Applied Thermal Engineeringashkanhh.mit.edu/sites/default/files/images/ATE1.pdf · 2016. 5. 23. · Shock and vibration protection of submerged jet impingement cooling systems: Theory

lable at ScienceDirect

Applied Thermal Engineering 73 (2014) 1076e1086

Contents lists avai

Applied Thermal Engineering

journal homepage: www.elsevier .com/locate/apthermeng

Shock and vibration protection of submerged jet impingement coolingsystems: Theory and experiment

Ashkan Haji Hosseinloo a, *, Siow Pin Tan b, Fook Fah Yap c, Kok Chuan Toh b

a Department of Mechanical Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139, USAb Temasek Laboratories, Nanyang Technological University, 637553, Singaporec School of Mechanical and Aerospace Engineering, Nanyang Technological University, 50 Nanyang Avenue, Singapore 639798, Singapore

h i g h l i g h t s

� A novel double-chamber jet impingement cooling system was designed and fabricated.� Comprehensive set of random vibration and shock tests are conducted.� The isolation system proved to protect the cooling system properly against mechanical failure.� Cooling system performance was not significantly affected by the input random vibration and shock.

a r t i c l e i n f o

Article history:Received 17 June 2014Accepted 26 August 2014Available online 9 September 2014

Keywords:Cooling systemJet impingementVibration isolationRandom vibrationShock

* Corresponding author. Tel.: þ1 617 308 9776.E-mail address: [email protected] (A. Haji Hosse

http://dx.doi.org/10.1016/j.applthermaleng.2014.08.051359-4311/© 2014 Elsevier Ltd. All rights reserved.

a b s t r a c t

In the recent years, advances in high power density electronics and computing systems have pushedtowards more advanced thermal management technologies and higher-capacity cooling systems. Amongdifferent types of cooling systems, jet impingement technology has gained attention and been widelyused in different industries for its adaptability, cooling uniformity, large heat capacity, and ease of itslocalization. However, these cooling systems may not function properly in dynamically harsh environ-ment inherent in many applications such as land, sea and air transportation. In this research article, anovel double-chamber jet impingement cooling system is fabricated and its performance is studied inharsh environment. Using the authors' previous studies, isolators with optimum properties are selectedto ruggedize the chassis containing the cooling chamber against shock and random vibration. Experi-ments are conducted on both hard-mounted and isolated chassis and the cooling performance of thesystem is assessed using the inlet, and impingement surface temperatures of the cooling chamber. Theexperimental results show the isolation system prevents any failure that otherwise would occur, and alsodoes not compromise the thermal performance of the system.

© 2014 Elsevier Ltd. All rights reserved.

1. Introduction

High-tech equipment operation usually brings about higherlevels of unwanted heat generation such as excessive heat gener-ated in high-power embedded computing systems. This conse-quently, has raised the need for more sophisticated and higher-capacity cooling systems. Among the different types, jet impinge-ment cooling systems have been widely used in many industrialsystems such as gas turbine cooling, and high-density electricalequipment coolingmainly for their adaptability, cooling uniformity,large heat capacity, and ease of cooling localization. According to

inloo).

9

application of the main equipment, the accompanying coolingsystem could be subjected to shock and vibration. Shock and vi-bration are inherent to harsh environments pertaining to sea, land,and air transportation, automotive industries, and etc. Severe shockand vibration usually result in a degradation of the system per-formance, thus the systemneeds to be ruggedized against the harshenvironment.

Actually the choice of the “submerged jet” (i.e. the jet is in atotally liquid-filled chamber, as opposed to a free jet which is intoan air filled chamber) impingement system is dictated by theintended application: an airborne system subjected to changes inorientation and high g-forces. In general, a free jet will have betterheat transfer enhancement when it impinges on a surface, but willbe deflected when the orientation changes or under g-forces,causing dramatic changes in its performance. The shear forces

mailto:[email protected]://crossmark.crossref.org/dialog/?doi=10.1016/j.applthermaleng.2014.08.059&domain=pdfwww.sciencedirect.com/science/journal/13594311http://www.elsevier.com/locate/apthermenghttp://dx.doi.org/10.1016/j.applthermaleng.2014.08.059http://dx.doi.org/10.1016/j.applthermaleng.2014.08.059http://dx.doi.org/10.1016/j.applthermaleng.2014.08.059

A. Haji Hosseinloo et al. / Applied Thermal Engineering 73 (2014) 1076e1086 1077

between the liquid and the submerged jet will however hold it inplacee hence making it more suitable for the intended application.Also, other heat enhancement mechanisms, e.g. evaporative spraycooling and immersion boiling techniques, are considered notsuitable for the same reason, even though they may be better froma heat transfer viewpoint.

