Publication de 4 numéros par an
ISSN Imprimer: 2152-2057
ISSN En ligne: 2152-2073
Indexed in
LINEAR DYNAMIC NEURAL NETWORK MODEL OF A VISCOELASTIC MEDIUM AND ITS IDENTIFICATION
RÉSUMÉ
For identification of the behavior of viscoelastic media with small deformations the linear dynamic neural network model is suggested. The model realizes the principle of an adaptively hierarchical superstructure. In order to reach the specified level of the identification error (10−12) the model changes its structure automatically from the 3rd to the 24th order of complexity. The neural network model, compared to other known phenomenological models of viscoelastic media, possesses a higher operation speed, allows use of parallel computational procedures, and realizes an adaptively hierarchical principle of construction. A small error of training the linear nonstationary dynamic model without feedback can be reached only in the presence of a huge initial massif of experimental data.
-
Joseph, D. D., Fluid Dynamics of Viscoelastic Liquids.
-
Basistov, Yu. A. and Yanovsky, Yu. G., Ill-posed problems under identification of non-linear rheological models of state.
-
Vainberg, M. M., Variatsionnyi metod i metod monotonnykh operatorov v teorii nelineinykh uravnenii (Variational Method and Method of Monotonic Operators in the Theory of Nonlinear Equations).
-
Yanovsky, Yu. G., Basistov, Yu. A., and Filipenkov, P. A., Problem of identification of rheological behavior of heterogeneous polymeric media under finite deformation.
-
Basistov, Yu. A. and Yanovsky, Yu. G., Identification of mathematical models of viscoelastic media in rheology and electrorheology.
-
Yanovsky, Yu. G., Polymer Rheology: Theory and Practice.
-
Hagan, M. T., Demuth, H. B., and Beale, M. H., Neural Network Design.
-
Chokshi Sagar, Gohil Piyush, Patel Darshan, Experimental investigations of bamboo, cotton and viscose rayon fiber reinforced Unidirectional composites, Materials Today: Proceedings, 28, 2020. Crossref