM) for
the management of energy generation based on experimental finding an efficient management of these sources of electric
work. The SVM model is consist of two inputs and one power generation is more needed [3]. Solar energy could be
output. The inputs are solar radiation, and ambient utilized through photovoltaic PV or thermal systems. Solar
temperature; the output is the photovoltaic current. The energy technology and especially solar cells and photovoltaic
practical implementation of the proposed SVM model is have been experimented for more than 70 years and it shows
achieved a final MSE of (0.026378744) in training phase and good success. The PV system prices get down and at the same
(0.035615759) in cross validation phase. The experiments
time, the PV efficiency is improved [4].
achieved a value of (R squared) equal to 0.0774 which
indicates the predicting model is very close to the regression
Machine learning is a subdivision of artificial intelligence that
line and a well data fitting to the statistical model. Besides, the
proposed model achieved less MSE in comparison with other is developed as a result of studies in pattern classification and
related work. recognition for finding mathematical models for various real
life problems. Machine learning investigates the construction
Keywords: Solar energy prediction; machine learning; of algorithms that can learn from the previous data and help in
support vector machine; Oman. finding a forecast on the data in the current and future time.
The machine learning applied iterative and interactive
statistical methods in the construction of computational
INTRODUCTION
Currently the subject of generating energy from renewable models to obtain the desired results [5]. The factors like
sources is considering as one of the important topics that are efficiency of learning algorithms, the complexity of the
growing promptly as a result of its many benefits. The current problem, methods of representation of data are the most
step of technological development makes the extraction of important factors affect the accuracy of the results and future
renewable energy from various sources such as the sun, wind, forecasting for data.
geothermal energies, and many other sources commercially
Several machine learning methods are used for designing and
viable process [1].
implementing different phases of a renewable energy power
Renewable energy started competing with non-renewable systems based on the problem requirements and its
energy sources due to many reasons (i.e., fluctuation of oil characteristics [6, 7 & 8]. Hence, the adaptation of optimal
prices, independency, reduction in renewable energy location and structures of renewable power plants is one of the
technology prices and improvement in efficiency, etc.) [2]. important implementation of learning machine methods.
Renewable energy generation capacity significantly affected Recently, the Support Vector Machines (SVMs) have been
by fluctuations in the atmosphere for example, solar energy widely implemented into several problems of renewable
generated in the morning for the presence of sunlight during energy power systems [9, 10, 11, and 12].
the day and the energy generated from the wind depends on
the wind speed and geographic location. These fluctuations in
10166
International Journal of Applied Engineering Research ISSN 0973-4562 Volume 11, Number 20 (2016) pp. 10166-10172
© Research India Publications. http://www.ripublication.com
10167
International Journal of Applied Engineering Research ISSN 0973-4562 Volume 11, Number 20 (2016) pp. 10166-10172
© Research India Publications. http://www.ripublication.com
EXPERIMENTAL SETUP
1. Practical measurement
In the current study, 24 PV modules have been installed in
Sohar University in Oman. The rating of PV module is 140W Figure. 2. (b) Meteorological station
as shown in Table 1. Three PV systems configurations;
standalone, grid connected and tracking systems have been
designed and evaluated as shown in Figure 2 (a). The
meteorological data has been measured and recorded using
weather station as shown in Figure 2 (b). The grid connected
system data (voltage, current, power and energy) has been
recorded and monitored. Also, the measured environmental
parameters are linked to the productivity of the PV system.
This research is a part of a research on grid connected PV
system (see Figure 3) and consequently the production of the Figure. 3. Solar photovoltaic grid connected system
PV system here is needed.
2. Machine Learning
Table 1: Installed solar system specifications The Machine Learning is one of the applications of artificial
intelligence to simulate the thinking capabilities of human
PV array [23]. It is concerning with the techniques and methods that
help the machine to learn. Machine learning is developed as a
PV panels (24 panels) 140 Wp (3.36 kWp) result of studies in pattern classification and recognition for
“Maximum voltage” 17.7 finding mathematical models for various problems. Machine
“Maximum current” 7.91 learning investigates the construction of algorithms that can
“Open circuit voltage” 22.1 learn from the previous data and help in finding a forecast on
the data in the current and future time. The machine learning
“Short circuit current” 8.68 applied iterative and interactive statistical methods in the
“Efficiency” 13.9% construction of computational models to obtain the desired
Temperature coefficient of Vo.c -0.36 %/k results [24]. Several techniques and methodologies are
established for machine learning tasks. The new development
Temperature coefficient of Is.c 0.06 %/k of machine learning techniques is mainly utilizing Kernel-
based methods such as support vector machines (SVM),
Gaussian processes, etc. [6]. A SVM is considered as a
supervised learning technique which is presented by Refs. [23
& 25] in COLT-92 as presented in Figure 4. It is considering
as a classification and regression prediction techniques. SVM
is deployed in many applications, such as face analysis and
detection, hand writing recognition, pattern classification and
regression etc. A classification technique comprises of
training and testing data attributes. The main benefits of SVM
are minimizing the classification error by implement an
iterative training algorithm and maximizing the margin
between two hyper-planes. Figure 5 show the classification
Hyper-planes (wx+b>1 and wx+b<-1). The SVM separates
the cases of different class labels of hyper-planes of
Figure. 2. (a) Grid connected PV system
multidimensional space into binary cases (Positive, Negative)
as illustrated in Figure 5.
