Wyniki wyszukiwania - BazTech

Ograniczanie wyników

Znaleziono wyników: 2

Liczba wyników na stronie

Wyniki wyszukiwania

Sortuj według:

Ogranicz wyniki do:

The modelling of layered rocks using a numerical homogenisation technique and an artificial neural network

Urbański Aleksander, Ligęza Szymon, Przecherski Piotr

Technical Transactions

2023

Vol. 120, iss. 1

art. no. e2023007

A method of creating a constitutive model of layered rocks based on an artificial neural network (ANN) is reported in this work. The ANN gives an implicit constitutive function Ʃⁿ⁺¹= F( Ʃⁿ , ΔE), relating the new state of homogenized stresses Ʃⁿ⁺¹ with the old state Ʃⁿ and with the increment of homogenized strains ΔƩ. The first step is to repeatedly run a strain- controlled homogenisation on an uni-dimensional finite element model of a periodic cell with elastic-plastic models (Drucker-Prager) of the components. Paths are created in (Ʃ, E) space, from which, a set of patterns is formed to train the ANN. A description of how to prepare this data and a discussion on ANN training issues are presented. Finally, the procedure based on trained ANN is put into a finite-element code (ZSoil.PC) as a user-delivered constitutive function. The approach is verified by comparing the results of the developed model basing on ANN with a direct (single-scale) analysis, which showed acceptable accuracy.

Multi-scale modelling of brick masonry using a numerical homogenisation technique and an artificial neural network

Urbański Aleksander, Ligęza Szymon, Drabczyk Marcin

Archives of Civil Engineering

2022

Vol. 68, nr 4

179--197

A new method of creating constitutive model of masonry is reported in this work. The model is not an explicit orthotropic elastic-plastic one, but with an artificial neural network (ANN) giving an implicit constitutive function. It relates the new state of generalised stresses Σn+1 with the old state Σn and with an increment of generalised strains ΔE (plane-stress conditions are assumed). The first step is to run a strain- controlled homogenisation, repeatedly, on a three-dimensional finite element model of a periodic cell, with elastic-plastic models (Drucker-Prager) of the components; thus a set of paths is created in (Σ, ΔE) space. From these paths, a set of patterns is formed to train the ANN. A description of how to prepare these data and a discussion on ANN training issues are presented. Finally, the procedure based on trained ANN is put into a finite-element code as a constitutive function. This enables the analysis of arbitrarily large masonry systems. The approach is verified by comparing the results of the developed model basing on ANN with a direct (single-scale) one, which showed acceptable accuracy.

W pracy przedstawiono sposób tworzenia makro-modelu konstytutywnego muru ceglanego. Przyjmuje się założenia płaskiego stanu naprężenia. Tworzony model nie jest modelem ortotropowym sprężysto-plastycznym, ale jest zbudowany jako sztuczna sieć neuronowa (SSN) dająca niejawną funkcję konstytutywną. Wiąże ona nowy stan naprężeń uogólnionych (sił membranowych) Σn+1 z poprzednim stanem Σn oraz przyrostem odkształceń uogólnionych ΔE. Forma tak utworzonego makro-modelu konstytutywnego jest zgodna z analizą przyrostową problemu statyki w przypadku nieliniowości materiałowych. Składniki muru (cegła i zaprawa) są opisane modelami sprężysto-plastycznymi Druckera-Pragera. Parametry materiałowe składników muru oraz geometria komórki powtarzalnej stanowią dane wejściowe, służące budowie makro-modelu muru.