CBRFNN-based short-term nodal load forecasting for middle voltage distribution networks
Yingzhe Yu, Jianzhong Wu · Proceedings of the CSEE · 2004
A case-based-reasoning fuzzy-neural-network (CBRFNN) is presented based on cognitive science and Parallel Distributed Processing (PDP) model. The principle of CBRFNN is analyzed, the elementary architecture of CBRFNN is defined, and a hybrid (supervised/unsupervised) learning algorithm is also proposed, which equips CBRFNN with a good generalization capability. All nodes in a CBRFNN are created dynamically by the rapid and incremental learning procedure. CBRFNN can withstand the effect of bad data effectively through network self-organizing. The proposed method can solve the short-term nodal load-forecasting problem for middle voltage distribution network, which belongs to the kind of problems that are characterized by incomplete and inaccurate information.