An effective MILP model for the optimal design of microgrids with high-reliability requirements

This work presents a Mixed Integer Linear Programming formulation to optimize the design of microgrids and aggregated energy systems, including all the required constraints to guarantee N-1 reliability. In comparison to previous studies, the proposed model accounts for not only the optimal redundancy of generators but also the necessary power reserves (spinning and upward) during the start-up period of spare generators, considering also potential fluctuations in renewable energy production and electric demand. Previous studies have not taken these reserve requirements into account, leading to overoptimistic cost estimates. The methodology is applied to the optimization of two real-world off-grid case studies of industrial relevance: (i) a remote natural gas liquefaction plant and (ii) an off-grid airport. The proposed N-1 reliability approach yields optimal solutions with a lower renewable capacity compared to classic methods. Additionally, it requires a larger storage size as the battery is not only operated jointly with renewables, but it is also used as a power bank to provide spinning and upward reserves. The increase in total annual costs due to the reliability requirements is estimated at 2.8 % and 10.4 % for the two case studies, resulting from higher investment and operating costs. Compared to conventional N-1 reliability approaches, which do not include reserves for the start-up time of spare generators, the cost increase ranges from 2 to 5 %. • MILP model for the optimal design of off-grid plants with reliability requirements • Ad-hoc reserve constraints to cover the start-up time of the spare generator • The approach is applied to two real-world off-grid plants of industrial relevance • The reliability and reserve constraints cause a 3–10 % cost increase.