Robustness of convergence demonstrated byparametric-guiding andcomplex-root-tunneling algorithms for Bratu’s problem

Purpose This study aims to propose the parametric-guiding algorithm, the complex-root (CR) tunneling algorithm and the method that integrates both algorithms for the heat and fluid flow (HFF) community, and apply them to nonlinear Bratu’s boundary-value problem (BVP) and Blasius BVP. Design/methodology/approach In the first algorithm, iterations are primarily guided by a diminishing parameter that is introduced to reduce magnitudes of fictitious source terms. In the second algorithm, when iteration-related barriers are encountered, CRs are generated to tunnel through the barrier. At the exit of the tunnel, imaginary parts of CRs are trimmed. Findings In terms of the robustness of convergence, the proposed method outperforms the traditional Newton–Raphson (NR) method. For most pulsed initial guesses that resemble pulsed initial conditions for the transient Bratu BVP, the proposed method has not failed to converge. Originality/value To the best of the authors’ knowledge, the parametric-guiding algorithm, the CR tunneling algorithm and the method that integrates both have not been reported in the computational-HFF-related literature.

Some New Characterizations of The Harmonic and Harmonic 1-Type Curves in Euclidean 3-Space

A Laplace operator and harmonic curve have very important uses in various engineering science such as quantum mechanics, wave propagation, diffusion equation for heat, and fluid flow. Additionally, the differential equation characterizations of the harmonic curves play an important role in estimating the geometric properties of these curves. Hence, this paper proposes to compute some new differential equation characterizations of the harmonic curves in Euclidean 3-space by using an alternative frame named the N-Bishop frame. Firstly, we investigated some new differential equation characterizations of the space curves due to the N-Bishop frame. Secondly, we firstly introduced some new space curves which have the harmonic and harmonic 1-type vectors due to alternative frame N-Bishop frame. Finally, we compute new differential equation characterizations using the N-Bishop Darboux and normal Darboux vectors. Thus, using these differential equation characterizations we have proved in which conditions the curve indicates a helix.

Double-Layer Metal Foams for Further Heat Transfer Enhancement in a Channel: An Analytical Study

A local thermal non-equilibrium analysis of heat and fluid flow in a channel fully filled with aluminum foam is performed for three cases: (a) pore density of 5 PPI (pore per inch), (b) pore density of 40 PPI, and (c) two different layers of 5 and 40 PPI. The dimensionless forms of fully developed heat and fluid flow equations for the fluid phase and heat conduction equation for the solid phase are solved analytically. The effects of interfacial heat transfer coefficient and thermal dispersion conductivity are considered. Analytical expressions for temperature profile of solid and fluid phases, and also the channel Nusselt number (NuH) are obtained. The obtained results are discussed in terms of the channel-based Reynolds number (ReH) changing from 10 to 2000, and thickness ratio between the channel height and sublayers. The Nusselt number of the channel with 40 PPI is always greater than that of the 5 PPI channel. It is also greater than the channel with two-layer aluminum foams until a specific Reynolds number then the Nusselt number of the channel with two-layer aluminum foams becomes greater than the uniform channels due to the higher velocity in the outer region and considerable increase in thermal dispersion.

Deep learning or interpolation for inverse modelling of heat and fluid flow problems?

Purpose The purpose of this study is to compare interpolation algorithms and deep neural networks for inverse transfer problems with linear and nonlinear behaviour. Design/methodology/approach A series of runs were conducted for a canonical test problem. These were used as databases or “learning sets” for both interpolation algorithms and deep neural networks. A second set of runs was conducted to test the prediction accuracy of both approaches. Findings The results indicate that interpolation algorithms outperform deep neural networks in accuracy for linear heat conduction, while the reverse is true for nonlinear heat conduction problems. For heat convection problems, both methods offer similar levels of accuracy. Originality/value This is the first time such a comparison has been made.