Periodic autoregressive forecasting of global solar irradiation without knowledge-based model implementation

Abstract

Reliable forecasting methods increase the integration level of stochastic production and reduce cost of intermittence of photovoltaic production. This paper proposes a solar forecasting model for short time horizons, i.e. one to six hours ahead. In this time-range, machine learning methods have proven their efficiency. But their application requires that the solar irradiation time series is stationary which can be realized by calculating the clear sky global horizontal solar irradiance index (CSI), depending on certain meteorological parameters. This step is delicate and often generates additional uncertainty if conditions underlying the calculation of the CSI are not well-defined and/or unknown. As a novel alternative, we introduce a so-called periodic autoregressive (PAR) model. We discuss the computation of post-sample point forecasts and forecast intervals. We show the forecasting accuracy of the model via a real data set, i.e., the global horizontal solar irradiation (GHI) measured at two meteorological stations located at Corsica Island, France. In particular, and as opposed to methods based on CSI, a PAR model helps to improve forecast accuracy, especially for short forecast horizons. In all the cases, PAR is more appropriate than persistence, and smart persistence. Moreover, smart persistence based on the typical meteorological year gives more reliable results than when based on CSI.

Benchmarks for Solar Radiation Time Series Forecasting

“…It is important to remember that the goal of the predictive methodologies described and tested here is not to be the best forecasting models, but the simplest ones (sometimes very simple) that would allow arbitration of the classification of more complex models. For example, forecasts made with artificial neural networks of the multilayer perceptron type on this data set are slightly less than 20% for 1-h horizon [40] but are too complex to be used as reference.…”

Evaluation of net radiation using the autoregressive models with higher orders over Nigeria

“…The performance of the ( ) model developed in this study was assessed using the mean squared error (MSE), root mean squared error, coefficient of determination (R 2 ), mean absolute percentage error (MAPE) and Akaike's information criterion (AIC) as expressed in Equations ( 5) to (9). Several studies used ( ) techniques to evaluate the global solar radiation [17]- [20], total ozone content [21] and ocean current [22]. The method was found to predict the variables very well with great accuracy compared to ground measurement. )…”

Non-linear analysis of bursty workloads using dual metrics for better cloud resource management

“…It also includes a fruitful discussion on nonstandard analysis. 2 Persistence and scaled, or smart, persistence are quite often discussed elsewhere in the literature, essentially perhaps in meteorology (see, e.g., [49], [50]) and climatology (see, e.g., [38]). 3 For the estimation techniques, see also [22], [44], and [36] for more mathematical details.…”