Exploring the effects of interdependencies on energy systems in smart communities: A multi-domain modeling and quasi-Monte Carlo sensitivity analysis

An efficient smart and connected community (SCC) depends on the interconnectivity of essential infrastructure systems. However, current modeling tools are unable to determine which interconnections are most important to include, particularly as system dynamics become more complex with high-order effects. To bridge this gap, we propose a comprehensive framework that incorporates multi-layers, multi-blocks, and multi-agents to model interdependent infrastructure systems. Interconnections span cyber, physical, and logical aspects, including human interactions. With the equation-based object-oriented language Modelica, we model energy, transportation, communication, and water systems for a hypothetical SCC and assess higher-order interdependency effects during normal operation. Additionally, we develop a quasi-Monte Carlo sensitivity analysis framework and use variance-based sensitivity metrics to assess the impact of interdependencies on energy system operation. Compared to the decoupled baseline system, the energy consumption of logical interdependency cases varied by 23.3%, the cyber interdependency by 2.0%, and the nested global interdependency by 21.5%. The sensitivity analysis further revealed that interrelationships are not linear nor quadratic, but involve higher-order interactions between parameters. Specifically, occupancy and cyber delays had significant first-order effects. Road delays had a significant higher-order effect, which corresponded to a stronger influence on other model parameters. By modeling higher-order cascading dependencies, our proposed framework has the potential to improve the planning, operation, and co-design of SCCs by quantifying the importance of complex system interactions.