Cascaded deep NN-based customer participation by considering renewable energy sources for congestion management in deregulated power markets

Abstract

Continuously varying loading conditions and the cost-based operation of a competitive power market lead to the problem of congestion as one of the most crucial issues. In day-ahead power market operation (PMO), customer participation (CP) and generation rescheduling (GR) are the most effective techniques preferred by the system operator to eliminate congestion. In this paper, a cascaded Deep Neural Network (DNN) module has been presented for estimating customer participation and power generated by Wind Energy Source (WES) as on-site generation (OSG) to manage congestion. The proposed module is a cascade combination of Artificial Neural Network (ANN) as a filtering module (FM) and DNN as a congestion management (CM) module. The CM module estimates the customer participation for all receptive costumers, power supplied by wind energy sources under uncertain conditions and generation rescheduling of all generators with minimum cost for all unseen congested power system loading patterns. The proposed CM approach provides an instant and efficient solution to manage congestion with minimum cost. The developed module has been examined on IEEE 30-bus power system. The maximum error found in the testing phase is 1.1865% which is very less and within the acceptable limit.