Hybrid Deep Neural Network-Based Generation Rescheduling for Congestion Mitigation in Spot Power Market

In the open-access power market environment, continuously varying loading and accommodation of various bilateral multilateral transactions, sometimes leads to congestion, which is not desirable. a day ahead or spot market, generation rescheduling (GR) one most prominent techniques be adopted by system operator (SO) release congestion. this paper, novel hybrid Deep Neural Network (NN) developed for projecting rescheduled dispatches at all generators. The proposed cascaded combination modified back-propagation (BP) algorithm based ANN as screening module NN GR module. segregates congested non-congested scenarios resulting due bilateral/multilateral efficiently accurately. However, projects re-scheduled active generating units minimum congestion cost unseen instantly. present approach provides ready/instantaneous solution manage in market. During training, Root Mean Square Error (RMSE) evaluated minimized. effectiveness method has been demonstrated on IEEE 30-bus system. maximum error incurred during testing phase found 1.191% within acceptable accuracy limits.