DC microgrid operation with hybrid energy storage considering islanding constraints and demand response coordination: A bi-level Stackelberg game approach

Abstract

DC microgrid (DC$\mu$G) is becoming popular for niche applications due to multiple advantages over AC microgrids ($\mu$G). However, operation of a DC$\mu$G is challenging due to uncertainties of renewable energy source (RES) generation and load demands, limited availability of controllable generation, and unintended islanding events. Sectoral coupling between electricity and hydrogen ($H2$), hybrid energy storage system (HESS), and demand response (DR) implementation address the challenges and enhance the techno-economic benefits of DC$\mu$G operation. Further, incorporating islanding constraints in the scheduling strategy improves the security of system operation. The objective of this paper is to develop an energy management scheme (EMS) for an electricity-$H2$ grid-connected DC$\mu$G with a HESS incorporating islanding constraints and DR implementation in an uncertain environment with correlated and uncorrelated input uncertainties to maximize the profit of the DC$\mu$G operator (DC$\mu$GO), minimize the electricity usage cost of consumers, and ensure secure operation after unintended islanding using bi-level optimization.DC$\mu$G network level, equipment level, and consumer’s apparatus level operating security constraints are considered in the EMS. Uncertainties of input random variables (RV) and their correlation are modelled using Copula theory and incorporated in the EMS. The DC$\mu$G consists of a gas turbine (GT), power to hydrogen (P2H), hydrogen to power (H2P), HESS (comprising battery energy storage system (BESS) and hydrogen storage system (HSS)), wind power generation (WPG), solar power generation (SPG), and consumers. The consumers have non-flexible and flexible loads (thermostatically controlled load (TCL) and plug-in hybrid electric vehicles (PHEV)). The proposed EMS is modelled using a bi-level leader–follower Stackelberg game (SG) architecture, in which the DC$\mu$GO is the leader and the consumers are followers. The DC$\mu$GO optimally schedules flexible resources within its control and sets the retail power price (RPP) to maximize the operating profit. Consumers participate in the DR program by adjusting flexible demands according to the RPP to minimize the cost of electricity use. The dynamic RPP acts as the bridge between the upper and lower-level problems. The bi-level EMS is reformulated as a single-level mixed-integer linear programming (MILP) problem by successively using Karush–Kuhn–Tucker (KKT) conditions, the big-M method, and the strong duality theory. The MILP problem is solved in the MATLAB environment with the YALMIP toolbox and GUROBI solver. Simulation studies reveal that the proposed approach balances the interests of the DC$\mu$GO and the consumers, ensures secure operation after unintended islanding, reduces RES curtailment, reduces the electricity usage cost of the consumers, and enhances the profit of the DC$\mu$GO. For the system under study, the profit of the DC$\mu$GO increases by $\sim 2.22\text{\%}$ while the cost of energy use of the flexible consumers is reduced by $\sim 18.05\text{\%}$.