Evaluating market designs in power systems with high wind penetration

In this paper we assess the incidence of strategic behavior of demand and supply units in market set ups which should enable the efficient in-feed of renewable energy sources. Therefore, a sequence of energy and reserve power auctions is modeled in an agent-based framework in order to approximately come up with the drawbacks of the different market designs. The main focus lies on the reserve power market. The generation companies as well as representative consumers are facilitated with Q-learning in order to exploit market flaws. The representative network includes a renewable in-feed, which is characterized by low marginal generation costs. In terms of congestion management, we assume in a first step uniform marginal pricing, in a second step nodal pricing for the energy power auction.