Scheduling power-to-ammonia plants considering uncertainty and periodicity of electricity prices

Developing affordable and scalable energy storage solutions are essential to decarbonizing power systems. The conversion of renewable electricity into chemical energy carriers such as ammonia has attracted extensive attention from academia and industry. Many Power-to-Ammonia (PtA) plants have been conceptualized and developed worldwide in recent years. The PtA plant is an integration of multiple electrochemical processes, each with a distinct set of operational constraints and cost structure. One of the problems in the operation of PtA plants is the optimal scheduling of the hydrogen buffer in PtA plants considering the operational characteristics of electrochemical processes and the volatility and uncertainty of electricity prices. In this paper, a two-stage Markov-Decision-Process (MDP) approach is proposed. The computational challenges brought by the infinite optimization horizon and non-concavity of cost functions are resolved. The first stage solution is based on the periodic MDP approach, which captures the periodic structure of electricity prices. The second stage solution gives optimal real-time decisions based on a rolling-horizon MDP approach. Numerical results show that the accurate representations of the cost functions and the optimization horizon using the proposed method are necessary, while the linearization of cost functions and the truncation of the optimization horizon lead to notable deviations from the optimality.