AN INTEGRATED DEEP-LEARNING MODEL-BASED PRECISE FORECASTING MODEL FOR SUSTAINABLE ENERGY SYSTEMS

Abstract

Worldwide, Sustainable Energy Systems (SES) and regulations have been advocated to shift from fossil fuel sources to ecologically SES, including Wind Power (WP), Solar Power (SP), and Fuel Cells (FC). WP and SP sources must be more consistent and accessible when integrated into SES; hence, caution is required in their implementation and associated legislation. This paper formulates an energy forecasting model incorporating SES, serving as a basis for policy, utilizing the Korean model. Deep Learning (DL) predicts variable changes in power needs and generations, that is essential for SES and which traditional models cannot do. The gated recurrent unit has a higher forecasting ability than the other forecasting methods. Hence, it is chosen as the foundational model to analyze four distinct SES. The possibilities are assessed based on financial-ecological cost evaluation. The ideal situation employs an integrated gasified paired cycle, onshores and offshores wind turbines, SP locations, and FC facilities; this situation exhibits minimal economic-environmental expenses, produces reliable power to meet demand, and attains a 100% SE policy. The ideal scenario is evaluated by studying strengths, shortcomings, possibilities, and threats, examining domestic and global techno-economic and ecological power conditions.