Towards mega-scale decarbonized industrial park (Mega-DIP): Generative AI-driven techno-economic and environmental assessment of renewable and sustainable energy utilization in petrochemical industry

Abstract

In alignment with climate change mitigation plans, renewable and sustainable energy integration into the industrial sector should be investigated for decarbonization purposes by 2050. This study aims to examine the feasibility of decarbonization of mega-scale industrial parks with two emerging technologies; first, an integrated energy system involving an air separation unit (ASU) and a liquid air energy storage (LAES)-based power generation system (ALPG) with the assistance of a large-scale liquified natural gas (LNG) plant and second, the green electricity predicted by various renewable energy (VRE) scenarios using an AI-driven generative model concerning techno-economic and environmental analyses. Firstly, the electricity demand of mega-scale petrochemical industrial complexes was allocated, and remote sensing data of VRE (wind and solar) were collected in South Korea. Then, the ALPG was modeled to consider the actual operation of an LNG plant near target complexes. Subsequently, the CC-informed VRE scenario generation model (CC-VRES) based on an adversarial autoencoder (AAE) algorithm was developed to generate VRE scenarios considering the CC effects. Finally, techno-economic and environmental assessments were conducted to target the mega-DIP by 2050. The results showed that the levelized cost of electricity of the proposed ALPG was 150 $/MWh, which is enough to supply the required sustainable electricity, considering the Korean energy plan and CC effects in the CC-VRES model. Therefore, the excess VRE can be used to decarbonize the target industrial parks; it is anticipated that 26,084 GW h/yr of renewable and sustainable electricity will be generated and cover the total electricity demand of Yeosu and Ulsan target complexes while reducing 13,943-kilo t CO2 eq./yr through 2050.