A Systematic Approach to Predict Solar Energy using Artificial Neural Networks

Abstract

The impact of climate change on national and municipal governments is leading to an increasing global recognition and adoption of sustainable development and renewable energy. The EU 2030 agenda aims to ensure universal access to affordable, reliable, and sustainable energy in the future. An obstacle to attaining this objective is the restricted dependability of certain renewable energy sources. Although individual citizens and governmental entities are striving to generate sufficient renewable energy to fulfil their requirements, the exact amount of investment required to mitigate the unpredictability resulting from natural factors such as fluctuations in wind speed and daylight remains uncertain. A technology that accurately forecasts the energy production of renewable sources over the course of a year for a specific location can greatly enhance the effectiveness of investments in sustainable energy. This study employs Internet of Things (IoT) sensors, which are distributed across Europe, along with open data sources to construct a tool that utilizes artificial neural networks. We examine the impact of different factors on the estimation of energy generation and explore the possibility of utilizing public data to forecast the anticipated output of renewable sources. We provide consumers with the required information to make investment decisions based on the energy production requirements specific to their location. Our approach provides a specialized level of abstraction that prioritizes energy generation rather than radiation statistics. It can be tailored and adapted for various locations using publicly available data, distinguishing it from state-of-the-art alternatives.