In a jet impingement cooling system, heat is dissipated byimpinging high-velocity fluid jets onto the heating source. To thisend, the working fluid that is usually water, is pumped into achamber housing where it is pushed through tiny nozzles drilledinto a so-called nozzle plate and then it impinges onto an inter-mediate plate called impingement plate that is mounted back toback to the heating source, and hence the heat is dissipated. Theheated water is drained back to a heat rejection unit to cool downbefore it circulates back to the cooling chamber in a closed loop.Fig. 1 shows the exploded view of a double-chamber jet impinge-ment cooling chamber considered in this study and also the ther-mal diagram of the entire cooling system.

The double-chamber design was chosen for two reasons: 1) it isa space-saving design as it enables the sharing of the pressurizedfeeding chamber and 2) the diametrically opposed jets will enableany changes in the heat transfer performance due to the orientationeffect to be picked up.

Despite the far-reaching applications of the jet impingementcooling system in harsh environment, the investigations on effect ofshock and vibration on this cooling system is scarce. Among thesefew works, most of them have focused on the pulsating (synthetic)jets rather than vibration of the target heating source or vibration ofboth. Synthetic jets are generally formed by flow moving back andforth through a small opening. They are produced by periodicejection and suction of fluid from an orifice induced by movementof a diaphragm inside a cavity among other ways [1]. In general, theresearch studies show that synthetic jets enhance the heat transferas compared to conventional continuous jets [2e5]. The studiesalso show that the heat transfer coefficient is found to bemaximumat the resonance frequency of the cavity [6,7]. Kataoka et al. [8]explored and explained that the mechanism for the enhancementof stagnation point heat transfer using a pulsed jet lies in produc-tion of a turbulent surface renewal effect. Using electro-hydrodynamic (EHD) oscillatory gas flows is another use ofvibration concept to enhance the heat transfer in impingementsystems [9e11]. The main idea of these enhancement techniques isto have additional mixing of the fluid and destabilization of thethermal boundary layer using secondary flows, and thus to increasethe heat transfer coefficient.

Documented studies on effect of vibration of target heat sourceson thermal performance of the jet impingement cooling systemsare even scarcer. Wen [12] studied the flow characteristics of around jet impinging on constant-heat-flux flat surface undergoing

Fig. 1. (a) Exploded view of the assembly of the jet impingement chamber: 1. chamber houcooling system.

forced vibrations. His experimental results show that by increasingthe frequency and amplitude of the heating plate, the stagnationpoint and average Nusselt numbers increase. His results tend tosuggest that at higher vibration frequencies/amplitudes, the inter-face between the solid and the air becomes more-turbulent, whichresults in an increase in the specific interfacial area resulting inbetter gasesolid contact and better heat transfer.

It should be noted that the intention of this study is not to usethe vibration to enhance the thermal performance (otherwise thesystem could have been designed differently); but rather to preventthe vibration from having an undesirable effect on the systemperformance.

Although the studies show that pulsated jets or low-frequencyvibration of the heating source could enhance the thermal perfor-mance, no studies have investigated the effect of vibration on thewhole cooling system including the jets and the heating sources.This is a crucial study because when the cooling system is used inan application with inherent harsh dynamic environment, thewhole cooling system along the impingement surfaces (heatingsources) are subjected to shock and vibration. Moreover, theworking fluid used inmost of the above-mentioned literature is air;however, using liquid-based jet impingement cooling systems,particularly in electronics industries, is increasing more than ever,and thus their performance under harsh environment should beinvestigated as well. And last but not least, all the studies on effectof vibration on these cooling systems have considered only sinu-soidal excitations while the excitation sources in real applications,particularly in transportation and other harsh environments, aretotally different. In these applications, the excitation sources aremainly base shock and base random vibration that have never beenaddressed in the literature up to date. Apart from shock and vi-bration effect on the thermal performance of the cooling system,these harsh environment result in mechanical catastrophic or fa-tigue failure of the system components, and hence need to be maderugged. In addition, the ruggedization system should not compro-mise the cooling performance of the system if not enhancing it.

