Modeling and Forecasting Renewable Energy Resources for Sustainable Power Generation: Basic Concepts and Predictive Model Results

The utilization of renewable energy is an essential tool for the mitigation of the negative impacts of a changing climate on water resources, ecosystems and human lives. However, the intermittent nature of renewable energy poses a practical challenge for its wider applicability such as in electrical power grid utilization. Accurate modeling and forecasting of renewable energy resources, such as wind and solar energy are necessary to make the energy generation process easy and relatively reliable for utilization in grid energy systems. As the development of physics-based models can be economically costly and includes many assumptions and constraints, the emergence of data-driven and machine learning modeling approaches are becoming attractive and viable alternative tools. This chapter outlines the respective phases required for the development of machine learning models in the renewable energy generation sector, including the pertinent concepts and definitions that are used in time-series forecasting approaches. Such information provides a handy tool for engineers and renewable energy practitioners, as well as novice forecasters who wish to explore the practicality of methods for real-life forecasting of power. To provide insights into how machine learning could be used in the energy modeling sector, a 10-min wind speed forecasting case study for Rakiraki, Fiji is carried out using the widely adopted artificial neural network (ANN) model, and the results are compared with multiple linear regression (MLR) models. This serves as an example of how the different phases of model development need to be implemented in a power forecasting study. The present study reveals a better performance of the ANN model over the MLR model in a 10-min wind speed forecasting problem.