Managing Wind-based Electricity Generation with Storage and Transmission Capacity

In the past decade wind-based electricity generation has grown rapidly, and is likely to grow even more in the future. Since generation sites—wind farms—tend to be remote, they require transmission lines to transport electricity to markets. These wind farms can be augmented by co-locating them with an electricity storage facility, for example, an industrial battery. This storage facility enables the combined system to stockpile generation that exceeds the transmission capacity, to time-shift the sale of generated electricity to periods of more favorable prices, and to buy electricity from the market for future resale. We consider the problem of operating such a combined generation, storage, and transmission system. We formulate this problem as a finite-horizon Markov decision process (MDP) with stochastic electricity prices and wind energy. We show that an optimal policy for our MDP can be complex, and thus examine heuristics. One heuristic policy ignores the buying option; we prove that this policy has an appealing dual-threshold basestock structure. Our numerical analysis, based on real data, shows the following: (i) the dual-threshold policy is nearly optimal; (ii) there can be significant value from using optimization versus a näıve policy; (iii) incorporating future price and wind information as well as their uncertainty when making operational decisions is relevant; and (iv) the value of storage can be substantial.