The problem of signals’ parameters estimation in power systems based on neural network approach is considered. Architectures of neurons, networks and their real time learning algorithms are proposed. The results allow to improve the reliability and effectiveness of monitoring and diagnostics systems, especially under conditions of emerging damages in the power networks in the presence of outliers.
artificial neuron network, signal processing, technical diagnostics
“Neirosetevoi podkhod k otsenke parametrov signalov v elektroenergeticheskikh sistemakh”,
Information Processing Systems,