Dispatch with Energy Storage through Constructing the Optimal Cost-to-go Function

The stochastic future can be reduced to a cost-to-go value function of the system state variable for fast and optimal real-time decision making, following the concept of dynamic programming, this serves as a potential solution to the ongoing real-time energy storage dispatch problem, however finding the optimal cost-to-go function for the complex power system remains a challenge. This study provides an algorithmic insight into constructing a piece-wise linear cost-to-go function for energy storage state-of-charge through scenario-based numerical simulations. The proposed method has much lower computation complexity and can be implemented in parallel, allowing the power system economic dispatch to adapt to the newest forecast at the earliest possibility.