Nonlinear autoregressive prediction model for VAWT power supply network energy management

Due to technological breakthroughs and improvements in the wind energy industry over the past few decades, wind energy is seen as one of the viable alternatives to non-renewable energy resources (such as fossil fuels). Researchers have recently focused on traffic-generated wind since it can be captured by vertical axis wind turbines (VAWT) and transformed into useful energy with sustainability. This paper presents a nonlinear autoregressive (NAR) approach to offer an energy management prediction model to forecast energy management choices for VAWTs with batteries and energy purchased from the power grid. Compared to earlier heuristic and computational methods, the developed approach is more novel and robust as it integrates MILP with ANNs, which further enables finer convergence for complex optimization functions. To identify the elements that have a major influence on the prediction’s accuracy, the prediction’s accuracy is evaluated in a variety of circumstances. Wind speeds, turbine swept area, battery size, and seasonality of energy demand are some of these variables. The findings indicate that seasonal variations and energy demand significantly affect the accuracy of the forecast by 69.23%, while the swept area of the turbine has the least effect, affecting the accuracy of the prediction by 15.38%. Furthermore, according to the forecast results, using VAWT under steady state conditions as a supplementary energy supply can reduce carbon emissions by up to 231 tons yearly and save up to $52,831 in energy expenses annually.