Rollout strategies for real-time multi-energy scheduling in microgrid with storage system

The rising climatic concerns force the conventional power system to minimise the gas emissions. Microgrid integrating with renewable energy plays a key role in utilising power systems in a more environment friendly way. Compared with single renewable energy system, the microgrid with multi-energy systems is more attractive in practice. The schedule problem is considered to schedule wind power, combined heating and power generation (CHP), battery, and power grid in order to satisfy the electricity load and heat loads in the microgrid with the minimal cost in real time. Owing to the stochastic nature of wind power generation, it is a challenge to solve the multi-energy scheduling problem when using the obtained real-time information. In this study, the joint scheduling of the microgrid with multi-energy systems is formulated as a rolling horizon Markov decision process (MDP). Moreover, then the rollout algorithm is applied to solve the problem in order to deal with the large state and decision space of MDP, which is caused by the increase of the number of CHPs. The feasible base policy needed in the rollout algorithm is constructed using a greedy algorithm. Numerical results demonstrate that the algorithm can achieve the requirement of real-time scheduling effectively.