Abstract:

With the rapid growth of the world’s population and the decrease in fossil fuels reserves, the demand for renewable energy, especially solar energy, is increasing every day. Smart networks have been developed to meet the increasing energy demand in a smart structure form with renewable energy sources. One of the main objectives of smart networks is to include and increase the use of renewable energy sources in the energy production process. However, the inclusion of renewable energy sources that are interrupted and uncontrollable in existing electricity network systems brings a number of challenges. For determining when the demand for variable electricity will be met by which renewable sources, the forecast of energy production is one of the most important challenges. Therefore, in this study, a model was developed that used atmospheric measurements to predict the power plant-specific and hour-based energy production. This model was created by machine learning approaches and energy production was predicted. A variety of methods such as multi-linear regression, Powell optimization and possible programming based on the Markov Chain Monte Carlo simulations have been used and the predictive abilities have been compared with each other. The energy production is estimated with an analytical approach of 80% accuracy, while the probability approach has been successfully estimated to determine the upper and lower limits in the 95% hour-based confidence range. The results obtained in this study showed that using machine learning and atmospheric data, variable energy production can be predicted in advance. It is also estimated that predictive algorithms will facilitate the integration of renewable energy systems into the existing network and make the smart network more widespread.

Keywords: