Researcher 2017;9(12) http://www.sciencepub.

net/researcher

Fuzzy logic based modeling and estimation of global solar energy using meteorological parameters

Babak Mohammadi

Department of water Engineering, university of Tabriz, Tabriz. Iran

[email protected]

Abstract: Global solar energy data is considered as the most important parameter in smart grid applications,
particularly for sizing the photovoltaic system and demand driven supply. However the data of global solar energy is
rarely available on hourly basis, even for those stations where measurement has already been done. Due to lack of
such measured data, the estimation of global solar energy at the earth’s surface is an important study in the present
scenario to meet the energy requirement from green energy sources. This paper is based on fuzzy logic approach for
modeling and estimating the global solar energy using mean duration sunshine per hour, temperature, latitude,
longitude, altitude and months of the year as input parameters. Fortunately, these important for accurate parameters
estimation of solar energy are commonly available. Results obtained from fuzzy logic approach are used for the
prediction of SPV system output for smart grid application.
[Babak Mohammadi. Fuzzy logic based modeling and estimation of global solar energy using meteorological
parameters. Researcher 2017;9(12):91-95]. ISSN 1553-9865 (print); ISSN 2163-8950 (online).
http://www.sciencepub.net/researcher. 13. doi:10.7537/marsrsj091217.13.

Keywords: Fuzzy; logic based modeling; estimation; global solar energy; meteorological parameter

1. Introduction estimate the accurately using

The rapid depletion of fossil fuel reserves and mathematical/regression/stochastic models. Therefore,
large greenhouse emissions are becoming the major due to uncertainty in weather conditions, fuzzy logic
challenge for the today’s power sector. Due to based models are proposed by researchers to estimate
increasing population and urbanization, the demand of the solar energy at a given location using different
electricity is increasing exponentially. Under such meteorological pa- rameters [12e18]. The uncertainty
circumstances it is highly desirable to adopt the in atmosphere may occur due to the existence of the
environment friendly energy sources for power following: dust, moisture, aerosols, clouds, or
generation. In order to achieve the target, renewable temperature differences in the lower atmosphere.
energy sources particularly solar and wind sources can Among these factors, clouds can cause the maximum
play the important role. In addition, solar energy is one losses in the extraterrestrial solar energy reaching at
of the most promising and more predictable than other earth’s surface. The atmosphere causes a reduction of
renewable energy sources. Therefore, high penetration the extraterrestrial solar input by about 30% on a very
of solar technologies into electric power system can clear day to nearly 100% on a very cloudy day [13].
meet the desired goals. It is well known fact that, India Further, the use of solar energy estimation in smart
is located on the equatorial sun belt of the earth has grid application is rarely available in the literature.
great potential of solar energy with an average of Keeping in view of aforesaid fuzzy logic based model
about 300 solar days per year. Further, the government for the estimation of global solar energy has been
is also providing the in- centives and other benefits for developed considering latitude, longitude, altitude of
installing the power plants based on solar energy. the location, months of the year, mean duration
Hence, to exploit the environment friendly source of sunshine per hour (it is the ratio of average daily actual
energy, analysis and estimation of solar energy at sunshine duration at the location to the theoretical
different stations of the country is utmost important sunshine duration), and temperature as input parame-
[1e3]. Keeping in view of the aforesaid, a thorough ters. Obtained results are further simulated for smart
literature review is carried out and it is found that grid appli- cations using fuzzy logic approach/toolbox.
number of mathematical models for the estimation of Nomenclature
global solar energy under cloudless skies is available D day of the year
in the literature [4e5]. Regression models and H monthly mean daily irradiance on horizontal
stochastic models for estimating the solar energy are surface
also presented by researchers [6e11]. The results H0 mean clear sky daily irradiance
obtained using these models were satisfactory, but H/Ho clearness index
applicable only for clear sky weather condition. In Hmeasured measured value of monthly mean
India around 300 days in a year are clear sky and daily irradiance on horizontal surface
remaining days are cloudy, so it is very difficult to

