Використання фізично-обґрунтованої нейромережі для визначення концентрації компонентів скрапленого нафтового газу
Б. П. Книш, О. В. Лесько, B. Knysh et al. · Repository of Vinnytsia National Technical University (Vinnytsia National Technical University) · 2026
A method for determining the concentration of liquefied petroleum gas components using a physics-informed neural network is proposed. The approach combines experimental measurements of mixture density with physical relations describing the temperature dependence of component densities. Incorporation of physical constraints into the loss function of the neural network ensures consistency with mixture laws and improves the accuracy of estimating propane and butane concentrations.
Cited by 0 works. Open-access PDF