This paper proposes a model to carry out analysis of
storage facilities operation including a transmission network. The model
represents short-term storage operation in an approxi-mated way that reduces
computational requirements, which makes it suitable for medium and long-term
operational planning in power systems with a high level of renewable energy
penetra-tion.
In the proposed model, we cluster hourly data using the
so-called system-states framework developed in previous work. Within this
framework, non-consecutive similar time periods are grouped, while
chronological information is represented by a tran-sition matrix among states.
We extend the system-state framework from a single-bus system to a transmission
network.
We define and analyze two alternative sets of
representative variables for clustering hours to obtain system states when the
transmission network is considered. This extension of the system states
framework allows us to evaluate the impact of transmission congestions in
medium- and long-term planning models in a rea-sonable computation time.
A case study shows that the proposed model is 235 times
faster than an hourly approach, used as benchmark, whereas the overall system
cost is approximated with less than 2% error. The overall charging/discharging
trends are similar enough to those of the hourly model, being hydro storage
better approximated than fast-ramping batteries. Besides, for the analyzed case
study, it is shown how congestion in the transmission network in fact improves
the accuracy of the proposed approach.