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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

This paper presents a machine learning methodology to improve the predictionsof traditional RANS turbulence models in channel flows subject to strongvariations in their thermophysical properties. The developed formulationcontains several improvements over the existing Field Inversion MachineLearning (FIML) frameworks described in the literature, as well as thederivation of a new modelling technique. We first showcase the use of efficientoptimization routines to automatize the process of field inversion in thecontext of CFD, combined with the use of symbolic algebra solvers to generatesparse-efficient algebraic formulas to comply with the discrete adjoint method.The proposed neural network architecture is characterized by the use of aninitial layer of logarithmic neurons followed by hyperbolic tangent neurons,which proves numerically stable. The machine learning predictions are thencorrected using a novel weighted relaxation factor methodology, that recoversvaluable information from otherwise spurious predictions. The study uses theK-fold cross-validation technique, which is beneficial for small datasets. Theresults show that the machine learning model acts as an excellent non-linearinterpolator for DNS cases well-represented in the training set, and thatmoderate improvement margins are obtained for sparser DNS cases. It isconcluded that the developed machine learning methodology corresponds to avalid alternative to improve RANS turbulence models in flows with strongvariations in their thermophysical properties without introducing priormodeling assumptions into the system.

Machine Learning for RANS Turbulence Modelling of Variable Property Flows