Solar Irradiance Forecasting Using Triple Exponential Smoothing

Abstract

Owing to the growing concern of global warming and over-dependence on fossil fuels, there has been a huge interest in last years in the deployment of Photovoltaic (PV) systems for generating electricity. The output power of a PV array greatly depends, among other parameters, on solar irradiation. However, solar irradiation has an intermittent nature and suffers from rapid fluctuations. This creates challenges when integrating PV systems in the electricity grid and calls for accurate forecasting methods of solar irradiance. In this paper, we propose a triple exponential-smoothing based forecasting methodology for intrahour forecasting of the solar irradiance at future lead times. We use time-series data of measured solar irradiance, together with clear-sky solar irradiance, to forecast solar irradiance up-to a period of 20 minutes. The numerical evaluation is performed using 1 year of weather data, collected by a PV outdoor test facility on the roof of an office building in Utrecht University, Utrecht, The Netherlands. We benchmark our proposed approach with two other common forecasting approaches: persistence forecasting and average forecasting. Results show that the TES method has a better forecasting performance as it can capture the rapid fluctuations of solar irradiance with a fair degree of accuracy.

Modelo predictivo inteligente para la generación de energía eléctrica en un sistema fotovoltaico

“…El análisis del error fue realizado aplicando el MAE (Mean Absolute Error o Error Medio Absoluto) y el MAPE (Mean Absolute Percentage Error o Error Porcentual Medio Absoluto) como técnicas de evaluación de precisión por ser de los más utilizados en sistemas inteligentes. MAE y MAPE están descritas por las ecuaciones 8 y 9, respectivamente, donde N es la cantidad total de muestras, s es la muestra a considerar, Pm es el valor de potencia real o medida y Pe es la potencia estimada por el ANFIS (Dev et al, 2018;Pitalúa-Díaz et al, 2019;Ruz-Hernandez et al, 2019 A partir de la Figura 6 y Figura claramente se logra apreciar un comportamiento lineal de los datos puntuales (verde) el cual se comprueba con la línea de tendencia (roja) indicando que la estimación del modelo implementado es lineal y directamente proporcional a los valores reales. A su vez, la Tabla 2 detalla los valores obtenidos por cada método evaluado.…”

Meteorological Variables’ Influence on Electric Power Generation for Photovoltaic Systems Located at Different Geographical Zones in Mexico

“…As seen in Section 1 and according to Reference [51], the output power of a PV array greatly depends, among other parameters, on solar radiation; however, this variable has an intermittent nature and suffers from rapid fluctuations. In Reference [41,51], the above is considered and some parameters besides the solar radiation such as clear sky or weather data are added; nonetheless, for these cases, the solar radiation is either the estimated output or an estimated input, contrary to this paper where this variable is measured on-site.…”

“…Two kinds of errors were applied, where "P m " was the measured power, "P e " was the estimated power, "s" was the sample in consideration, and "N" was the total amount of samples. The first one was called Mean Absolute Error (MAE) defined by Equation (11) and was used as a standard statistical metric to measure the model performance in meteorology, air quality, and climate research studies [42,51,52]. The second one was the percentage of the MAE known as Mean Absolute Percentage Error (MAPE) defined by Equation (12).…”

Predicting Solar Irradiance in Singapore

“…The Triple Exponential Smoothing technique has already been used in different time series data [3][4][5][6]. In [3], the TES method has been used for solar irradiance prediction for the region of Utrecht, the Netherlands. In this paper, we use this technique for forecasting the solar irradiance in the tropical region viz.…”