Shock and vibration isolation is probably the most commontechnique for dynamic protection of sensitive components. Intraditional isolation system design, the whole system that isintended to be protected, is taken as a rigid system and togetherwith the isolation system form a single-degree-of-freedom (sdof)model for which the isolation system properties are optimized.Although the traditional design may fulfill the shock or vibrationrequirements in some applications [13e19], it does not alwaysresult in an optimum design. This is because it ignores dynamics ofthe internal components that could be dynamically very sensitiveand responsive to the excitation. Therefore, to have better isola-tion systems more sophisticated models of the systems taking intoaccount their sensitive internal components, have been adopted

sing 2. nozzle plate 3. impingement plate (b) Thermal diagram of the jet impingement

Fig. 2. (a) CAD model of the ATR chassis with all the included components: 1. chassis 2. wedge locks 3. heaters 4. micro pumps 5. cooling chamber 6. isolators 7. inlet tubing 8. outlettubing. (b) Top view of the chassis: 1. heater thermocouples 2. outlet thermocouples 3. heaters 4. micro pumps 5. cooling chamber.

A. Haji Hosseinloo et al. / Applied Thermal Engineering 73 (2014) 1076e10861078

for isolation system design in different applications such as pro-tection of printed circuit boards [20,21], infrared equipment[22,23], and vibration isolation from hand-held percussion ma-chines [24e26].

According to the literature, there is no research article publishedup to date investigating the effects of shock and random vibrationon a liquid-based jet impingement cooling system. Furthermore,there is no theoretical or experimental study on design and analysisof shock and random vibration isolation system for these coolingsystems (except the authors' recent theoretical studies). In theirrecent studies [27e29], the authors developed discrete andcontinuous mathematical models of jet impingement cooling sys-tems and presented a generic design and optimization procedurefor a shock and vibration isolation system for them. Here in thiscurrent research, the authors experimentally investigate the effectsof shock and randomvibration on a double-chamber submerged jetimpingement cooling system. In addition, in view of their previousstudies, they select and experiment an isolation system that notonly protects the cooling system against any catastrophic or fatiguefailure but also guarantees no compromise in terms of thermalperformance of the system.

2. Mathematical modeling and theory

The cooling chamber is mounted in an air transport rack (ATR)chassis along with two micro pumps, heating sources and all thetubing (Fig. 2). The chassis itself is isolated from the base excitationsusing four elastomeric isolators. The excitations considered in thisstudy are Gaussian random vibration and terminal-peak sawtoothshock that may typically be encountered in an airborne equipmentoperating in a harsh environment.

2.1. Random base excitation

The power spectral density (PSD) profiles of random excitationare shown in Fig. 3. The randomvibration takes place in three steps:step 1 with root mean square (rms) acceleration of 8.8 grms, step 2with rms acceleration of 4.4 grms, and step 3 with rms accelerationof 31 grms. All the PSD profiles have an initial ramp from 20 Hz to100 Hz and a flat profile from 100 Hz to 2000 Hz. The chassis issubjected to the random excitation in all three X, Y, and Z directions.In-house experiments are conducted using Ling model A395shaker; however, this shaker is incapable of reaching step 3 of therandom vibration in horizontal directions due to the massive sliptable. For this reason a reduced profile is used for step 3 with rmsacceleration of 16 grms (Fig. 3).

2.2. Shock excitation

The shock base excitation is a terminal-peak sawtooth acceler-ation pulsewith duration of 11ms and amplitude of 20 g. The shockprofile is depicted in Fig. 3(b). This profile is used as a standardshock test [30] that could mimic a range of real-world shock phe-nomena such as those in military applications or transportation.The shock is applied to the chassis in both positive and negativedirections of each of the three axes.

