Comparative Evaluation of Algorithms for Achieving Ultrapacked Thermal Greases: Microstructural Models and Effective Behavior

2020 ◽  
Vol 142 (4) ◽  
Author(s):  
Sukshitha Achar P. L ◽  
Huanyu Liao ◽  
Ganesh Subbarayan

Abstract In this work, we develop and evaluate algorithms for generating ultrapacked microstructures of particles. Simulated microstructures reported in the literature rarely contain particle volume fractions greater than 60%. However, commercially available thermal greases appear to achieve volume fractions in the range of 60–80%. Therefore, to analyze the effectiveness of commercially available particle-filled thermal interface materials (TIM), there is a need to develop algorithms capable of generating ultrapacked microstructures. The particle packing problem is initially posed as a nonlinear programming problem, and formal optimization algorithms are applied to generate microstructures that are maximally packed. The packing efficiency in the simulated microstructure is dependent on the number of particles in the simulation cell; however, as the number of particles increases, the packing simulation is computationally expensive. Here, the computational time to generate microstructures with large number of particles is systematically evaluated first using optimization algorithms. The algorithms include the penalty function methods, best-in-class sequential quadratic programming method, matrix-less conjugate gradient method as well as the augmented Lagrangian method. Heuristic algorithms are next evaluated to achieve computationally efficient packing. The evaluated heuristic algorithms are mainly based on the drop-fall-shake (DFS) method, but modified to more effectively simulate the mixing process in commercial planetary mixers. With the developed procedures, representative volume elements (RVE) with volume fraction as high as 74% are demonstrated. The simulated microstructures are analyzed using our previously developed random network model to estimate the effective thermal and mechanical behavior given a particle arrangement.

Author(s):  
Sukshitha Achar P. L. ◽  
Huanyu Liao ◽  
Ganesh Subbarayan

Abstract As device power density increases, there is a need to dissipate generated heat by increasing particle volume loading in thermal interface materials. In this work, we develop and evaluate algorithms for generating ultrapacked microstructures of particles. Simulated microstructures reported in the literature rarely contain particle volume fractions greater than 60%. However, commercially available thermal greases claim to achieve volume fractions in the range of 60–80%. Therefore, to analyze effectiveness of commercially available particle-filled thermal interface materials, there is a need to develop algorithms capable of generating ultrapacked microstructures. The particle packing problem is initially posed as a nonlinear programming problem (NLP), and formal optimization algorithms are applied to generate microstructures that are maximally packed. Since accuracy of the simulated behavior is dependent on the number of particles in the simulation cell, efficiently simulating large number of particles is imperative. However, the packing simulation is computationally expensive. Therefore, various optimization algorithms are systematically evaluated to assess the computational efficiency as measured by the time to generate the microstructures for a system containing a large number of particles. The evaluated algorithms include the penalty function methods, best-in-class sequential programming method, matrix-less conjugate gradient method as well as the augmented Lagrangian method. In addition, heuristic algorithms are also evaluated to achieve computationally efficient packing. The evaluated heuristic algorithms are mainly based on the Drop-Fall-Shake method, but modified to more effectively simulate the mixing process in commercial planetary mixers. With the developed procedures, Representative Volume Elements (RVE) with volume fraction as high as 74% are demonstrated. The simulated microstructures are analyzed using our previously developed random network model to estimate the effective thermal and mechanical behavior given a particle arrangement.


Author(s):  
Ke Niu ◽  
Armin Abedini ◽  
Zengtao Chen

This paper investigates the influence of multiple inclusions on the Cauchy stress of a spherical particle-reinforced metal matrix composite (MMC) under uniaxial tensile loading condition. The approach of three-dimensional cubic multi-particle unit cell is used to investigate the 15 non-overlapping identical spherical particles which are randomly distributed in the unit cell. The coordinates of the center of each particle are calculated by using the Random Sequential Adsorption algorithm (RSA) to ensure its periodicity. The models with reinforcement volume fractions of 10%, 15%, 20% and 25% are evaluated by using the finite element method. The behaviour of Cauchy stress for each model is analyzed at a far-field strain of 5%. For each reinforcement volume fraction, four models with different particle spatial distributions are evaluated and averaged to achieve a more accurate result. At the same time, single-particle unit cell and analytical model were developed. The stress-strain curves of multi-particle unit cells are compared with single-particle unit cells and the tangent homogenization model coupled with the Mori-Tanaka method. Only little scatters were found between unit cells with the same particle volume fractions. Multi-particle unit cells predict higher response than single particle unit cells. As the volume fraction of reinforcements increases, the Cauchy stress of MMCs increases.


