作者: J.M. Molina , P. Isasi , A. Berlanga , A. Sanchis
DOI: 10.1016/S0952-1976(00)00009-9
关键词:
摘要: Abstract The use of Neural Networks (NN) is a novel approach that can help in taking decisions when integrated more general system, particular with expert systems. In this paper, an architecture for the management hydroelectric power plants introduced. This relies on monitoring large number signals, representing technical parameters real plant. composed Expert System and two NN modules: Acoustic Prediction (NNAP) Predictive Maintenance (NNPM). NNAP based Kohonen Learning Vector Quantization (LVQ) order to distinguish sounds emitted by electricity-generating machine groups. NNPM uses ART-MAP identify different situations from plant state variables, prevent future malfunctions. addition, special process generate complete training set has been designed module. developed deal absence data about abnormal situations, neural nets trained backpropagation algorithm.