2.3. Mathematical modeling

Themain purpose of the isolation system for this cooling systemis to protect the tubing, pumps and the impingement chamber inorder to prevent any loss in the cooling performance, and to protectit against any catastrophic or fatigue mechanical failure. Therandom vibration could make the nozzle and impingement platesvibrate that is unwanted. Comprehensive modeling and simulationwere done in the authors' previous studies to analyze and design anisolation system for jet impingement cooling systems [27e29]. Forthe case of the double-chamber cooling system, a continuousmodel (Fig. 4) was developed and used for isolation system design.Using this model, isolator properties (kc and cc) are chosen suchthat the acceleration at the center of the nozzle and impingementplates is minimized. In this model, acceleration PSD profile at thechassis (S€xc ðuÞ) could be calculated as.

S€xcðuÞ ¼ S€xb ðuÞ��

�u2c þ 2juuczc

��u2c � u2 þ 2juuczc

��2

; (1)

where, uc and zc are isolation system undamped natural frequencyand damping ratio, respectively, and S€xb ðuÞ is the input base ac-celeration PSD profile. This acceleration profile could then be usedas the boundary excitation of the nozzle and impingement plates tocalculate the mean square of the plate (nozzle or impingement)absolute acceleration at any given point ðx; yÞ on the plate as

Eh€w2a

�x; y; t

�i¼

Z∞

0

jHwaðx; y;uÞj2S€wbðuÞdu: (2)

In Eq. (2), E½:� shows the expected value, €wa denotes absoluteacceleration on a given point on the plate, Hwa ðx; y;uÞ representsfrequency response function of the plate from its boundary to anygiven point on the plate, and S €wb ðuÞ designates acceleration PSDprofile at the plate boundary that is the same as the acceleration

Fig. 3. (a) PSD profiles of the base random acceleration (b) terminal-peak sawtooth shock profile of the base excitation.


profile at the chassis level (S€xc ðuÞ). Readers are referred toRefs. [28,29] for details on mathematical modeling.

It was shown that there is an optimum isolation damping ratiothat varies from 0.13 to 0.21 for isolation frequency range of [20,50] Hz which maximizes the attenuation factor at the nozzle andimpingement plates when subjected to random vibration [29].Attenuation factor is defined as the ratio of the input rms acceler-ation to the rms acceleration at the target point. In the case of shockinput, peak acceleration is used to evaluate attenuation factor. Itwas also shown that there is an optimum damping which mini-mizes the change in impingement height (clearance between thenozzle and impingement plates); this value varies from 0.33 to 0.44for isolation frequency range of [20, 50] Hz. Furthermore, it wasshown that for the shock response a single-degree-of-freedommodel will be adequate enough to represent the dynamics of thesystem [27]. There is an optimum isolation damping ratio of 0.25for isolation frequencies below 40 Hz that minimizes the nozzleplate acceleration and deflection when the system is subjected tobase sawtooth shock excitation [27].

According to the findings mentioned above, an isolation systemwith isolation frequency of [20, 50] Hz with damping ratio of0.15e0.25 would be an appropriate isolator. There are a couple ofreasons why choosing an isolator system with an exact isolationfrequency and damping ratio is almost impossible; first of all, most

Fig. 4. Schematic of the isolated chassis with continuous plate models of the nozzleand impingement plates [29]: 1. base 2. isolation system 3. chassis with the pumps 4.chamber housing 5. impingement plate 6. nozzle plate.

of the highly-damped isolators are elastomeric isolators and elas-tomeric materials are usually not very linear. Secondly, the center-of-gravity (cg) of the cooling system does not lie in the plane of theisolators which consequently, brings about the rocking motion thatitself changes the translational isolation frequencies. The latterphenomenon is more severe in the lateral directions than the ver-tical one in this case study.

