Solar Powered Electricity Grids-A data driven approach towards creating a decision support system

Renewable energy is being increasingly incorporated in our city power grids. This not only reduces the dependence on fossil fuels but also lowers the carbon emissions. However, there is a need to address how to incorporate renewable energy in an efficient manner and reduce power losses and infrastructural damage. This paper proposes an ensemble method for solar power prediction. Solar output power usually has a complex and nonlinear characteristic due to intermittent and time varying behaviour of solar irradiance. So different machine learning models have been applied to best predict solar power generation and to mitigate the ‘Duck Curve’. The solar power prediction models developed in this paper are part of a larger Decision Support System (DSS) for energy grid management. By improving forecast accuracy, these models enable more effective decision-making, allowing grid operators to better address the Duck Curve problem and optimize energy storage and distribution.