Title :
Application of neural networks to the non-stationary heat conductivity problems
Author :
Trushevskyy, Valeriy
Author_Institution :
Ivan Franko Nat. Univ. of Lviv, Lviv, Ukraine
Abstract :
The paper is devoted to application of recurrent artificial neural networks to solution of non-stationary heat conductivity problems. A recurrent neural network has been constructed for implementation of the multi-grid method. Space-time discretization of the problem was established using the projective-grid scheme of the finite element method (FEM).
Keywords :
finite element analysis; neural nets; parabolic equations; partial differential equations; physics computing; thermal conductivity; FEM; artificial neural networks; finite element method; multigrid method; nonstationary heat conductivity problems; projective grid scheme; recurrent ANN; space-time discretization; Conductivity; Heating; delamination; evolutionary algorithms; finite element method; unilateral contact; variational inequalities;
Conference_Titel :
Modern Problems of Radio Engineering Telecommunications and Computer Science (TCSET), 2012 International Conference on
Conference_Location :
Lviv-Slavske
Print_ISBN :
978-1-4673-0283-8
