文档下载 文档下载导航
设为首页 | 加入收藏
搜索 请输入内容:  
 导航当前位置: 文档下载 > 所有分类 > 自然科学 > 物理 > Two phase convective cooling for ultra-high power dissipation in microprocessors
侵权投诉

Two phase convective cooling for ultra-high power dissipation in microprocessors

超高压电

Engineering Equation Solver (EES) [6] to identify the impact of coolant selection on predicted performance. Results indicate that when application is limited to employing dielectric coolants, fluids with higher saturation pressure tend to display better heat transfer performance.

Thermodynamic Model

A simple thermodynamic model is applied to the system depicted in Fig. 1.

Fig. 2 Mass flux (blue curve, right axis) and outlet density

(black curve, left axis) as a function of pressure drop The model is comprised of algebraic equations for mass,

for R134a with inlet conditions saturated liquid at energy and momentum (excluding friction) conservation under

358.15 K, and full vaporization of coolant (outlet the assumptions of constant microgap cross sectional

condition saturated vapor) as determined by solving geometry, i.e., Eqs. (1)-(3), respectively:

Eq. (3).

G constant (1)

saturated vapor, is an implication of momentum conservation,

Q

(2) and is independent of microgap geometry. Including energy G(hout hin)

Hconservation, Eq. (2), determines microgap height for a given

11 heat input rate per unit depth. (3) G2 Pin Pout out in

The conservation equations are supplemented by equations of state, which are implemented implicitly in EES. The thermodynamic state at the inlet is specified by the a minimum height exists as a function of pressure drop, temperature and the vapor mass quality: Tin = 358.15 K and occurring close to the same pressure drop which yields the x = 0. The temperature is close to, but below, typical design maximum mass flux. The minimum gap height occurs at a limits for wall temperature for silicon based microprocessors. slightly different pressure drop than the maximum mass flux A saturated, as opposed to sub-cooled, liquid avoids having as a result of variation of enthalpy change from inlet to outlet any section of the microgap experience relatively ineffective as a function of pressure drop. We find that for the heat input single phase cooling. The outlet condition is saturated vapor, rate (required power dissipation) of interest, 1 kW, and for i.e., x = 1, which is the upper bound for two phase heat coolants with relatively low (~ 100 kPa) saturation pressures transfer (in real systems, critical heat flux limits quality below at the inlet temperature of 358.15 K, the minimum height is unity). With these conditions, specification of the mass flux, unexpectedly large. For water, for instance, the minimum gap G, determines the outlet pressure through Eq. (3), and Eq. (2) is 600 m. In other words, for a microgap smaller than determines the microgap height, H, required for a given heat 600 m, with water entering at 358.15 K (Pin =57.8 kPa) and

for a heat input rate of 1 kW, the coolant will exit as transfer rate per unit depth, Q'.

superheated vapor regardless of mass flow rate. A small gap is

Results & Discussion of Thermodynamic Analysis expected to be beneficial due to the inverse scaling of heat The pressure drop, P Ptransfer coefficients with gap height to some power exponent, in Pout, required to maintain a

given mass flux, G, with complete vaporization of the coolant, n, and, because heat transfer coefficients also scale directly specified by Eq. (3), is, surprisingly, not unique (e.g., Fig. 2). with the thermal conductivity of the coolant, we present the

n

This is a result of the fact that the density at the outlet is a results of our thermodynamic analysis in terms of k/H in Fig. function of the outlet pressure, which introduces non-linearity 3 (with n chosen to be unity for simplicity and lack of a better into the system of equations. Practically, one implication is choice without diverting to empiricism).

As seen in Fig. 3 the highest heat transfer coefficients that in numerical implementation of the model described by

Eqs. (1)-3), we specify the pressure drop to explicitly and should occur at the highest system pressures. Water and uniquely find the resulting non-negative mass flux, i.e., treat methanol have anomalously superior performance at low mass flux as a function of outlet pressure, and not vice versa. pressure due to combined effects of their larger latent heats For each coolant we investigated, we find that there is a and thermal conductivities as compared to the dielectric fluids. maximum attainable mass flux for a given inlet temperature However, they are not suitable for many electronics cooling and under the constraint that the coolant starts as saturated applications due to their appreciable electrical conductivities

(unless they are extremely pure or continuously de-ionized), liquid and exits as saturated vapor.

The result described above and depicted in Fig. 2, that a and, in the case of methanol, safety concerns. The trend of maximum mass flux is attainable for a given inlet temperature improved performance at higher system pressure is a direct when the coolant state progresses from saturated liquid to

Two phase convective cooling for ultra-high power dissipation in microprocessors

result of the small achievable gaps at higher pressure. The

Fig. 1 System definition for thermodynamic analysis.

上一页第2页

你可能喜欢

热门文档

相关文档

站点地图 | 文档上传 | 侵权投诉 | 手机版
本站所有资源均来自互联网,本站只负责收集和整理,均不承担任何法律责任,如有侵权等其它行为请联系我们.
文档下载 Copyright 2013 doc.guandang.net All Rights Reserved.  -  email