Detailed Investigation of Persistence Technique for Microgrid Load and Renewable Energy Forecasting

Abstract

A micro-grid system, whether linked to  the utility grid or self-contained, often comprises of a  combination of renewable and non-renewable production,  controllable or non-controllable loads, and Energy  Storage Systems (ESSs) such as batteries or flywheels. To  estimate how much power is used from controlled  resources such as ESS, diesel generators, micro-turbines,  or gas turbines, we must first identify how much demand  exists or how much renewable energy sources provide,  which is performed by forecasting techniques. Due to the  intermittent nature of renewable resources such as wind  energy or solar energy, precise forecasting of wind power  or solar power is challenging. These projections are  heavily reliant on weather predictions. It is obvious that  forecasting any data based on forecasting other factors  would result in increased inaccuracy, even if the  relationship between the inputs and outputs could be  predicted using regression methods. As a result, this  research demonstrates a method for producing short-term  forecast results using historical power data rather than  numerical weather projections. Forecasting power  generation from renewable energy sources (RESs) has  become critical in micro-grid applications to improve  asset scheduling and dispatching.