Dynamic Control of Pressure Drop Oscillation in a Microchannel Cooling System

2020 ◽  
pp. 1-16 ◽  
Author(s):  
Qi Jin ◽  
John T. Wen ◽  
Shankar Narayanan
2014 ◽  
Vol 136 (2) ◽  
Author(s):  
Yubai Li ◽  
Dongzhi Guo ◽  
Shi-Chune Yao

This study goes beyond the common microchannel cooling system composed of uniform parallel straight microchannels and proposed a three-stage design approach for spatially thermal-aware microchannel cooling of 2D multicore processors. By applying effective strategies and arranging key design parameters, stronger cooling is provided under the high power core area, and less cooling is provided under the low power cache area to effectively save the precious pumping power, lower the hot spot temperature and lower temperature gradients on chip. Two microchannel cooling systems are specifically designed for a 2 core 150 W Intel Tulsa processor and an 8 core 260 W (doubled power) Intel Nehalem processor with single phase HFE7100 as coolant. For the Tulsa processor, a strategy named strip-and-zone is used. The final design leads to 30 kPa pressure drop and 0.094 W pumping power while maintains the hot spot temperature to be 75°C. For the Nehalem processor, a split flow microchannel system and a widen-inflow strategy are applied. A design is achieved to cost 15 kPa pressure drop and 0.0845 W pumping power while maintains the hot spot temperature to be 82.9°C. The design approach in this study provides the basic guide for the industrial applications of effective multicore processor cooling using microchannels.


Author(s):  
Daniel Lo¨rstad

The main parts of the annular combustor liner walls of the Siemens gas turbine SGT-800 are convectively cooled using rib turbulated cooling. Due to the serial system of cooling and combustion air there is a potential of further reduction of total combustor pressure drop by improvements of the cooling system. Apart from the rib cooling, also the cooling channel bypass entrance is related to a significant part of the total cooling system pressure drop. In this study, an investigation is performed for a rib cooled channel which is related to the considered combustor liner and where empirical correlations are available in order to evaluate the methodology used. The study includes an assessment of the Reynolds Averaged Navier-Stokes (RANS) and Large Eddy Simulation (LES) models available within commercial Computational Fluid Dynamics (CFD) codes and includes also an investigation of model size when using periodic boundaries for LES simulations. It is well known that a small geometrical distance in the direction of the periodic boundaries may have a strong effect on the flow field but is often neglected in practice in order to speed up LES calculations. Here the effect is assessed in order to show what size is required for accurate results, both for time averaged and transient results. In addition too small domains may be affected by spurious low frequencies originating from the periodic boundaries requiring additional simulation time for time converged statistics, but also the averages may be significantly affected. In addition the simulation period for time converged statistics is evaluated in order to show that larger model size in the periodic direction does not necessarily require longer practical simulation time, due to the fact that larger volumes may be used for the combined time and space averaging. The aim is to obtain practical guidelines for LES calculations for internal cooling flows. Then the study is extended step by step to investigate the importance due to high Reynolds number, variable fluid properties and large temperature gradients in order to cover the ranges and specifics required for SGT-800 engine conditions.


Author(s):  
Dong-Il Kim ◽  
Ki-So Bok ◽  
Han-Bae Lee

To seek the fan operating point on a cooling system with fans, it is very important to determine the system impedance curve and it has been usually examined with the fan tester based on ASHRAE standard and AMCA standard. This leads to a large investment in time and cost, because it could not be executed until the system is made actually. Therefore it is necessary to predict the system impedance curve through numerical analysis so that we could reduce the measurement time and effort. This paper presents how the system impedance curve (pressure drop curve) is computed by CFD in substitute for experiment. In reverse order to the experimental principle of the fan tester, pressure difference was adopted first as inlet and outlet boundary conditions of the system and then flow rate was calculated. After determining the system impedance curve, it was compared with experimental results. Also the computational domain of the system was investigated to minimize computational time.


