Explainable AI-driven net-zero carbon roadmap for petrochemical industry considering stochastic scenarios of remotely sensed offshore wind energy

Abstract

Recently, several countries have focused on industrial decarbonization to ensure that the net emissions of carbon dioxide reach zero and contribute to decreasing the global mean temperature. This study aims to investigate the feasibility of a net-zero carbon roadmap for the petrochemical industry by 1) developing an explainable artificial intelligence (XAI)-based generative model to produce stochastic scenarios for offshore wind power in an elec-trical grid, and 2) conducting techno-economic and environmental assessments for forecasting models that define offshore wind power networks. Firstly, data processing techniques were utilized on remotely sensed offshore wind speed datasets and energy data obtained from petrochemical industrial parks. Second, a generative model was designed using a variational autoencoder (VAE) to produce different forecast scenarios of offshore wind power. Third, stochastic scenarios were developed by considering behavioral characteristics of offshore wind power with stochastic scenario uncertainties. Finally, we investigated the techno-economic and environmental assessments of the proposed renewable energy networks. Comparing to the case of offshore wind power in 2022, The proposed XAI-driven net-zero carbon roadmap indicated that the total cost of electricity generation and fossil fuel cost with carbon capture and storage (CCS) can be reduced by 4% and 42% with 41% reduction of CO2 emission. Ultimately, this may be implemented in industrial parks to aid in efforts towards carbon neutrality.