Long-Term Hourly Scenario Generation for Correlated Wind and Solar Power combining Variational Autoencoders with Radial Basis Function Kernels

Abstract

Accurate generation of realistic future scenarios of renewable energy generation is crucial for long-term planning and operation of electrical systems, especially considering the increasing focus on sustainable energy and the growing penetration of renewable generation in energy matrices. These predictions enable power system operators and energy planners to effectively manage the variability and intermittency associated with renewable generation, allowing for better grid stability, improved energy management, and enhanced decision-making processes. In this paper, we propose an innovative method for generating long-term hourly scenarios for wind and solar power generation, taking into consideration the correlation between these two energy sources. To achieve this, we combine the capabilities of a Variational Autoencoder (VAE) with the additional benefits of incorporating the Radial Basis Function (RBF) kernel in our artificial neural network architecture. By incorporating them, we aim to obtain a latent space with improved regularization properties. To evaluate the effectiveness of our proposed method, we conduct experiments in a representative study scenario, utilizing real-world wind and solar power generation data from the Brazil system. We compare the scenarios generated by our model with the observed data and with other sets of scenarios produced by a conventional VAE architecture. Our experimental results demonstrate that the proposed method can generate long-term hourly scenarios for wind and solar power generation that are highly correlated, accurately capturing the temporal and spatial characteristics of these energy sources. Taking advantage of the benefits of RBF in obtaining a well-regularized latent space, our approach offers improved accuracy and robustness in generating long-term hourly scenarios for renewable energy generation.