Login Register
 
Home Accepted Papers
HYBRID ARMA-ARIMA MODEL FOR SHORT-TERM POWER FORECASTING IN GRID-CONNECTED PHOTOVOLTAIC PLANTS
Paper ID : 1306-ICESE
Authors:
احسان فرازمند *1, سید محمد تقی المدرسی2, علی اکبر تدین2
1خیابان فردوسی نبش لبخندشکوفه پلاک 410
2دانشگاه یزد
Abstract:
In this paper, we propose a 10-minute forecast for a grid-connected photovoltaic system in smart grid. The model we employ is hybrid i.e. we choose between autoregressive-moving average (ARMA) and autoregressive integrated moving average (ARIMA) models through a Kwiatkowski, Phillips, Schmidt, and Shin (KPSS) test. The stochastic behaviour of the renewable resources such as solar and wind power is noteworthy to attain a stable and reliable integration of these resources. To deal with this issue, we may study forecasting of power generation of photovoltaic systems for different forecast horizons. There is a number of forecasting methods; nevertheless, statistical time series ones such as ARMA and ARIMA are suggested for short-term forecasts due to high level of correlation among the successive samples of data. Simulation results, obtained from a grid-connected photovoltaic plant, show that the hybrid proposed method has a good performance in sunny days, while its forecasts for cloudy days are comparable with pure ARMA or ARIMA methods.
Keywords:
ARMA, ARIMA, hybrid model, photovoltaic power output forecasting, unit root test.
Status : Paper Accepted