Experimental investigation of a novel heat pipe cold plate for electronics cooling

Transcription

1 Journal of Scientific & Industrial Research Vol. 68, October 2009, pp. MA & YAO: NOVEL HEAT PIPE COLD PLATE FOR ELECTRONICS COOLING 861 Experimental investigation of a novel heat pipe cold plate for electronics cooling Zhe-Shu Ma* and Shou-Guang Yao School of Naval Architecture and Ocean Engineering, Jiangsu University of Science and Technology, No. 2, Mengxi Road, Zhenjiang City, Jiangsu Province, , The People s Republic of China Received 22 July 2008; revised 11 June 2009; accepted 29 June 2009 In this study, to deal with multi-heat-source and high-heat-flux heat transfer problem in confined space within electronic equipments, cold plate equipment employing heat pipe technology has been designed. Acetone-aluminum heat pipe construction composed of 8 vertical pipes with their upper condensation section and lower evaporation section connected to let working liquid and vapor flow through each other when equipment works. Evaporation section of connective heat pipe construction is embedded in aluminum-made cold plate and cooling water flowing through water jacket cools condensation section of construction. Electronic-heating heat sources (16) are evenly arranged on two bigger vertical surfaces of cold plate to simulate heat generation of real array antennas and heat is eventually transferred to cooling water and then to outer environment. Heat pipe cold plate equipment has been formed and possesses excellent heat-transfer performance, startup performance and temperature evenness and can solve high heat-flux problem determining working reliability of array antennas effectively. Keywords: Electronics cooling, Heat pipe cold plate, Heat transfer, Startup performance Introduction Cooling technologies and developments related to high power electronic devices include a variety of processes ranging from phase-change cooling, forced convection cooling, natural convection cooling and micro heat exchanger cooling 1-3. Heat pipe electronic cooling is a kind of phase-change cooling technology possessing large equivalent thermal conductance, excellent packaging flexibility, passive operation and high reliability 4-9. Heat pipes including two-phase closed thermosyphons are promising phase-change electronic cooling technology for advantages such as high heat transfer efficiency, simple construction and reliable operational performance 2-3. Cooling of array antennas of ship borne radar equipment is a multi-heat-source and high heat-flux heat transfer problem in confined space. In current employed forced-convection water-cooling cold plate equipment, cold plate is main part of heat exchanger and its two big vertical surfaces are assembly surfaces of array antennas. Long-time application indicates that heattransfer performance of cold plate equipment is difficult *Author for correspondence mazheshu@sina.com to meet cooling requirement. On the other hand, uneven plastic deformation of cold plate surfaces caused by uneven temperature distribution further affects effective running of antennas and cooling effect. This study presents design of cold plate equipment employing acetone-aluminum heat pipe. Experimental Design of Heat Pipe Cold Plate Equipment Based on thermal environment, wherein array antennas of shipborne radar equipment works well, design constraints of heat pipe cold plate equipment are as follows: 1) total power of heat transfer of equipment must be more than 12 kw; 2) maximum heat flux are more than 24 W/cm 2 ; and 3) maximum temperature of equipment surface must be lower than 70 C; and 4) temperature difference on equipment surface must be lower than 10 C under different working conditions. In newly designed heat pipe cold plate (Fig. 1), acetone-aluminum heat pipe, which is composed of 8 vertical pipes with upper end and lower end connected to let working liquid flow through each other, is embedded in aluminum-made cold plate. Selection of acetone as working fluid in aluminum-based heat pipe cold plate is

