Intelligent dynamic grid forecasting algorithm for a grid-connected solar PV based microgrid

This paper outlines the enhancement in performance characteristics of microgrids operating in rural and semi-urban regions with the inclusion of intermittent grid to the existing solar PV microgrid installations. The paper presents an Intelligent Dynamic Grid Forecasting Algorithm (IDGFA) that predicts day-ahead requirement of power from the utility grid, providing for load scheduling for the grid operator. This results in dynamic and strategic resource allocation in real time and enables maximum load servicing by the microgrids. The IDGFA runs at 7:00 a.m. every day wherein day-ahead shortages are tracked based on hourly State Of Charge (SOC) estimates made for a 48 hour duration. The SOC estimates are made with inexpensive, accurate, hourly solar PV forecasts made available for the same period. The instance and duration of grid requirement are intelligently predicted by minimizing the shortages for next day, leading to maximizing load servicing efficiency for a particular PV size. This proves to be an inexpensive option for enhanced load servicing as alternatively, oversizing PV systems for the same purpose leads to much higher capital costs and excessive wastage of energy. Rural microgrid with intermittent grid connectivity that are examined in this paper are for conditions in central India and it is observed that the system efficiency is above 97% in all cases. The excess energy contributed by the intermittent grid is also found to be negligible.