Prediction of hourly global solar radiation incident on a horizontal surface in Kampala using Markov Chain Model.

Abstract

The present investigation delves into the feasibility of predicting global solar radiation (GSR) on horizontal surfaces within Kampala using the Markov Chain Model (MCM). This predictive data is crucial for informing governmental policy formulation and implementation, providing vital technical and market support for the solar energy sector. Hourly data of global horizontal irradiation (GHI), gathered over a two-year period from January 2018 to December 2019, has been meticulously analyzed, organized into tables and graphs, and subjected to in-depth discussion. The GHI data utilized in this study was meticulously measured and documented through the utilization of a Kipp and Zonen CMP6 Pyranometer, horizontally installed on the rooftop of the Department of Physics at Makerere University in Kampala, Uganda. The study incorporated both first-order and second-order transition probability matrices from the Markov chain. The performance evaluation of the model was conducted using a suite of statistical measures, encompassing the root mean square error (RMSE), mean bias error (MBE), and the correlation coefficient (r). This analysis yielded a correlation coefficient of 94.0%, an MBE of -6.883 MJm −2 hour −1 , and an RMSE of 2.81 MJm −2 hour −1 using the first-order MCM. Meanwhile, the second-order MCM produced a correlation coefficient of 91.6%, an MBE of -5.457 MJm −2 hour −1 , and an RMSE of MJm −2 hour −1 . Comparisons between the performance of the MCM, an artificial neural network (ANN) model employed in this study, and other GSR prediction models corroborated that the MCM stands as a viable alternative with substantial predictive capacity. Notably, the MCM requires only a solitary input parameter to generate dependable forecasts. Given its stochastic nature, the MCM can proficiently accommodate the inherent stochasticity of GSR with remarkable precision. This predictive capability extends to areas lacking measurement instruments but sharing similar climatic conditions and topography as Kampala.