Cyber-resilient energy modeling in water-energy nexus: Computational intelligence techniques and optimization for promoting agricultural systems sustainability

Abstract

Smart farms, exemplified by smart water-energy nexus (WEN) system, are evolving to integrate renewable energy sources (RESs) such as biomass boilers and wind turbines (WTs) for thermal and electrical demands, alongside water desalination systems to ensure a sustainable fresh water supply. As critical nodes in the water-energy management system (WEMS), these systems face increasing risks from sophisticated cyber threats, which can compromise resource efficiency, stability, and environmental sustainability. This study explores the resilience of a WEN that incorporates battery-based storage, heat retention systems, and stored water reserves facilities, highlighting its representation of smart farms as WEN for optimized resource management. A novel undetectable false data injection (FDI) attack is simulated, targeting electricity and water demands to exploit system vulnerabilities and destabilize the WEMS. To mitigate these risks, a residual long short-term memory (LSTM)-based data modification method, leveraging a residual neural network (ResNet) approach with LSTM model and residual connections, is proposed to restore the integrity of tampered time-series data. The results reveal that FDI attacks can disrupt electricity demand by 2.32% to 21.46% and water demand by 6.29% to 21.47%, jeopardizing the stability of the WEMS. Implementing the proposed correction approach reduces the error, measured as the percentage discrepancy between reconstructed and original values, to 10.31% for power usage and 3.19% for water consumption estimates, showcasing its efficacy in stabilizing the WEMS. The research highlights the value of intelligent algorithms within energy nexus in resource optimization and sustainability. Integrating renewable technologies and data analytics enhances WEN reliability, efficiency, and cyber resilience.