Masters Thesis Work
Intelligent Computational Analysis of Magnetohydrodynamic Double-Diffusive Mixed Convection in Nanofluids
Under the Supervision of -
Dr. Mohammad Arif Hasan Mamun
Professor
Department of Mechanical Engineering, BUET
Objectves
In this work, one of the major objectives is to create a novel ANN model using
the simulation data gathered with numerical simulation using FEM for prediction of
different performance parameters in present framework. The objectives of this
investigation can be summarized as follows –
Implementing machine learning techniques along with numerical simulation to
investigate mixed convection phenomena in lid-driven trapezoidal enclosure
filled with different nanofluids and hybrid nanofluids with double rotating
cylinder inside.
Using simulation data to develop data-based Artificial Neural Network (ANN)
Architecture to predict results for this problem and a comparative study between
FEM and ANN technique.
Finding the required parameter values of volume fractions, dimensionless
rotating speed (|Ω|), Richardson Number (Ri), Reynolds Number (Re),
Hartmann Number (Ha), Lewis Number (Le) as well as the inclination angle (λ)
of the magnetic field with magnetic flux density Bo and buoyancy ratio (BR) to
get maximum heat and mass transfer.
Observing dimensionless temperature (Θ), velocity (U, V) and mass
concentration (C) across the computational domain.
Developing various novel Artificial Neural Network (ANN) architectures and
identifying the best architecture among those for the present investigation to
predict and compare results obtained by FEM and ANN.