Proceedings of the 4th World Conference on Climate Change and Global Warming
Year: 2024
DOI:
[PDF]
The Use of Machine Learning In Assessing Future Sustainability of Newly Developed Solar Thermal Systems
Lisa Gobio-Thomas, Mohamed Darwish, Valentina Stojceska
ABSTRACT:
The future environmental impact of an innovative solar thermal system called ASTEP was estimated using artificial neural network (ANN) regression model in MATLAB R2022a. ASTEP consists of three main elements: a novel rotary Fresnel Sundial, thermal energy storage (TES) and the controls. It supplies solar thermal energy to industrial processes of maximum 400°C. This technology has been implemented to two end-users, Mandrekas (MAND) and Arcelor Mittal (AMTP). The initial GHG emissions of MAND & AMTP’s ASTEP systems were 3.45 and 4.6 kgCO2eq/kwh. The following features were used as inputs to the ANN model; plant capacity, plant lifespan, direct normal irradiation (DNI), energy consumption, energy generation, TES, carbon intensity and plant installation year, whilst the output is the GHG emissions. A total of 30 datasets were used in training and testing of the ANN model. 80% of the dataset was used to train the ANN model and 20% used for testing the model. The model achieved a mean square error (MSE) of 5.2. The carbon intensity of grid electricity in the ANN model was changed to 125 kgCO2eq/kwh as predicted for 20230 for the carbon intensity of EU grid electricity. This resulted in ASTEP’s GHG emissions of 1.99 and 2.94kgCO2eq/kwh for 2030 for MAND and AMTP, respectively. The lower GHG emissions can be attributed to the predicted reduction in grid carbon intensity for 2030-2050 as the EU integrates a higher proportion of renewable energy into the electricity grid to achieve net-zero carbon emissions by 2050.
keywords: GHG emissions; environmental impact; artificial intelligence; solar thermal plants; artificial neural network (ANN) regression