So, the positive zone is defined by (w. x+ + b = +1), while the
negative zone is defined by (w. x- + b = -1). If the negative
zone is subtract from the positive zone, then w. (x+-x-) = 2.
Now, the margin M is defined by equation 1.
( x x ) w
M (1)
w
10168
International Journal of Applied Engineering Research ISSN 0973-4562 Volume 11, Number 20 (2016) pp. 10166-10172
© Research India Publications. http://www.ripublication.com
The maximum margin linear classifier is defined as in training phase are used for the sake of maximizing the chance
equation 2. of classification and prediction of data. The Gaussian Axon is
2 implemented as a transfer function.
Maximize M (2)
w
Formulate a Quadratic Optimization Problem in order to solve
for w and b as defined in equation 3.
Minimize
1
𝜑 (𝑊) = 𝑤 𝑡 𝑤 (3)
2
Subjected to 𝑦𝑖 (𝑤𝑥𝑖 + 𝑏) ≥ 1 𝑓𝑜𝑟 ∀𝑖
Then the classifying function is defined as in equation 4.
𝑓(𝑥) = ∑ 𝛼𝑖 𝑦𝑖 𝑋𝑖𝑇 𝑋 + 𝑏 (4)
10169
International Journal of Applied Engineering Research ISSN 0973-4562 Volume 11, Number 20 (2016) pp. 10166-10172
© Research India Publications. http://www.ripublication.com
0.5
0.4
0.3
MSE
0.2
0.1
0
1 100 199 298 397 496 595 694 793 892 991
Epoch
Figure 7. The graph of training and cross validation MSE Figure 9. The graph of trend line of output and desired
results.
Desired Output and Actual Network The future individual values prediction could be shown
Output through the MSE efficiency of the developed model (SEE
Table 2). A large positive MSE means a large deviation in the
O_Upv_soll O_Upv_soll Output
predicted value from the real value. The developed model
500 found to be able to predict the individual values since its MSE
450 is 0.035. Also, Table 3 illustrates a comparison of MSE for
400
the proposed model and models in literature, which shows that
350
the proposed model is more accurate than other models in
300
Output
250 China, UAE and Malaysia. Also, it is found that SVM is more
200 efficient in comparison with other ANN models.
150
100
50
Table 2: The Minimum MSE & Final MSE of SVM network
0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29
Best Networks Training Cross Validation
Exemplar
Epoch # 1000 1000
Minimum MSE 0.026378744 0.035615759
Figure 8. The graph of training and cross validation MSE Final MSE 0.026378744 0.035615759
y= -0.0001 x4 + 0.0307 x3 – 1.269 x2 + 15.557 x +130.38 (6)
The predicting system is generating an output closely to the Table 3: MSE comparison of the proposed model and models
desired output y1. The desired output is defined as in equation in literature
7. Reference Model type Site MSE
Proposed model SVM Oman 0.026-0.035
y1= -0.00001x4 +0.0255x3 –1.257x2 +16.969x +115.08 (7) J. Zeng et al, SVM, SVM China 0.060-0.137
2013 [8] and Wavelet
The coefficient of determination (R) is used to measure how
Tamer et al, FFMLP Malaysia 0.330-0.530
close the data from the regression line. Usually determine its 2011 [26]
value between 0 and 1, and the closer the value of 1 gave an Tamer et al, FFMP Malaysia 0.342-0.419
indication that the data correspond to the regression line. The 2012 [27]
experiment got a value of (R squared) equal to 0.0774 which Maitha et al, MLP UAE 0.350
indicates the predicting model is very close to the regression 2013 [28]
line and a well data fitting to the statistical model.
10170
International Journal of Applied Engineering Research ISSN 0973-4562 Volume 11, Number 20 (2016) pp. 10166-10172
© Research India Publications. http://www.ripublication.com
CONCLUSIONS AND FUTURE WORK [7] Zhao P, Xia J, Dai Y and He J, Wind speed
A prediction model of solar photovoltaic system using SVM prediction using support vector regression Industrial
has been proposed for Oman. This prediction was based on Electronics and Applications (ICIEA), pp. 882 –886,
collected data from installed system in location called Sohar. 2010.
