Download or read book Photovoltaic and Load Forecasting Predictions for a Microgrid Using Machine Learning written by Nelson Mauricio Flores and published by . This book was released on 2021 with total page 0 pages. Available in PDF, EPUB and Kindle. Book excerpt: Load forecasting has always been an integral part of plant operation. Energy usage is tied to economics and national prosperity. The power grid is in a perpetual state of balance, with the customers consuming the exact amount of energy the utility provides. Knowing the load forecast allows for proper planning and dispatching of energy generation. Over the past several decades, power planners have only been worried about the unknown being load demand. Generation of energy was always assumed to be dispatchable on-demand, like coal, natural gas, etc., due to the abundant supply, which helps also keep energy prices low. Predicting solar irradiation and demand is a challenging task. In this thesis, several methods used in the literature for forecasting load and sun irradiance, such as Auto-Regressive Integrated Moving Average, Moving Average, Long Short-Term Memory, and Support Vector Machines are investigated. Atwo-layer battery management algorithm is used with historical data for load and sun irradiance of a large facility in Downtown Los Angeles to compare these methods in a practical setting. Both load and sun irradiance forecasting methods have been further categorized into short-term and long-term to be used in each layer of the controller. After finding the best algorithm in each category, they have been further optimized to improve their accuracy, and then the two-layer algorithm with optimized forecasts is compared with a conventional peak shaving method.