Load Forecasting Analysis Using Contextual Data and Integration With Microgrids Used For Off Grid EV Charging Stations

Electricity is an essential component of the smooth working of every sector. If a successful prediction of how much electricity will be required for say the next 24 hours or 48 hours can be made, it will not only help in efficiently planning the activities and operations but also help in minimizing the cost incurred. In this thesis the same is being attempted, first, a model is created that can predict the energy consumption of households using various tools available. To achieve this, historical data of the past 5 years that has been recorded in London has been used. Secondly, a model is created that can forecast future energy requirements of EV charging stations, which will help in optimizing their working in off-grid areas. To achieve this, data that has been recorded for more than 100 charging stations across the Salt Lake area for around 4 years has been used.The second part of this thesis involves integrating the load forecasting model with the solar energy forecasting model and microgrid optimization. Since microgrid can disconnect from the energy source and work autonomously, solar energy can be used to generate power that can be used instead of the energy that is bought from the grid. This will help in building an efficient model that maximizes the profit by making sure energy is bought at a minimum price and the same energy is being utilized when the demand is high.