Optimal management of a hybrid and isolated microgrid in a random setting

Nowadays, governments and electricity companies are making efforts to increase the integration of renewable energy sources into grids and microgrids, thus reducing the carbon footprint and increasing social welfare. Therefore, one of the purposes of the microgrid is to distribute and exploit more zero emission sources. In this work, a Stochastic Unit Commitment of a hybrid and isolated microgrid is developed. The microgrid supplies power to satisfy the demand response by managing a photovoltaic plant, a wind turbine, a microturbine, a diesel generator and a battery storage system. The optimization problem aims to reduce the operating cost of the microgrid and is divided into three stages. In the first stage, the uncertainties of the wind and photovoltaic powers are modeled through Markov processes, and the demand power is predicted using an ARMA model. In the second stage, the stochastic unit commitment is solved by considering the system constraints, the renewable power production, and the predicted demand. In the last stage, the real-time operation of the microgrid is modeled, and the error in the demand forecast is calculated. At this point, the second optimization problem is solved to decide which generators must supply the demand variation to minimize the total cost. The results indicate that the stochastic models accurately simulate the production of renewable energy, which strongly influences the total cost paid by the microgrid. Wind production has a daily impact on total cost, whereas photovoltaic production has a smoother impact, shown in terms of general trend. A comparison study is also considered to emphasize the importance of correctly modeling the uncertainties of renewable power production in this context.