91
 Researcher 2017;9(12) http://www.sciencepub.net/researcher

Hestimated predicted value of monthly mean requirements to support overall development, hence
daily irradiance on horizontal surface the opportu- nities for building smart grid is immense
Is short circuit current of PV module at STC (A) [3,19e20].
Isc solar constant ¼ 1367 W/m2 Smart grid is used to predict and intelligently
Pmax maximum power of module at MPP respond to the behavior and actions of all users
S monthly mean daily hours of bright sunshine connected to it. Further, it is used to efficiently deliver
So monthly mean of maximum possible daily the reliable, economic, and sustainable final customer.
hours of bright sunshine The main advantage of smart grid implementation in
S/So mean fraction possible sunshine terms of utility benefits include reduced perturbations
hours and outages; minimal power losses and blackout
T monthly mean hourly temperature ( C) prospects; lower maintenance and operational cost;
To monthly mean hourly maximum possible lower greenhouse gas emis- sions; increased energy
temperature efficiency; increased large scale renew- able energy
( C) and distributed generation integration; enabled micro-
T/To ratio of monthly mean hourly grid applications and energy management systems
temperature to monthly mean hourly maximum envi- ronmental benefits and economic growth through
possible temperature clean power markets [19].
Ta ambient temperature ( C) In the present grid system generation is following
Tcell cell temperature ( C) the load, but in the smart grid system the load will
Tstc temperature of PV module at STC follow the generation. For such a system, where load
Voc open circuit voltage of PV module at follows the generation the solar energy estimation at
STC the site is the driving contributor to the power output
Greek symbols calculations for these systems. If a system is able to
u hour angle predict solar energy with a good accuracy at a
us hour angle at sunset particular location then load scheduling, economic
f latitude of the location load dispatch, battery sizing, time of use and pricing
d declination angle can be done intelligently and optimally including the
hc efficiency of the PV module supply of critical loads for those times when the
This paper is organized as follows: Section 2 sufficient amount of po- wer is available. In addition to
presents the basics of smart grid operation. Section 3 this when the power available from the solar energy
describes the fuzzy logic based model for solar energy based system is less then only critical loads can be
estimation. Fuzzy model for the prediction of SPV supplied. With the correct data, solar collectors are
system output for smart grid application is presented in fairly easy components to size, install, and begin
Section 4. Results and discussions are presented in generating power.
Section 5. A Conclusion followed by the references is
discussed in Section 6. 3. Fuzzy logic approach for global solar energy
estimation
2. Smart grid In the global solar energy estimation,
The global electricity sector and its customers are geographical features like latitude, longitude, altitude
faced with a number of challenges that are and meteorological parameters like mean duration
unparalleled since the advent of widespread sunshine per hour and temperature play an important
electrification. Challenges including climate change, role. There are other meteorological parameters also
escalating energy prices, energy security and energy which may affects the global solar energy such as
efficiency are converging to drive fundamental change relative humidity and wind speed. These parameters
in the way of energy produced, delivered and utilized. do not affect the global solar energy significantly.
Keeping in view of aforesaid the future electricity Therefore these parameters are not considered as input
system must produce and distribute electricity which parameters in the proposed model.
should be reliable, affordable and clean. To achieve Keeping in mind different climatic conditions,
the same, both the electricity grid and the existing four stations such as New Delhi, Jodhpur, Kolkata and
regulatory system must be smarter. Hence there arises Shillong are chosen for the present study that cover the
a need of grid that should be smart i.e. smart grid. different climatic zones.
Further, world is venturing into renewable energy re- New and Renewable Energy [21].
sources like wind and solar. With such unpredictable The input parameters are latitude, altitude,
energy sources, feeding the grid must be highly longitude, months of the year, temperature ratio (T/T0)
adaptive in terms of supply and demand. A good and mean duration sunshine per hour (S/S0). The
electric supply is one of the key infrastructure output parameter is the clearness index (H/H0). The

