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

Personal portfolio

PARAMETER OPTIMIZATION OF THE INTELLIGENCE DIAGNOSTIC SYSTEM OF HIGH-VOLTAGE SWITCHES

Dubrov V.I. 1
1 South Russian State Technical University (Novocherkassk Polytechnic Institute)
It describes the steps for creating the universal intelligent diagnostic algorithm based on neural networks and show its application for the diagnosis of high-voltage switchgear. The main attention is paid to the final stage of the development of diagnostic algorithm, which is to optimize the neural network parameters such as the number of neurons in the layer, the number of hidden layers in the neural network, the speed of network training, the number of epochs (iterations) for network learning. It contains the following stages of development: a diagnostic algorithm, methods of information signal processing, the solution of optimization problems. Presents the developed algorithm of informative range selecting from excess measurement data flow and optimization capabilities of genetic algorithms applied to optimize the neural network of diagnostic system. The algorithms are simulated in the environment of Octave, high-voltage switch of type MCP 110M was chosen as the object of diagnosis.