3. Experimental set-up and tests

Fig. 5 shows the experimental set-up for vibration and shocktests. The chassis shown in parts (a) and (b) of this figure is inhard-mounted, and isolated conditions, respectively. The chassisis subjected to shock and vibration with respect to the referenceaxes shown in Figs. 5(a) and 2(a), with and without isolators, andthe thermal performance of the system is assessed using thetemperatures measured by the thermocouples. The thermocou-ples are mounted in the inlet and outlet of the cooling chamberand also at nine points in the impingement plate. The shakershown in Fig. 5 is a Ling model A395 shaker with frequency rangeof 5e3000 Hz for structural and environmental vibration testing.The sine and random force ratings of this shaker are 6000 lb (y26.7 kN) and 5750 lb (y 25.6 kN), respectively. The slip tableshown in Fig. 5 weighs 147 lb (y 66.7 kg). The accelerometersused in the experiments to measure the vibration/shock accel-erations are from B&K Type 4393. The charge sensitivity has anerror of ±2%.

Fig. 2(b) depicts a view inside the chassis that contains thecooling chamber, the two heaters and the two pumps. The pumpsused are I-Drive IEG micro pumps with speed range of500 rpme5500 rpm. The heater plates are in fact two stainless steelplates with serpentine heating elements. Each of the heaters canprovide up to 900 W over an area of 100 mm � 150 mm. Graphitesheets and insulation plates are used to improve the heatspreading, and to reduce the heat loss, respectively.

The temperatures are measured at inlet and outlet tubes toand from the chamber, and also at nine points on each of theimpingement plates. The thermocouples used are 24 AWG T-type.To reduce the effect of heat conduction along the thermocouplewires on the temperature readings, shrink tubes are used to coverthe copper wires. These shrink tubes also prevent the formationof secondary junctions and strengthen the stiffness of the wires.The acquisition of the temperatures is conducted by an Agilent34972A LXI data acquisition unit through three Agilent 34901A20-channel multiplexers. The data acquisition process iscontrolled with HP BenchLink Data Logger3 software. The sam-pling time is set to 10 s. The thermocouples used have an accu-racy of ±0.5 �C.

Fig. 5. Vibration and shock tests set-up of the cooling system (a) view (a) with the reference coordinate system attached to the chassis cg: 1. shaker 2. slip table 3. controlaccelerometer 4. chassis including the cooling chamber and the two micro pumps 5. voltage transformers 6. DC power supplies 7. data acquisition unit for thermocouples 8. heatrejection unit (b) view (b): 1. isolator (one of the four) 2. control accelerometer 3. charge amplifier of the control accelerometer 4. shaker amplifier 5. shaker controller 6. computerinterface of the controller.


The isolators used are HT1-20 manufactured by LORD. Thisisolator is an elastomeric and relatively highly damped isolator;according to the manufacturer company, it has a rated dampingratio of 0.18.

4. Results and discussion

The shock and vibration tests are conducted on the hard-mounted chassis to check the effects of the excitations on themechanical failure and thermal performance of the system. Thenthe same tests are carried out with the isolated chassis to investi-gate the performance of the isolation system for prevention of anymechanical failure and also to check that the thermal performanceof the system is not compromised.

4.1. Isolation system performance

First a double sinesweep test is conducted to investigate theisolation properties of the isolators and check if they are in thedesired range. Fig. 6 shows the transmissibility curves of the chassisin three X, Y, and Z directions obtained from the sineweep tests.These curves are obtained by finding the ratio of acceleration at thechamber to the input acceleration (from the control accelerometer).The first and the largest peak corresponds to the isolation

101 1020

0.5

1

1.5

2

2.5

3

3.5

freque

Tran

smis

sibi

lity

rocking

Fig. 6. Transmissibility curves for the isolated chassis in three

frequency while the second one corresponds to the rocking mode.Rocking motion is the result of the center-of-gravity (cg) of thecooling system not lying in the plane of the isolators. All the otherpeaks are mostly related to the structural modes of the chassis.These modes are more easily excited in the lateral directions (out-of-plane direction for the chassis walls) than in the vertical direc-tion (in-plane direction for the chassis walls).

Embedded nonlinear leastesquare curve fitting tool in MAT-LAB is used to curve-fit the experimental curves with theoreticalsdof transmissibility function. The curve fitting results in isola-tion frequencies of 25.4 Hz, 30.4 Hz, and 51.8 Hz, and isolationdamping ratios of 0.18, 0.19, and 0.19 in the X, Y, and Z directions,respectively.