92
 Researcher 2017;9(12) http://www.sciencepub.net/researcher

estimated global solar energy is obtained by selected and operated at maximum power point
multiplying the esti- mated clearness index by H0. tracking.
Where H is the monthly mean global irradiance on Average power output of SPV system for smart
horizontal surface, and H0 is the extraterrestrial solar grid application can be intelligently predicted using the
irradiance on the 15th day of month, S is the monthly fuzzy logic based model presented. Fuzzy rules for the
mean of the daily hours of bright sunshine; S0 is the prediction of SPV system output are presented. In the
maximum daily hours of sunshine or day length. The proposed model solar insolation and cell temperature
ratio of S/S0 is the fraction of maximum possible are taken as input parameters and the output of the
numbers of bright sunshine hours and H/H0 is the model is PV power generated by the given module.
atmo- spheric transmission coefficient, commonly This data can then be used for variety of purposes
known as clearness index. The value of S0 can be which among others include reliability of supply,
computed from Cooper’s formula. intelligent load management, optimized resource
Data of meteorological parameters like clearness utilization, grid safety and grid stability.
index (H/H0), mean duration sunshine per hour (S/S0)
and temperature ratio (T/T0) for New Delhi and 5. Results and discussions
Jodhpur stations . Fuzzy logic based model is developed and
Similarly, the data of meteorological parameters presented for the estimation of global solar energy
mentioned above for Kolkata. using the above mentioned data for New Delhi,
The impact of mean duration sunshine per hour Jodhpur, Kolkata and Shillong stations. The output of
and tempera- ture is more significant as reported in the the model is clearness index (H/H0). Obtained results
literature [11e12,15]. Therefore, it is necessary to are multiplied by the extraterrestrial solar irradiance
select them as input parameters to get more accurate (H0) to get the global solar irradiance (H). Results of
output. Hence, the sunshine duration and temper- ature fuzzy logic based models are compared with the well
have been selected as input meteorological parameters established REST2 (Reference Evaluation of Solar
for the estimation of global solar energy. Further, we Transmittance, 2 band) model [4] for validation.
have also taken into consideration the effect of Further ANN based model considering the same input
geographical parameters like lati- tude, longitude and parameters is also developed and presented in
altitude as inputs, though these parameters remain Appendix. The estimated global solar energy us- ing
constant for a particular location. Hence these the above mentioned models is compared with the
parameters have not been included in defining the measured data for New Delhi, Jodhpur, Kolkata and
rules. Shillong stations are presented respectively. The fuzzy
The fuzzy system contains a set of rules which logic and ANN based models are developed in
are developed from qualitative descriptions. In fuzzy MATLAB toolboxes. The perfor- mance of the models
systems, rules may be fired of the sources particularly is evaluated on the basis of mean absolute relative
wind and solar. This problem can be overcome by error in percentage. The mean absolute relative error is
predicting the renewable energy sources intelligently. calculated as very close to each other. It shows that the
In addition, the generation and load forecasting can performance of the model is satisfactory. Obtained
provide output consequent is the clearness index results of solar insolation are further simulated along
(H/H0). The consequents of the rules are shown in the with cell temperature as input parameters for the
shaded part of the matrix [22]. Ob- tained results are prediction of output power of the given SPV system.
multiplied by the extraterrestrial solar irradiance (H0) As the cell temperature of a PV module affects the
to get the global solar energy. The rules are output of the PV system. The predicted output of SPV
summarized in the fuzzy decision matrix. system in comparison with the computed power for
The fuzzy variables, mean sunshine duration per New Delhi is presented. Due to paucity of space the
hour (S/S0), temperature ratio (T/T0) and clearness results of other stations are not presented in this paper.
index (H/H0) are described by the fuzzy terms high, The predicted power of SPV system in comparison
highemedium/normal, medium/normal, with computed power is quite accurate. The mean
lowemedium/normal or low. These fuzzy variables absolute relative error. The mean percentage error
described by linguistic terms are represented by using fuzzy logic for New Delhi, Jodhpur, Kolkata and
membership functions. The membership functions for Shillong are 5.11%. 5.17%, 5.53% and 5.67%
sunshine duration, temperature ratio. respectively whereas it is 4.87%, 4.89%, 4.81% and
In this paper, prediction of SPV system output is 5.97% respec- tively using ANN. The results obtained
done on the basis of data acquired from fuzzy model from REST2 model are 5.54%, 5.70%, 6.26% and
and cell temperature. This is done by predicting the 6.47% respectively. As REST2 model is a clear sky
output of PV module for a given insolation level and model and does not incorporate the uncertainties in the
cell temperature. In this paper, 120 W PV module is weather, therefore intelligent models for the estimation

