Short-Term Congestion Forecasting in Wholesale Power Markets

Short-term congestion forecasting is highlyimportant for market participants in wholesale power marketsthat use Locational Marginal Prices (LMPs) to managecongestion. Accurate congestion forecasting facilitates markettraders in bidding and trading activities and assists marketoperators in system planning. This study proposes a new shorttermforecasting algorithm for congestion, LMPs, and otherpower system variables based on the concept of systempatterns—combinations of status flags for generating units andtransmission lines. The advantage of this algorithm relative tostandard statistical forecasting methods is that structural aspectsunderlying power market operations are exploited to reduceforecast error. The advantage relative to previously proposedstructural forecasting methods is that data requirements aresubstantially reduced. Forecasting results based on a NYISO casestudy demonstrate the feasibility and accuracy of the proposedalgorithm.[...]