Author(s):  
Gerardo L. Augusto ◽  
Alvin B. Culaba ◽  
Laurence A. Gan Lim

The design criteria of converter cooling system for a 2.5 MW permanent magnet direct-drive wind turbine generator were investigated. Two (2) distribution networks with pipe sizes of DN40 and DN50 were used as basis for fluid flow analysis. The theoretical system pressure drop and system volume flow rate of converter cooling system were calculated using the governing equations of mass conservation, pump performance curve and distribution network characteristics. The system of nonlinear equations was solved using multivariable Newton-Raphson method with the solution vector determined using LU decomposition method. Numerical results suggest that the DN50 pipe provides a pressure drop limit of less than 300 Pa/m in the converter cooling system better than the pressure drop obtained from a DN40 pipe. The system volume flow rate of DN50 pipe was found to be above the operating limit of heat exchanger requirement of 135.30 L/min which needs to dissipate heat with a minimum of 50 kW.


2013 ◽  
Vol 10 (1) ◽  
pp. 40-47 ◽  
Author(s):  
Aparna Aravelli ◽  
Singiresu S. Rao ◽  
Hari K. Adluru

Increased heat generation in semiconductor devices for demanding applications leads to the investigation of highly efficient cooling solutions. Effective options for thermal management include passing of cooling liquid through the microchannel heat sink and using highly conductive materials. In the author's previous work, experimental and computational analyses were performed on LTCC substrates using embedded silver vias and silver columns forming microchannels. This novel technique of embedding silver vias along with forced convection using a coolant resulted in higher heat transfer rates. The present work investigates the design optimization of this cooling system (microheat exchanger) using systems optimization theory. A new multiobjective optimization problem was formulated for the heat transfer in the LTCC model using the log mean temperature difference (LMTD) method of heat exchangers. The goal is to maximize the total heat transferred and to minimize the coolant pumping power. Structural and thermal design variables are considered to meet the manufacturability and energy requirements. Pressure loss and volume of the silver metal are used as constraints. A hybrid optimization technique using sequential quadratic programming (SQP) and branch and bound method of integer programming has been developed to solve the microheat exchanger problem. The optimal design is presented and sensitivity analysis results are discussed.


Author(s):  
Jacques du Plessis ◽  
Michael Owen

Abstract As direct dry-cooling systems are becoming more popular for thermal power plants, there is a demand to increase the flexibility of the application and performance of these cooling systems. A novel hybrid (dry/wet) dephlegmator (HDWD) cooling system is being developed, and at this stage in the development of the HDWD, the performance analysis and optimization of the HDWD is currently subject to uncertainties in a number of parameters. One of the parameters is the confidence in the correlations to predict the steam-side pressure drop over the wide range of full to partial condensation conditions expected in the system as a result of the design. This study makes use of an experimental apparatus to measure steam pressure drop over a range of partial to full condensation inside a circular horizontal tube. The experiment is conducted by measuring the steam flow and steam pressure drop in a horizontal primary condenser tube with the presence of a secondary condenser tube. The primary condenser has a tube length of 2.5 m and an inside tube diameter of 19.3 mm similar to the proposed HDWD design. Existing correlations for pressure drop in condensing flow are compared with the results to assess the applicability of the correlations for the HDWD case. It was found that the correlation of Lockhart and Martinelli’s with the Chisholm parameter fits the experimental data the best with a mean error of ±15.6%. A parametric study also indicated that there is a prominent increase in the frictional pressure drop at low partial condensation ratios (i.e., high steam through flow) as expected with wave drag at the vapor and condensate interface due to the difference in velocity.


Author(s):  
Tom Saenen ◽  
Martine Baelmans

A one dimensional dynamic system model is developed to accurately simulate a two-phase microchannel electronics cooling loop. This model is based on the single component mixture equations for mass, momentum and energy. These equations are solved numerically using a finite volume method in conjunction with the SIMPLE algorithm. To calculate the pressure losses and heat transfer state of the art empirical correlations are used. Furthermore size effects of a typical microchannel cooling system are investigated with the new model. Special attention is given to the accumulator size and its limitations for portable applications. A simple model to investigate the accumulator size effect on the loop is developed and compared to numerical results obtained from the system model. The influence of various loop parameters and possible improvements are also investigated. Finally the effect of using different coolants is studied.


Sign in / Sign up

Export Citation Format

Share Document