Scientific journal
Scientific Review. Technical science
ISSN 2500-0799
ПИ №ФС77-57440

Personal portfolio

INTELLIGENT DIAGNOSIS OF ELECTROMAGNETIC MECHANISM BASED ON NEURAL NETWORK TECHNOLOGY

Dubrov V.I. 1
1 South Russian State Technical University (Novocherkassk Polytechnic Institute)
Describes stages of adaptation of the intelligent diagnostic algorithm of high-voltage switchgear for the diagnostic of electromagnetic mechanisms. In the article research the possibility of the application of wavelet spectral processing of characteristics of electromagnetic mechanisms. Proposed adapted diagnostic algorithm of electromagnetic mechanisms. In the developed algorithm is introduced informative function, which greatly improves the accuracy of diagnosis in the next stage with the use of the intellectual apparatus of the neural network. To confirm the results given neural network learning error calculated using different electromagnetic characteristics. The algorithms are simulated in the mathematical modeling system “Matlab”, as the object of diagnosis was chosen electromagnet of Inventio type. The measured properties are obtained with the use of modern measuring equipment - a device MagHyst, providing sufficient accuracy current I (t) and flux linkage Psi (t) in the coil of the electromagnet.