4.1.1. Random vibration testsSince the elastomeric materials are essentially nonlinear,

transmissibility curves, and hence isolation frequencies anddampings in random vibration tests are expected not to be exactlythe same as those obtained in sinesweep tests (Fig. 6). Fig. 7 showsthe transmissibility curves of the cooling chamber in the X-direc-tion for different steps of the input random excitations. As could beseen from the figure, the isolation frequencies are around 20 Hzthat is slightly smaller when compared to the sinesweep test resultshown in Fig. 6 (25.4 Hz).

103ncy(Hz)

X directionY directionZ direction

modes

structural modes

directions of X, Y, and Z obtained from sinesweep tests.

20 100 1000 20000

0.5

1

1.5

2

2.5

3

3.5

frequency(Hz)

Tran

smis

sibi

lity

step 1step 2reduced step 3

rocking mode

Fig. 7. Transmissibility curves of the isolated chassis in the X-direction for different random input excitations.


Figs. 8 and 9 depict the transmissibility curves of the coolingchamber in the Y and Z directions, respectively. According toFigs. 7e9, the isolation frequencies when the system is subjectedto random excitations are lower than those obtained from thesinesweep tests plotted in Fig. 6. Furthermore, as the excitationlevel increases i.e. at higher grms levels, the isolators becomesofter that subsequently results in lower isolation frequencies. Itcould also be seen from the transmissibility curves that thelateral isolation frequencies are lower than that of the verticaldirection. There are two reasons; first, the lateral to axial stiffnessratio of the isolators is smaller than one, resulting in a softerstiffness in the lateral X and Y directions. Secondly, there is acoupled rocking mode in the lateral directions that affects thetranslational mode and decreases the corresponding lateral nat-ural frequencies. The rocking mode is almost negligible in thevertical (Z) direction as the isolators are mounted almost sym-metrically with respect to the projected cg position in the XeYplane.

According to Figs. 7e9, the isolation frequencies and dampingratios vary in the X direction from 21.6 Hz to 18.7 Hz, and from0.18 to 0.25, in the Y direction from 23.7 Hz to 19.2 Hz, and from0.19 to 0.23, and in the Z direction they vary from 35.1 Hz to31.4 Hz and from 0.19 to 0.18, respectively. These damping values

20 1000

0.5

1

1.5

2

2.5

3

frequen

Tran

smis

sibi

lity

rocking mode

Fig. 8. Transmissibility curves of the isolated chassis in th

are quite close to the optimum damping found in Ref. [29] tominimize the vibration of the nozzle plate, impingement plateand the chassis. The optimum damping ratio found in Ref. [29] tominimize the plates acceleration varies from 0.13 to 0.19 forisolation frequency of 20 Hze35 Hz. Having the exact optimumdamping for a given isolation frequency using off-the-shelfelastomeric isolators is almost impossible and usually notnecessary. For instance, according to the results in Ref. [29], achange of damping ratio from 0.13 to 0.18 for isolation frequencyof 20 Hz reduces the attenuation factor at the nozzle andimpingement plates by only 4%.

Table 1 tabulates the attenuation factors at the cooling chamberin different directions and excitation steps. It is impossible tomount an accelerometer inside the chamber to measure the ac-celeration at the nozzle and impingement plates, thus the attenu-ation factor is reported only at the chamber. Consequently, the sdofmodel with parameters obtained from curve fitting of the experi-mental data, is used to calculate the theoretical attenuation factors.According to Table 1, there is a good agreement between theexperimental and theoretical attenuation factors. It could also beseen that the isolation system attenuates the input random exci-tations significantly; the lowest attenuation factor in the table is 5.1and the highest is 16.1.

1,000 2000cy(Hz)

step 1step 2reduced step 3

e Y-direction for different random input excitations.

20 100 1,000 2,0000

0.5

1

1.5

2

2.5

3

3.5

frequency(Hz)

Tran

smis

sibi

lity

step 1step 2reduced step 3step 3

rocking mode

Fig. 9. Transmissibility curves of the isolated chassis in the Z-direction for different random input excitations.