2 862 J SCI IND RES VOL 68 OCTOBER 2009 based on its appropriate boiling point and other thermodynamic properties 10. Electronic-heating heat sources (8) are evenly arranged on each of two bigger vertical surfaces to simulate heat generation of array antennas and heat is eventually transferred to cooling water surrounding condensation segment and then to outer environment. Branches of heat pipe are connected with each other. Different temperature and pressure in branches caused by different antennas working combinations can be effectively reduced and quickly maintained to a new balanced temperature and pressure level and thus equipment can quickly work stably. Thus, temperature evenness performance of equipment under different heating conditions becomes better. For good weldability, water jacket has been used for cooling aluminum alloy plate. Inlet and outlet of cooling water are on opposite side of cooling water jacket and there is a hole on top of jacket to allow thermocouples measuring temperatures of condensation segment through. Fig. 1 Construction of heat pipe cold plate equipment Working Principle of Experimental System Pure aluminum blocks, which shape same as real array antennas, are employed to simulate heat generation of real array antennas. Each block is machined with one groove to embed electronic heating piece generating heat. Heating conditions of experiment are realized by electronic heating pieces employing different coil densities in different heat-generation areas. Pure aluminum blocks contact closely with cold plate surface and are stuck solidly with conductive silicon glue. Further, a pair of iron clamps is employed to ensure better fixation. Vertical surfaces (Fig. 2) of cold plate are marked positive and negative. Simulative antennas (8) on positive and negative surface are marked A, B, C, D, E, F, G, and H. Thermocouples employed to measure temperatures of high heat flux areas are marked T1 and thermocouples employed to measure temperatures of low heat flux areas are marked T2. Thus, measuring point under simulative antenna (A) and on high flux area of positive surface can be nominated positive (A1). At the same time, 8 thermocouples, marked a, b, c, d, e, f, g and h, are welded at middle of condensation segment on each branch heat pipes. Thus, there are 40 thermocouples in experimental system (Fig. 2). To realize dynamic data acquisition and graphic display, data acquisition and processing system is employed. Hardware system is set up based on a PCL- 812PG A/D card and three serial PCLD-789D cards Fig. 2 Distribution of thermocouples and software is developed based on Kingview6.01 desktop. Data acquisition and processing system can acquire dynamic data from each channel and display data varying curves. In addition, data acquired are stored as Access 2000 database to be convenient for analysis. Results and Discussion Startup and temperature evenness performance of heat pipe cold plate are tested under full heat load condition, under which all simulative antennas work, and partial heat load conditions, under which more than one

3 MA & YAO: NOVEL HEAT PIPE COLD PLATE FOR ELECTRONICS COOLING 863 Fig. 3 Startup performance of heat pipe cold plate (P=4000 antenna doesn t work. Startup performance includes startup time and how other factors affect startup time. In this paper, experiments under different cooling water flow rate with same heat generation power and different heat generation power with same cooling water flow rate are processed and reported. In all results, temperatures of thermocouples T1 and thermocouples T2 are average temperatures of 8 branches at same height. Fig. 4 Temperature evenness of heat pipe cold plate (P=4000 Startup and Temperature Evenness Performance of Heat Pipe Cold Plate under Full Load Condition Startup curves of heating power (P=4000 W) are drawn at cooling water flow rate of m=200 L/h (Fig. 3a), 300 L/h (Fig. 3b) and 400 L/h (Fig. 3c) alternatively. Temperatures of T1 are higher than that of T2. Temperatures of T1 increase more quickly than

4 864 J SCI IND RES VOL 68 OCTOBER 2009 Fig. 5 Startup performance of heat pipe cold plate (P=2000 W): a) m=200 L/h; b) m=300 L/h; c) m=400 L/h that of T2, as shown by slopes of start segment of curves. After 4 min, difference of temperatures of T1 and T2 becomes stable and temperatures are beyond 50 C, indicating that heat pipe cold plate equipment has already been startup. Startup performances under same heating power but different cooling water flow rate are similar to each other; working temperature of equipment is about 60 C and startup time becomes longer with increase of cooling water flow rate. Temperature difference of T1 and T2 on different simulative antenna under heating power (P=4000 W) are drawn at cooling water flow rate m=200 L/h (Fig. 4a), 300 L/h (Fig. 4b) and 400 L/h Fig. 6 Temperature evenness of heat pipe cold plate (P=2000 (Fig. 4c) alternatively. In these curves, distance direction is horizontal of cold plate, temperature values are temperatures of T1 and T2 on different simulative antennas. In Fig. 4a, difference of maximum (66.75 C) and minimum (60.44 C) surface temperature is 6.31 C (lower than 10 C). Surface temperature in middle is higher than that of two sides. Vapor pressure in middle branch is higher than that of other branches and thus acetone liquid in it is less than that of others. With increase of cooling water flow rate, working temperature lowers a little and difference between T1 and T2 becomes a little bigger (Fig. 4).

