Abstract
The description and forecasting of hourly solar resource is fundamental for the operation of solar energy systems in the electric grid. In this work, we provide insights regarding the hourly variation of the global horizontal irradiance in Medellín, Colombia, a large urban area within the tropical Andes. We propose a model based on Markov chains for forecasting the hourly solar irradiance for one day ahead. The Markov model was compared against estimates produced by different configurations of the weather research forecasting model (WRF). Our assessment showed that for the period considered, the average availability of the solar resource was of 5 PSH (peak sun hours), corresponding to an average daily radiation of ~5 kWh/m2. This shows that Medellín, Colombia, has a substantial availability of the solar resource that can be a complementary source of energy during the dry season periods. In the case of the Markov model, the estimates exhibited typical root mean squared errors between ~80 W/m2 and ~170 W/m2 (~50%–~110%) under overcast conditions, and ~57 W/m2 to ~171 W/m2 (~16%–~38%) for clear sky conditions. In general, the proposed model had a performance comparable with the WRF model, while presenting a computationally inexpensive alternative to forecast hourly solar radiation one day in advance. The Markov model is presented as an alternative to estimate time series that can be used in energy markets by agents and power-system operators to deal with the uncertainty of solar power plants.
Highlights
The performance of solar power plants depends essentially on an adequate characterization of the variations of the incoming solar radiation over land surface [1]
The assessment of the solar resource for the period March 2016 to February 2017 is performed with only daytime global horizontal irradiance (GHI) values, from 6 Local Time (LT) to 18 LT, since these are the limits of the interval in which finite values of GHI
We analyzed global horizontal irradiance data from a pyranometer station located in Medellín, Colombia, with records during the period March 2016 to February 2017
Summary
The performance of solar power plants depends essentially on an adequate characterization of the variations of the incoming solar radiation over land surface [1]. Markov-chain based approaches are not frequently used in the mid-term forecast (day-ahead forecasting) of the hourly series of GHI, which is a type of forecasting that is fundamental in the operation of generation systems based on solar radiation They are rather used for generating synthetic GHI series for solar system simulations, or to perform forecasts with time horizons near the time resolution of the model (i.e., few time steps into the future). As an alternative to obtain mid-term forecasts of the hourly GHI, we implemented a two-part model based on discrete Markov processes to estimate GHI in Medellín, Colombia, a highly populated city in the tropical Andes This model is a modified version of the formulations found in the works of [9,21,22,27] and is capable of forecasting diurnal series of hourly GHI for one day-ahead, by taking into consideration the current seasonal effects over the behavior of the variable. The magnitude of the errors obtained with the Markov model are comparable to the errors The magnitude of the errors obtained with the Markov model are comparable to the errors obtained with other models identified in the literature
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
