Forecasting of global horizontal irradiance by exponential smoothing, using decompositions

Abstract

Time series methods are frequently used in solar irradiance forecasting when two dimensional cloud information provided by satellite or sky camera is unavailable. Furthermore, satellite and sky camera based methods lose resolution at a 1-h time horizon. Exponential smoothing (ETS) has received great attention in the recent years due to the invention of its state space formulation. We explore 1-h ahead ETS forecasting of solar irradiance in this paper. Several knowledge based decompositions are considered to improve the forecast accuracy and computation speed.Three methods are proposed. The first method considers an additive seasonal-trend decomposition prior to the use of ETS. In such a way, the ETS state space is reduced, thus facilitates online forecasting applications. The second method decomposes the global horizontal irradiance (GHI) time series into a direct component and a diffuse component. These two components are used as forecasting model inputs separately; the results are recombined through the closure equation. We consider the time series of cloud cover index in the third method. ETS is first applied to the cloud cover time series to produce the forecast. The forecast cloud cover is then used to reconstruct GHI through regressions. It is found that the third method performs the best among three methods and all proposed methods outperforms persistence.

k-Nearest Neighbor Neural Network Models for Very Short-Term Global Solar Irradiance Forecasting Based on Meteorological Data

“…Salcedo-Sanz et al [5] worked on the prediction of daily global irradiation using a temporal Gaussian process, in which the study explains the suitability of Gaussian regression (GPR) for the estimation of solar irradiation compared to other machine learning regression algorithms. The GSI forecasting is not only used in stochastic modeling, but in other studies [6,7] attempted forecasting was analyzed using exponential smoothing combined with decomposition methods and least absolute shrinkage and selection operator model. Yang et al [6,7] studied the forecasting of global horizontal irradiance by exponential smoothing using decompositions, while on another study, they developed the least absolute shrinkage and selection operator model using irradiance very short-term forecasting.…”

“…The GSI forecasting is not only used in stochastic modeling, but in other studies [6,7] attempted forecasting was analyzed using exponential smoothing combined with decomposition methods and least absolute shrinkage and selection operator model. Yang et al [6,7] studied the forecasting of global horizontal irradiance by exponential smoothing using decompositions, while on another study, they developed the least absolute shrinkage and selection operator model using irradiance very short-term forecasting. Combining a forecasting model with GSI is important to get a better result, and forecasting GSI by the spatiotemporal pattern recognition method, ANN method, parametric models and decomposition models, has been described in [8][9][10].…”

Prediction of monthly average global solar radiation based on statistical distribution of clearness index

“…The level of clearness of sky was classified according to the classification by Khorasanizadeh et al [1] as follows: cloudy 0 < K T < 0:2 Partly cloudy 0:2 K T < 0:6 Sunny 0:6 K T < 0:75 Very sunny 0:75 K T < 1 (20) The result of the classification is shown in Table 3. The table reveals that the site was partly cloudy with the clearness index values between 0:2 K T < 0:6 in over 90% of times round the year.…”

“…Yang et al [20] combined ETS (Exponential Smoothing) with the knowledge-based heuristic time series decomposition methods to improve the accuracy and computational efficiency in forecasting global solar irradiation. Three decomposition methods were proposed: seasonal-trend decomposition procedure based on LOESS (LOcally-wEighted Scatter plot Smoothing), Decomposition using the closure equation and Decomposition using cloud cover.…”

Dynamic Coupling Analysis of Urbanization and Water Resource Utilization Systems in China

“…It typically includes three smoothing methods: single exponential smoothing, double exponential smoothing, and triple exponential smoothing. Studies show that using triple exponential smoothing ensures superior forecasts for trended time series data [34,35], that is, the forecasts are closer to the actuals using triple exponential smoothing. Thus, we choose the triple exponential smoothing method to forecast the degrees of coupling.…”