Author(s):  
Parisa Vaziee ◽  
Omid Abouali

Effectiveness of the microchannel heat sink cooled by nanofluids with various particle volume fractions is investigated numerically using the latest theoretical models for conductivity and viscosity of the nanofluids. Both laminar and turbulent flows are considered in this research. The model of conductivity used in this research accounts for the fundamental role of Brownian motion of the nanoparticles which is in good agreement with the experimental data. The changes in viscosity of the nanofluid due to temperature variation are considered also. Final results are compared with the experimental measurements for heat transfer coefficient and pressure drop in microchannel. Enhancement in heat transfer is achieved for laminar flow with increasing of volume fraction of Al2O3 nanoparticles. But for turbulent flow an enhancement of heat removal was not seen and using higher volume fractions of nanoparticles increases the maximum substrate temperature. Pressure drop is increased with using nanofluids because of the augmentation in the viscosity and this increase is more noticeable in higher Reynolds numbers.


2021 ◽  
Author(s):  
Ruifeng CAO ◽  
Taotao WANG ◽  
Yuxuan ZHANG ◽  
Hui WANG

Improved heat transfer in composites consisting of guar gel matrix and randomly distributed glass microspheres is extensively studied to predict the effective thermal conductivity of composites using the finite element method. In the study, the proper and probabilistic three-dimensional random distribution of microspheres in the continuous matrix is automatically generated by a simple and efficient random sequential adsorption algorithm which is developed by considering the correlation of three factors including particle size, number of particles, and particle volume fraction controlling the geometric configuration of random packing. Then the dependences of the effective thermal conductivity of composite materials on some important factors are investigated numerically, including the particle volume fraction, the particle spatial distribution, the number of particles, the nonuniformity of particle size, the particle dispersion morphology and the thermal conductivity contrast between particle and matrix. The related numerical results are compared with theoretical predictions and available experimental results to assess the validity of the numerical model. These results can provide good guidance for the design of advanced microsphere reinforced composite materials.


Author(s):  
W. M. Cho ◽  
Y. W. Kwon ◽  
C. T. Liu

This study investigated the effects of random and non-uniform particle distributions on the damage initiation and growth in particulate composites. Numerical specimens with either no crack or an existing crack were examined. For the cases with no crack, the effect of sizes of the representative area for non-uniform particle volume fractions was studied on the overall stress-strain curves and the results were compared with that of the specimen with uniform particle volume fractions. Other studies considered cracked specimens, either single edge crack or a center crack. The global-local approach was used along with multi-scale technique. The global analysis determined the deformations around the crack tip. Then, the local analysis evaluated the damage progress at the crack tip using the solution of the global analysis as boundary conditions. The results showed non-uniformed particle volume fractions in particulate composites caused the crack growth at lower applied loads than the uniform particle volume fraction. Statistical data were also plotted for the non-uniform particle volume fraction cases.


Fluids ◽  
2021 ◽  
Vol 6 (6) ◽  
pp. 217
Author(s):  
Liangquan Hu ◽  
Zhiqiang Dong ◽  
Cheng Peng ◽  
Lian-Ping Wang

The lattice Boltzmann method is employed to conduct direct numerical simulations of turbulent open channel flows with the presence of finite-size spherical sediment particles. The uniform particles have a diameter of approximately 18 wall units and a density of ρp=2.65ρf, where ρp and ρf are the particle and fluid densities, respectively. Three low particle volume fractions ϕ=0.11%, 0.22%, and 0.44% are used to investigate the particle-turbulence interactions. Simulation results indicate that particles are found to result in a more isotropic distribution of fluid turbulent kinetic energy (TKE) among different velocity components, and a more homogeneous distribution of the fluid TKE in the wall-normal direction. Particles tend to accumulate in the near-wall region due to the settling effect and they preferentially reside in low-speed streaks. The vertical particle volume fraction profiles are self-similar when normalized by the total particle volume fractions. Moreover, several typical transport modes of the sediment particles, such as resuspension, saltation, and rolling, are captured by tracking the trajectories of particles. Finally, the vertical profiles of particle concentration are shown to be consistent with a kinetic model.


