English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Multiscale computational modelling is challenging due to the highcomputational cost of direct numerical simulation by finite elements. Toaddress this issue, concurrent multiscale methods use the solution of cheapermacroscale surrogates as boundary conditions to microscale sliding windows. Themicroscale problems remain a numerically challenging operation both in terms ofimplementation and cost. In this work we propose to replace the localmicroscale solution by an Encoder-Decoder Convolutional Neural Network thatwill generate fine-scale stress corrections to coarse predictions aroundunresolved microscale features, without prior parametrisation of localmicroscale problems. We deploy a Bayesian approach providing credible intervalsto evaluate the uncertainty of the predictions, which is then used toinvestigate the merits of a selective learning framework. We will demonstratethe capability of the approach to predict equivalent stress fields in porousstructures using linearised and finite strain elasticity theories.

A Bayesian multiscale CNN framework to predict local stress fields in structures with microscale features