93
 Researcher 2017;9(12) http://www.sciencepub.net/researcher

of solar energy are required. It is fact that the References

performance of REST2 model is better as compared to 1. Rizwan M, Jamil M, Kothari DP. Solar energy
other models available in the literature. In addition, the estimation using REST model for PV-ECS based
data of transmittances is not available during monsoon distributed power generating system. Sol Energy
month (July September). Therefore it is difficult to Mater Sol Cells 2010;94(8):1324e8.
estimate the solar energy during above mentioned 2. Ahmed MA, Ahmed F, Akhter MW. Estimation
period. Further, the results ob- tained from ANN of global and diffuse solar ra- diation for
model are slightly better as compared to the proposed Hyderabad, Pakistan. J Basic Appl Sci
model but it can be compromised with other 2009;5:73e7.
advantages of fuzzy logic like flexibility, tolerant of 3. Gupta N, Jain A. Smart grids in India. Renew
imprecise data, model for nonlinear functions of Energy (Akshay Urja) 2011;5:38e41.
arbitrary complexity, blended with con- ventional 4. Gueymard CA. REST2: high performance solar
control techniques, based on natural language apart radiation model for cloudless sky irradiance,
from easy to develop and better understanding. In illuminance, and photosynthetically active
addition, the esti- mated values obtained from fuzzy radiation-validation with benchmark dataset. Sol
logic based model as compared to measured data are Energy 2008;82:272e85.
also represented graphically for New Delhi, Jodhpur, 5. Gueymard CA. Direct solar transmittance and
Kolkata and Shillong respectively. It is clearly seen energy predictions with broadband models. Sol
that the measured and estimated values are for New Energy 2003;74:355e79.
Delhi, Jodhpur, Kolkata and Shillong is 5.41%, 5.47%, 6. Rizwan M, Jamil M, Kothari DP. Estimation of
5.59% and 5.75% respectively. Therefore, the global solar irradiance using regression model for
proposed models may be used for the prediction of Jodhpur, India. In: Proceedings of International
solar energy and PV power generation. Further, the Sympo- sium on Photovoltaic Science &
obtained results may be used for smart grid Technology, Kanpur; 2010. p. 43.
application. 7. Kaplanis S, Kaplani E. Stochastic prediction of
hourly global solar radiation for Patra Greece.
6. Conclusions Appl Energy 2010;87:3748e58.
In the proposed work, a fuzzy logic based model 8. Kaplanis S, Kaplani E. A model to predict
for esti- mating the global solar energy for a particular expected mean and stochastic hourly global solar
station is pre- sented. The meteorological parameters I (h; nj) values. Renew Energy 2007;32:1414e25.
like mean duration sunshine per hour and temperature 9. Kaplanis SN. New methodologies to estimate the
and the geographical pa- rameters of the particular hourly global solar radiation; comparison with
location are considered as input pa- rameters. Since existing models. Renew Energy 2006;31:781e90.
there is always ambiguity in climatic conditions and it 10. Falayi EO, Adepitan JO, Rabiu AB. Empirical
is difficult to predict solar energy accurately by using models for the correlation of global solar
mathematical formulas and regression techniques at a radiation with meteorological data for Iseyin,
particular time and at a particular location that is Nigeria. Int J Phys Sci 2008;3:210e6.
utmost important in the design of solar based power 11. Akpabio LE, Etuk SE. Relationship between
station. However such problems are overcome by global solar radiation and sunshine duration for
fuzzy logic based techniques. Based on the data Onne, Nigeria. J Phys 2003;27:161e7.
generated, fuzzy logic is applied to estimate the solar 12. Sen Z. Fuzzy algorithm for estimation of solar
energy at four Indian stations that covers different irradiation from sunshine duration. Sol Energy
climatic as well as geographical conditions. 1998;63:39e49.
The results obtained from the proposed fuzzy 13. Gautam NK, Kaushika ND. A model for the
logic based model are compared with reference data estimation of solar radiation using fuzzy random
and it is found that the per- centage error is within variables. J Appl Meteorology; 2002:1267e76.
permissible limits that are about 5% for all stations 14. Paulescu M, Gravila P, Paulescu ET. Fuzzy logic
considered in this work. algorithm for atmospheric transmittances of use
By using proposed model other uncertainties may in solar energy estimation. Energy Convers
also be included easily in estimating solar energy more Manag 2008;49:3691e7.
accurate for any location that are required to set up 15. Paulescu ET, Paulescu M. Fuzzy modeling of
solar based power stations. Further, the estimation of solar irradiation using air tem- perature data.
solar energy and prediction of SPV output would be Theor Appl Climatol 2008;91:181e92.
helpful in smart grid application. 16. Gomez V, Casanovas A. Fuzzy logic and
meteorological variables: a case study of solar
irradiance. Fuzzy Sets Syst 2002;126:121e8.

94
 Researcher 2017;9(12) http://www.sciencepub.net/researcher

17. Gomez V, Casanovas A. Fuzzy modeling of solar 21. Solar radiation handbook. Pune: SEC & IMD;
irradiance on inclined sur- faces. Sol Energy 2008.
2003;75:307e15. 22. Ng HN, Salama MMA, Chikhani AY. Capacitor
18. Iqdour R, Zeroual A. Prediction of daily global allocation by approximate reasoning: fuzzy
solar radiation using fuzzy systems. Int J Sustain capacitor placement. IEEE Trans Power Deliv
Energy 2005;26:19e29. 2000;15: 393e8.
19. Hidayatullah NA, Stojcevski B, Kalam A. 23. Duffie JA, Beckmen WA. Solar energy thermal
Analysis of distributed generation systems, smart processes. New York: John Wiley & Sons, Inc;
grid technologies and future motivators 1980.
influencing change in the electricity sector. Smart
Grid Renew Energy 2011;2:216e29. Appendix
20. Kuschke M, Strunz K. Modeling of tidal energy For the estimation of global solar energy at New
conversion systems for smart grid operation. In: Delhi, Jodhpur, Kolkata and Shillong stations, an
Proceedings of IEEE Power and Energy Society ANN based model has been developed and presented
General Meeting, Detroit; 2011. pp. 1e2. in Figure 1.

Figure. 1. ANN model for the estimation of global solar energy.

12/25/2017

95