Price prediction based congestion management using growing RBF neural network

This paper proposes a growing radial basis function (GRBF) neural network based methodology for nodal congestion price (NCP) prediction for congestion management in emerging restructured power system. An unsupervised learning vector quantization (VQ) clustering has been employed as feature selection technique for GRBF neural network as well as for partitioning the power system into different congestion zones. This ensures faster training for proposed neural network and furnishes instant and accurate NCP values, useful for congestion management under real time power market environment. A case study of RTS 24-bus system is presented for demonstrating the computational efficiency and feasibility of this approach.