On the use of dynamic programming for optimal energy management of grid-connected reversible solid oxide cell-based renewable microgrids

Highlights: • Dynamic programming-based power split optimization in clean grid-connected rSOC-based microgrids. • Model order reduction is a key methodology to support the co-design of plant and related control strategies. • High temperature reversible fuel cell emerges as a key technological enabler of hydrogen economy. • rSOC capacity is better exploited if the scaling-down enabled by grid connection is properly accomplished. • Co-design exhibits high potential for medium-term deployment of rSOC-based clean polygenerators. In reversible solid oxide cell (rSOC)-based renewable microgrids connected to the network, control logic optimizing the power split with the grid is needed. The benefits associated with such configuration are assessed and compared to an islanded reversible microgrid, previously designed via a model-based approach. Multiple loads are met, including those of a residential complex as well as electric and fuel cell vehicle fleets. The energy storage system consists of a rSOC and hydrogen and thermal storage tanks. The targeted optimal power split annual trajectory is found through dynamic programming. Suitable simplifying assumptions are introduced to develop a fast reduced-order sub-model from the original islanded rSOC microgrid, thus conceiving the optimal control problem as a one state dynamic programming task. Due to the connection to the grid, the energy storage system no longer has to compensate for the difference between generation and demand, thus enabling economic rSOC sub-scaling. Finding the best control strategy and resizing jointly lead to capital and operating expense reductions. The optimization outcomes indicate how a 60% reduction in rSOC nominal power allows for a simple payback period of 40% less than for the islanded design, as well as for proper rSOC capacity exploitation.