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

Personal portfolio

CREATION NEURAL NETWORKS TO IDENTIFY THE STEAM QUALITY OF THE STEAM-WATER MIXTURE UNDER CERTAIN PRESSURE, TEMPERATURE AND ENTHALPY

Orehova E.E. 1 Abramov A.A. 1 Andreev V.V. 1
1 Nizhny Novgorod State Universiry n.a. R.E. Alekseev
In this paper considered tha application of neural networks to solving the problem of determination of the steam quality setam – water mixture. Were created and analysed neural networks to determinate the degree of dryness (steam content) for a know pressure, temperature and enthalpy of the stem – water mixture. In this study we investigated the neural networks type multilayered perceptron with one and two intermediate layers, with 1, 2 or 3 number of neurons in the layer. Just looked at various activation function the transfer function. Analysis of the obtained results. The analysis revealed architecture of neural networks, solving a task most closely to the existing methods. It is noticed, that the from the networks with one intermediate layer is the most optimal network with sigmoint function activation of networks with two one intermediate layer is the most optimal network with sigmoint function activation hypertangent. In this work was not carried out optimization of parameters of networks. This work is planned in future.