www.ijraset.com Volume 4 Issue III, March 2016
IC Value: 13.98 ISSN: 2321-9653
International Journal for Research in Applied Science & Engineering
Technology (IJRASET)
© IJRASET 2013: All Rights are Reserved 356
Forecasting Of Short Term Wind Power Using
ARIMA Method
Prashant Pant
1
, Achal Garg
2
1,2
Engineer, Keppel Offshore and Marine Engineering India Pvt. Ltd, Mumbai
Abstract- Wind power, i.e., electrical energy produced making use of the wind resource, is being nowadays constantly connected
to the electrical system. This has a non-negligible impact, raising issues like network stability and security of the supply. An
accurate forecast of the available wind energy for the forthcoming hours is crucial, so that proper planning and scheduling of
the conventional generation units can be performed. Also, with the liberalization of the electrical markets worldwide, the wind
power forecasting reveals itself critical to assure that the bids are placed with a minimum possible risk. The main application for
wind power forecasting is to reduce the need for balancing energy and reserve power, which are needed to integrate wind power
within the balancing of supply and demand in the electricity supply system. At times of maintenance it is required to know how
much power would have been generated and should be supplied by other source. This work addresses the issue of forecasting
wind power with statistical model, the Autoregressive Integrated Moving Average (ARIMA). The basic theory and the respective
application of these models to perform wind power prediction are presented in this paper. Furthermore, their forecasting abilities
are shown with the help of graphs.
Keywords- Autoregressive Moving Average, Energy, Forecasting, Model, Short Term, Wind Power.
I. INTRODUCTION
A wind power forecast corresponds to an estimate of the expected production of one or more wind turbines, referred to as a wind
farm in the near future. By production is often meant available power for wind farm considered with units kW or MW depending on
the wind farm nominal capacity. Forecasts can also be expressed in terms of energy, by integrating power production over each time
interval. Un-forecasted wind fluctuations increase requirements for spinning reserves and raise electricity system production costs.
Un-forecasted large ramp events can affect electricity system reliability. State of art forecasts have high economic value compared
to their cost. Wind power forecasts are essential for effective grid management with high wind penetrations (>5%). Forecasting
plays an important role in managing the variable output from wind farms on the grid making it appear more like conventional
energy sources. Reduces cost of integrating wind on the grid and so reduces energy costs, both financial and environmental, for
everyone. The advanced wind power forecasting methods are generally divided into two main groups; first is physical approach,
consists of several sub-models which together deliver the translation from the NWP forecast at a certain grid point and model level
to power forecast at the considered site and at turbine hub height. Every sub-model contains the mathematical description of the
physical processes relevant to the translation, and the second is statistical approach which consists of emulating the relation between
meteorological predictions, historical measurements and generation output.
II. THE FORECAST TIME HORIZONS AND METHODOLOGY
The very-short–term forecasting approach consists of statistical models that are based on the time series approach and includes such
models as the Kalman Filters, ARMA, ARX, and Box-Jenkins forecasting methods. These types of models only take as inputs past
values from the forecasted variable (e.g., wind speed, wind generation). At the same time, they can also use other explanatory
variables (e.g., wind direction, temperature), which can improve the forecast error. Since these methods are based solely on past
production data, they only outperform the persistence model (reference model) for forecast horizons between 3–6 hours. The
medium term forecasting is up to a time horizon of 72 hours. It is generally used for aggregate production planning, man power
planning and inventory. Several statistical models for day-ahead forecasts are clubbed together to decrease the forecast error. The
methods used for forecasting are; Neural Networks, Support Vector Machines, Regression Trees with Bagging, Random Forests,
Adaptive Neural Fuzzy System, Nearest Neighbour Search. In statistics and econometrics, and in particular in time series analysis
an autoregressive integrated moving average (ARIMA) model is a generalization of an autoregressive moving average (ARMA)
model. These models are fitted to time series data either to better understand the data or to predict future points in the series
(forecasting). They are applied in some cases where data show evidence of non-stationary, where an initial differencing step