A double-layer many-objective stochastic optimization model to handle many uncertainties in the operation of smart energy systems

With rising interactions and interdependencies among diverse energy carriers, higher renewable energy penetration, heightened cybersecurity risks, evolving energy market rules and decentralized energy models, electrification of transportation, and applying new government policies, regulations, and environmental standards, new uncertainties are introduced to the operation of smart energy systems (SESs) and the existing ones have been escalated. Traditional techniques require sufficient information on uncertainties (e.g., probability distribution functions) to examine their impacts on the operation of the SES. In practical cases, it is difficult and even impossible to access such information. Therefore, it is imperative to develop well-suited models based on efficient techniques to address these challenges. Hence, this work introduces a double-layer many-objective stochastic optimization model to handle many uncertainties in the operation of SESs. The upper layer simultaneously optimizes the horizon of many uncertainties considering their interactions. Uncertainties are electrical, heating, and cooling power demands, the price of electricity and gas, and the production capacity of the photovoltaic system. The boundaries set by minimum and maximum limits for the horizon of all uncertainties, along with the entirety of the lower-layer optimization problem, serve as constraints for the upper-layer many-objective optimization problem. Information-gap decision theory is applied at the upper layer to examine the adverse and beneficial impacts of many uncertainties on the operation of the SES by developing robustness and opportunity functions, respectively. Furthermore, the non-dominated sorting genetic algorithm III is used to tackle the upper-level many-objective optimization problem and find the six-dimension set of Pareto efficient solutions. Moreover, the optimal solution within Pareto efficient solutions is determined using a combination of the fuzzy satisfying method and the conservative methodology—the min-max formulation. In the lower layer, a computational core for the upper layer, the operation of the SES is formulated using mixed-integer linear programming to minimize the operation and emission costs subject to technical and logical constraints of the SES.