Jayathi Murthy

A Reynolds-averaged Navier-Stokes solver using unstructured mesh-based finite-volume scheme

A Reynolds-Averaged Navier-Stokes (RANS) solver for unstructured meshes is presented and validated for a number of turbulent flows. The RANS solver employs a cellcentered, second-order accurate finite-volume discretization based on linear reconstruction, in conjunction with a pressure-based segregated solution procedure, k-e models are used for turbulence closure with wall functions. Quadrilateral and triangular meshes with or without local refinement are employed in the validations. It is demonstrated that the proposed solver can predict the selected flows with good accuracy, and that the quality of the predictions is comparable to that of structured mesh based solvers. It is also shown that local mesh adaptation provides an economical means of resolving the critical regions of the flow.

Efficient automatic discrete adjoint sensitivity computation for topology optimization - Heat conduction applications

Purpose Topology optimization is a method for developing optimized geometric designs by distributing material pixels in a given design space that maximize a chosen quantity of interest (QoI) subject to constraints. The central objective of the paper is to develop a problem-agnostic automatic differentiation (AD) framework to compute sensitivities of the QoI required for density distribution based topology optimization in an unstructured co-located cell-centered finite volume framework. Using this AD framework, we develop and demonstrate the topology optimization procedure for multi-dimensional steady state heat conduction problems. Design/methodology/approach Topology optimization is performed using the well-established solid isotropic material with penalization (SIMP) approach. The method of moving asymptotes (MMA), a gradient based optimization algorithm, is used to perform the optimization. The sensitivities of the QoI with respect to design variables, required for optimization algorithm, are computed ...

Thermo-optical properties of packed nanoparticle thermal interface materials

Heat dissipation has become one of the biggest challenges to improving the performance, reliability and power density of electronic devices. One of the crucial components in electronics packaging that must be further developed to address this challenge is thermal interface materials (TIMs). Recently, packed SiO2 nanoparticles, which could serve as TIMs when embedded in a polymer matrix, have been shown to greatly enhance thermal transport due to near-field electromagnetic interactions. Here we investigate the optical excitations of these types of nanoparticle configurations by solving Maxwells equations by applying the Finite Difference Frequency Domain technique. We show that near-field thermal energy enhancement between the nanoparticles is affected by different configurations of nanoparticles. We also show that by changing the particle spacing, the surface plasmonic effects could be tuned to achieve high absorption in a narrow spectral range, corresponding to a high density of states of thermal emission. Results are compared to first-principle calculations of the near-field spectral thermal emission with excellent agreement. These results give insight into the electromagnetic interactions in high thermal conductivity packed particle beds and will aid in the design of new TIM materials.

Effect of Substrate and Nanoparticle Spacing on Plasmonic Enhancement in Three-Dimensional Nanoparticle Structures

Surface plasmon polaritons associated with light-nanoparticle interactions can result in dramatic enhancement of electromagnetic fields near and in the gaps between the particles, which can have a large effect on the sintering of these nanoparticles. For example, the plasmonic field enhancement within nanoparticle assemblies is affected by the particle size, spacing, interlayer distance, and light source properties. Computational analysis of plasmonic effects in three-dimensional (3D) nanoparticle packings are presented herein using 532 nm plane wave light. This analysis provides insight into the particle interactions both within and between adjacent layers for multilayer nanoparticle packings. Electric field enhancements up to 400-fold for transverse magnetic (TM) or Xpolarized light and 26-fold for transverse electric (TE) or Y-polarized light are observed. It is observed that the thermo-optical properties of the nanoparticle packings change nonlinearly between 0 and 10 nm gap spacing due to the strong and nonlocal near-field interaction between the particles for TM polarized light, but this relationship is linear for TE polarized light. These studies help provide a foundation for understanding micro/ nanoscale heating and heat transport for Cu nanoparticle packings under 532 nm light under different polarization for the photonic sintering of nanoparticle assemblies. [DOI: 10.1115/1.4037770]

Use of Detailed Particle Melt Modeling to Calculate Effective Melt Properties for Powders

Selective laser melting (SLM) is a widely used powder-based additive manufacturing process. However, it can be difficult to predict how process inputs affect the quality of parts produced. Computational modeling has been used to address some of these difficulties, but a challenge has been accurately capturing the behavior of the powder in a large, bed-scale model. In this work, a multiscale melting model is implemented to simulate the melting of powder particles for SLM. The approach employs a particle-scale model for powder melting to develop a melt fraction–temperature relationship for use in bed-scale simulations of SLM. Additionally, uncertainties from the particle-scale are propagated through the relationship to the bed scale, thus allowing particle-scale uncertainties to be included in the bed-scale uncertainty estimation. Relations, with uncertainty, are developed for the average melt fraction of the powder as a function of the average temperature of the powder. The utility of these melt fraction–temperature relations is established by using them to model phase change using a continuum bed-scale model of the SLM process. It is shown that the use of the developed relations captures partial melt behavior of the powder that a simple melting model cannot. Furthermore, the model accounts for both uncertainty in material properties and packing structure in the final melt fraction–temperature relationship, unlike simple melting models. The developed melt fraction–temperature relations may be used for bed-scale SLM simulations with uncertainty due to particle effects. [DOI: 10.1115/1.4038423]

Analysis of near-field thermal energy transfer within the nanoparticles

Nanoscale size effects bring additional near-field thermal considerations when heating nanoparticles under high laser power. Scanning electron micrographs of a typical copper nanoparticle powder bed reveal that the nanoparticles are distributed log-normally with 116 nm mean radius and 48 nm standard deviation. In this paper, we solve Maxwell’s equations in frequency domain to understand near-field thermal energy effects for different standard deviations. Log-normally distributed copper nanoparticle packings which have 116 nm mean radius with 3 different standard deviations (12, 48 and 84 nm) are created by using Discrete Element Model (DEM) in which certain number of particles are generated, specifying a position and radius for each. The solid particles interacting with the neighbouring particles are to be distributed randomly into the bed domain with an initial velocity and a boundary condition, which creates the particle packing within a defined time range under gravitational and weak van der Waals forces. Finite Difference Frequency Domain analysis, which yields the electromagnetic field distribution, is applied by solving Maxwells equations to obtain absorption, scattering and extinction coefficients. We show that different particle distributions create different plasmonic effects in the bed domain which results in non-local heat transport. We calculate the surface plasmon effect due to the electromagnetic coupling between the nanoparticles and the dielectric medium under the different distributions. This analysis helps to reveal how sintering quality can be enhanced by creating stronger laser-particle interactions for specific groups of nanoparticles.