Two inputs solar radiation and ambient temperature has been [8] Zeng J and Qiao W, Short-term solar power
used and one output, photovoltaic current. The developed prediction using a support vector
model has been evaluated using MSE. Results show that the machine, Renewable Energy, vol. 52, pp. 118 – 127,
predicting model is very close to the regression line and a well 2013.
data fitting to the statistical model. The MSE found to be [9] Banda d, Peña R, Gutiérrez G, Juárez E, Visairo N,
2.6%, which is more accurate in comparison with some Núñez C, Feasibility assessment of the installation of
studies in literature. a photovoltaic system as a battery charging center in
For the sake of reducing the MSE for the proposed network, a Mexican Mining Company, Power, Electronics and
alternative machine learning techniques need to be Computing (ROPEC), 2014 IEEE International
implemented for discovering the possibility of reducing the Autumn Meeting on Power, Electronics and
MSE rate like MLP with different hidden layers, SOFM, etc. Computing (ROPEC, 2014), 5 pages, 2014.
The sensitivity analysis need to be implemented to determine [10] Caudill M and Butler C, Understanding neural
the effects of the input variables on the output of network. networks: Computer explorations volume 1: Basic
networks. The MIT Press, Cambridge, Mass, USA,
1993.
[11] Sharma N, Sharma P, Irwin D and Shenoy P,
ACKNOWLEDGMENT Predicting solar generation from weather forecasts
The research leading to these results has received Research using machine learning. Proceedings of the 2011
Project Grant Funding from the Research Council of the IEEE International Conference on Smart Grid
Sultanate of Oman, Research Grant Agreement No. ORG SU Communications, pp. 528-533, 2011.
EI 11 010. The authors would like to acknowledge support [12] Hossain M R, Amanullah M T O, Ali A B M S,
from the Research Council of Oman. Hybrid prediction method of solar power using
different computational intelligence algorithms.
Proceedings of the 22nd Australasian Universities
REFERENCES Power Engineering Conference, pp. 1-6, 2012.
[13] Ministry of Tourism,
[1] Chaichan M T, Mohammed B A and Kazem H A, http://www.omantourism.gov.om
Effect of pollution and cleaning on photovoltaic [14] Chaichan M T, Kazem H A, Experimental analysis
performance based on experimental study, of solar intensity on photovoltaic in hot and humid
International Journal of Scientific and Engineering weather conditions, International Journal of
Research, vol. 6, No. 4, pp. 594-601, 2015. Scientific & Engineering Research, vol. 7, No. 3, 91-
[2] Al-Maamary H M S, Kazem H A, Chaichan M T, 96, 2016.
Changing the energy profile of the GCC States: A [15] Kazem H A, ChaichanM T, Saif S A, Dawood A A,
review, International Journal of Applied Engineering Salim S A, Rashid A A and Alwaeli A A,
Research (IJAER), vol. 11, No. 3, pp. 1980-1988, Experimental investigations of dust type effect on
2016. photovoltaic systems in north region, Oman,
[3] Chaichan M T, Kazem H A, Kazem A A, Abaas Kh International Journal of Scientific & Engineering
I, Al-Asadi K A H, The effect of environmental Research, vol. 6, No. 7, pp. 293-298, 2015.
conditions on concentrated solar system in desertec [16] H.A. Kazem, Renewable energy in Oman: Status and
weathers, International Journal of Scientific and future prospects, Renewable and Sustainable Energy
Engineering Research, vol. 6, No. 5, pp. 850-856, Reviews, vol. 15, pp. 3465–3469, 2011.
2015. [17] H.A. Kazem, T. Khatib, Techno-economical
[4] Kazem H A, Al-Waeli A H A, Al-Mamari A S A, assessment of grid connected photovoltaic power
Al-Kabi A H K, Chaichan M T, A photovoltaic systems productivity in Oman, Sustainable Energy
application in car parking lights with recycled Technologies and Assessments, vol. 3, pp. 61–64,
batteries: A techno-economic study, Australian 2013.
Journal of Basic and Applied Science, vol. 9, No. 36, [18] Kazem H A and Chaichan M T, Effect of humidity
pp. 43-49, 2015. on photovoltaic performance based on experimental
[5] Chen J L, Liu H B and Wu W, Estimation of monthly study, International Journal of Applied Engineering
solar radiation from measured temperatures using Research (IJAER), vol. 10, No. 23, pp. 43572-43577,
support vector machines. Renewable Energy, vol. 36, 2015.
No.1, pp. 413–420, 2011. [19] Authority for Electricity Regulation in Oman (2008).
[6] Schölkopf B, and Smola A J, Learning with Kernels: Study on Renewable Resources. Final Report, Oman.
Support Vector Machines, Regularization, [20] Public Authority for Electricity and Water
Optimization, and Beyond, MIT Press, Cambridge, [21]. Kazem H A, KhatibT and Sopian K, Sizing of a
MA, USA, 2011. standalone photovoltaic/ battery system at minimum
10171
International Journal of Applied Engineering Research ISSN 0973-4562 Volume 11, Number 20 (2016) pp. 10166-10172
© Research India Publications. http://www.ripublication.com
10172