CSEIT172215 | Received: 05 March 2017 | Accepted: 14 March 2017 | March-April-2017 [(2)2: 26-30]
International Journal of Scientific Research in Computer Science, Engineering and Information Technology
© 2017 IJSRCSEIT | Volume 2 | Issue 2 | ISSN : 2456-3307
26
Review of Forecasters Application to Solar Irradiance
Forecasting
M. Madhiarasan
*1
, S. N. Deepa
2
*
1
Research Scholar (Ph. D), Department of Electrical and Electronics Engineering, Anna University Regional Campus Coimbatore,
Coimbatore, Tamil Nadu, India
2
Associate Professor, Department of Electrical and Electronics Engineering, Anna University Regional Campus Coimbatore, Coimbatore,
Tamil Nadu, India
ABSTRACT
Solar irradiance forecasting is an ongoing research area, because of the improved and advanced utilization of
renewable energy resource plenty of researchers have turned their attention to constitute intelligent solar irradiance
forecasting tool. This paper provides the overview of the solar irradiance forecasting process, review the researches
regard to solar irradiance forecasting and suggested future works to the forthcoming researchers.
Keywords: Intelligent, Renewable Energy, Forecasting, Solar Energy, Solar Irradiance
I. INTRODUCTION
The deployment of solar energy for electricity
generation pursued to eliminate the environment
hazards due to global warming, stabilize the
concentration of GHG (greenhouse gas) and energy
crises. Solar power forecasting is directly obtained by
solar irradiance forecasting outcome with PV
(photovoltaic) material and orientation. Solar irradiance
is known as the rate at which solar electromagnetic flux
acquired by a surface per unit area. The unit of solar
irradiance is W/m
2
(watts per meter squared). The
beforehand planning of future electricity power
production from intermitted resources (solar, wind and
etc.) is a tough and critical process compared to
nonrenewable energy resource due to the
meteorological effect to avoid this fact of solar
irradiance forecasting is highly essential and significant
concerning to solar energy system. General pictorial
representation of the solar irradiance forecasting
process is showcased in Fig. 1.
First, gather related information or data (cloud cover,
precipitation of water content, sunshine hours, relative
humidity, temperature, wind speed and etc.) to process
the initial stage of solar irradiance forecasting after
completion of data collection perform data
preprocessing with the help most relevant methods
from the available methods like data cleaning (example:
resolve missing data issue (principle component
analysis)), data transformation (example: normalization
(min max)), data discretization (example: empirical
mode decomposition, prediction intervals) and data
integration (example: grouping (cluster)), followed by
the data preprocessing a suitable forecaster is selected
from the available forecasting methods like physical,
statistical, innovative methods. Using the individual
method alone is not sufficient to produce the better
forecast, thus suitable optimization algorithm is
adopted to build hybrid forecaster. At last the best
forecasting of solar irradiance is obtained and qualified
by error metrics like MAPE (mean average percentage
error), MSE (mean square error), MRE (mean relative
error) and RMSE (root mean square error). Complete
available forecasting methods are depicted in Fig. 2,
which convey the better perseverance concern to
forecasting methods. With the help of investigation of
historical data statistical methods are developed,
similarly with the help of reconstruction of physical
facts and historical data NWP methods are constructed.
Persistence is standard method suitable for very short
time scale forecasting. NWP (numerical weather
prediction method) is suitable for long time scale
forecasting. Statistical methods are suitable for medium;