Author(s):  
Diana Grandio ◽  
Drazen Fabris

In prior work an effective medium approach (EMA) has been developed to evaluate composite physical properties such as thermal conductivity, dielectric function or elastic modulus (C.-W. Nan, Prog. Mat. Sci. V. 37, 1993). This model combined with the Kapitza interface resistance can predict the effective thermal conductivity of randomly dispersed long fibers for a very low volume fraction (f < 0.01). The present study compares finite-element (FEA) computations and the EMA model for CNT-matrix compositions with low to moderate volume fractions, 0.001 to 0.02. The FEA results obtained show that the EMA model underestimates the effective thermal conductivity of the composite when the particles are very close to each other, even for small particle volume fractions. For aligned fibers the Kaptiza resistance cannot be neglected in the longitudinal direction. This paper proposes a general correction function for the dependence on particle to particle interaction based on the near neighbor distances and the number of near neighbors. This correction function reduces the EMA under prediction to within several percent (< 5%) in most cases.


Electronics ◽  
2018 ◽  
Vol 8 (1) ◽  
pp. 19 ◽  
Author(s):  
Alejandro García-Monzó ◽  
Héctor Migallón ◽  
Antonio Jimeno-Morenilla ◽  
José-Luis Sánchez-Romero ◽  
Héctor Rico ◽  
...  

A numerous group of optimization algorithms based on heuristic techniques have been proposed in recent years. Most of them are based on phenomena in nature and require the correct tuning of some parameters, which are specific to the algorithm. Heuristic algorithms allow problems to be solved more quickly than deterministic methods. The computational time required to obtain the optimum (or near optimum) value of a cost function is a critical aspect of scientific applications in countless fields of knowledge. Therefore, we proposed efficient algorithms parallel to Teaching-learning-based optimization algorithms. TLBO is efficient and free from specific parameters to be tuned. The parallel proposals were designed with two levels of parallelization, one for shared memory platforms and the other for distributed memory platforms, obtaining good parallel performance in both types of parallel architectures and on heterogeneous memory parallel platforms.


Author(s):  
X. Zhang ◽  
S. Kanuparthi ◽  
G. Subbarayan ◽  
B. Sammakia ◽  
S. Tonapi

Particle laden polymer composites are widely used as thermal interface materials in the electronics cooling industry. The projected small chip-sizes and high power applications in the near future demand higher values of effective thermal conductivity of the thermal interface materials (TIMs) used between the chip and the heat-spreader and the heat-spreader and heat-sink. However, over two decades of research has not yielded materials with significantly improved effective thermal conductivities. A critical need in developing these TIMs is apriori modeling using fundamental physical principles to predict the effect of particle volume fraction and arrangements on effective behavior. Such a model will enable one to optimize the structure and arrangement of the material. The existing analytical descriptions of thermal transport in particulate systems under predict (as compared to the experimentally observed values) the effective thermal conductivity since these models do not accurately account for the effect of inter-particle interactions, especially when particle volume fractions approach the percolation limits of approximately 60%. Most existing theories are observed to be accurate when filler material volume fractions are less than 30–35%. In this paper, we present a hierarchical, meshless, computational procedure for creating complex microstructures, explicitly analyzing their effective thermal behavior, and mathematically optimizing particle sizes and arrangements. A newly developed object-oriented symbolic, java language framework termed jNURBS implementing the developed procedure is used to generate and analyze representative random microstructures of the TIMs.


e-Polymers ◽  
2008 ◽  
Vol 8 (1) ◽  
Author(s):  
Wijittra Wichiansee ◽  
Anuvat Sirivat

AbstractElectromechanical properties of crosslinked polydimethylsiloxane (PDMS) and PDMS_PEDOT/PSS blends were investigated as potential actuator materials. Experiments were carried out under the oscillatory shear mode and with applied electric filed strength varying from 0 to 2 kV/mm. The storage modulus, G', of the crosslinked PDMS increases with increasing electric filed strength with a scaling behavior: ΔG' α Eα where the scaling exponent α is equal to 0.88. With PEDOT/PSS particles added at volume fractions of 5, 10, 15, and 20 vol%, the storage modulus G' of the polymer blends, are higher than that of the pristine crosslinked PDMS. The scaling exponent of the scaling behavior decreases from 0.88 to 0.10 as the volume fraction increases from 0 to 20 vol%. The storage modulus sensitivity,, attains maximum values of 8%, 9%, 10%, and 9%, at particle volume fractions of 5, 10, 15, and 20 vol%, respectively at the electric field strength of 2 kV/mm. Temporal response of the crosslinked PDMS system is irreversible, but the response of the PDMS_PEDOT/PSS system is evidently reversible


Sign in / Sign up

Export Citation Format

Share Document