Decarburization of building is a crucial part of the era of carbon neutrality. This paper considers the energy management of a sustainable residential cogeneration system consisting of fuel cell, battery pack, photovoltaic (PV), thermal energy system (TES) and heat pump. To overcome the difficulties caused by the uncertainties in PV power and loads, an optimal economic day-ahead scheduling problem is formulated, subject to the constraints on the balance of supply and demand of the heat and electricity. Random variables are modeled based on the time-varying transition matrix of Markov chain. Furthermore, stochastic dynamic programming (SDP) algorithm is used to minimize the system operation cost, i.e., total fuel cost of the fuel cell in 24 hours under uncertainty. Simulation results show that the derived control strategies of each stage are energy-efficient in managing the power output of fuel cell and heat pump.

fuel cell;sustainable cogeneration;stochasitc dynamic programming;optimal scheduling