5 MA & YAO: NOVEL HEAT PIPE COLD PLATE FOR ELECTRONICS COOLING 865 Startup and Temperature Evenness Performance of Heat Pipe Cold Plate under Partial Load Conditions Startup curves of heating power (P=2000 W) (simulative antennas marked A, C, E and G on both surfaces work while other simulative antennas don t work) are drawn at cooling water flow rate m=200 L/h (Fig. 5a), 300 L/h (Fig. 5b) and 400 L/h (Fig. 5c) alternately. Comparing Fig. 5 with Fig. 3, startup performances and variation of working temperature of cold plate equipment under partial load conditions (with same heating power but different cooling water flow rate) are similar. Startup time becomes longer with increase of cooling water flow rate. Temperature difference of T1 and T2 on different simulative antenna under heating power P=2000 W (simulative antennas marked A, C, E and G on both surfaces work while other simulative antennas don t work) are drawn at cooling water flow rate m=200 L/h (Fig. 6a), 300 L/h (Fig. 6b) and 400 L/h (Fig. 6c) alternately. Because A, C, E and G antennas on both surfaces work while other simulative antennas don t work under this heating arrangement, temperature curves fluctuate. On the other hand, for heat pipe branches that are connected with each other, heat pipe branches still work though simulative antennas don t work. Pressure in heat pipe construction goes to a new balanced value and thus temperatures of T1 and T2 under simulative antenna marked B, D, F and H are still close to that of appropriate others. It indicates that heat pipe cold plate equipment evens surface temperature very well. With increase of cooling water flow rate, working temperature of cold plate decreases and difference between T1 and T2 becomes a little bigger and thus heat pipe cold plate equipment evens surface temperature worse. Several limits (capillary, boiling, entrainment, sonic, and viscous) can inherently constrain operation and heat transfer performance of heat pipes 10. But when effective design calculation is done in accordance with appropriate working environment, such limits can be overcome. In present study, no limit was observed. For possible higher heat flux cooling application of novel heat pipe cold plate, in which power output and heat addition flux will be higher, novel heat pipe cold plate may meet some limits. e: Comparing with numerical simulation of liquid-vapor two-phase flow in heat pipe cold plate 11-12, newly designed heat pipe cold plate can provide excellent cooling of array antennas. Conclusions Heat pipe cold plate equipment under full load conditions and partial load conditions shows that startup time becomes longer as cooling water flow rate increases and startup time becomes shorter as heat generation power increases. Under full load conditions and partial load conditions, heat pipe cold plate equipment evens surface temperature of cold plate well and meets design requirements. For electronic heating pieces can t resist required design value of partial heat flux value (24 W/cm 2 ), experiments under design heating power are not carried out. But if new heating system can afford design heat generation power and realize appropriate power distribution, heat transfer performances (startup and temperature evenness) of designed heat pipe cold plate will be even better. References 1 Sathe B, Review of recent developments in some practical aspects of air-cooled electronic packages, J Heat Transf, Trans ASME, 120 (1998) Sauciuc B, Design and testing of the super fiber heat pipes for electronics cooling applications, in Annu IEEE Semiconductor Ther Measure & Manage Symp (San Jose, USA) November Namba K, Kimura N & Niekawa J, Heat-pipes for electronic devices cooling and evaluation of their thermal performance, IEEE Trans on Components and Packaging Technol, 23 (2000) Groll, M, & Khandekar, S, Pulsating Heat Pipes: A challenge and still unsolved problem in heat pipe science, Arch Thermodynamics, 23 (2002) Groll M & Rosler S, Operation principles and performance of heat pipes and closed two-phase thermosyphons, J Non-Equilibrium Thermodynamics, 17 (1992) Groll M & Khandekar S, Insights into performance modes of closed loop pulsating heat pipes and some design hints, in Proc 18th Nat & 7th ISHMT-ASME Heat Mass Transfer Conf, 2006, G- 440, Yang H, Khandekar S & Groll M, Performance characteristics of pulsating heat pipes as integral thermal spreaders, Int J Therm Sci, 48 (2009) Khandekar S & Groll M, An insight into thermo-hydrodynamic coupling in closed loop pulsating heat pipes, Int J Therm Sci, 43 (2004) Charoensawan P, Khandekar S & Groll M, Closed loop pulsating heat pipes - Part A: parametric experimental investigations, Appl Therm Eng, 23 (2003) Faghri A, Heat Pipe Science and Technology (Taylor & Francis, Washington DC) Ma Z S, Yao S G & Ming X, Numerical simulation of vapor-liquid two-phase flow and heat transfer within a new type of heat pipes for spaced heat sources, Power Eng, 24 (2004) Yao S G, Ma Z S & Chen R B, Numerical experimental study and analysis of performance for heat pipe cooling flat-plate with a newly structure, in Proc Chinese Soc Elect Eng, 25 (2005)