In addition to visual inspection, after each step of the vibrationtest, a sinesweep test is conducted to check for any abnormality inthe transmissibility curves that could be indicative of any failure inthe isolators. There was nomechanical failure either in the isolatorsor any part of the cooling system.

4.1.2. Shock testsThe chassis is subjected to the terminal-peak sawtooth base

shock with maximum acceleration of 20 g in all three X, Y, and Zdirections (both positive and negative directions). Since the shockduration is very small (only 11 ms), the mechanical response of thesystem and intactness of the cooling system and isolators are moreimportant than the short transient effect on the thermalperformance.

Figs. 10e12 illustrate acceleration and displacement time his-tories of the fixture (input), the cooling chamber and the approxi-mate isolator deflection when subjected to shock in positive andnegative X, Y, and Z directions. Each shock test is repeated threetimes and then the average response acceleration is recorded. Thedifference between the displacement of the accelerometer moun-ted on the chassis and that of the one mounted on the fixture (thecontrol accelerometer) is taken as an approximation for the iso-lators' deflection. This assumption is correct if the chassis is rigidand if it experiences only a translational motion. However, neitherthe chassis is completely rigid nor it goes through only a trans-lational motion. The chassis flexibility and rocking motion will in-crease the displacement of the accelerometer mounted on it, andhence will increase the isolators' approximate deflection. There-fore, the approximation of the isolator deflection should be slightlylarger than its actual value.

Table 2 summarizes the experimental and theoretical (sdofmodel) attenuation factors and approximated isolator deflectionin different directions of the shock tests. Here the attenuationfactor is defined as ratio of the maximum input acceleration i.e.20 g to the maximum acceleration experienced by the chassis.

Table 1Experimental and theoretical attenuation factors at the cooling chamber in different dir

X Y

Step 1 Step 2 Reduced step 3 Step 1

Attenuation factor (experimental) 14.8 9.7 16.1 15.3Attenuation factor (theoretical) 19.5 11.6 17.4 17.7

According to the table, there is a relatively good agreement be-tween the theoretical and experimental results except in the Zdirection. It is seen that shock input is attenuated in the lateraldirections but amplified in the Z direction. It is slightly amplified inthe þZ direction but it is amplified four times in the -Z direction.Although the amplification in -Z direction looks quite large, theduration of the amplification spike is so small that in turn resultsin a low shock energy. One explanation for excessive shock in the Zdirection could be that the heavy chassis causes a pre-deflection inthe isolators which could make the chassis to bottom out moreeasily in this direction.

According to Table 2, approximations for the isolators' deflectionis below 9 mm in all directions. Deflection approximations arelarger in the Y direction. The reason is that due to mounting pointlimitations, the accelerometer is mounted on the top of the coolingchamber for the test in this direction (larger distance from thesystem cg), and hence is affected more by the rocking motion thatconsequently results in larger displacement measured by thisaccelerometer. When excited in the X direction the accelerometer ismounted at the same height as the cg of the system so as tominimize the rocking motion effect on calculation of the isolators'deflection.

Just like the random vibration tests, visual inspection andsinesweep tests are conducted after each shock test to check for anymechanical failure. There was no mechanical failure either in theisolators or any part of the cooling system.

4.2. Thermal system performance

Thermal performance of the cooling system is assessed in termsof the temperature difference between the average impingementsurface temperature and the inlet water temperature. This tem-perature difference at different dynamic tests is measured andcompared to that of the static test to check for the effect of theshock and vibration.

ections and excitation steps.

Z

Step 2 Reduced step 3 Step 1 Step 2 Reduced step 3 Step 3

8.5 15.9 8.3 5.1 6.9 10.89.4 16.8 9.5 5.6 7.3 10.9

-0.04 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 0.04 0.05 0.06-4

-3

-2

-1

0

1

2

3

4

5

time(s)

disp

lace

men

t(mm

)

-X (displacement)

control(fixture)chamberisolator deflection

-40 -30 -20 -10 0 10 20 30 40 50 60-25

-20

-15

-10

-5

0

5

10

time(ms)

acce

lera

tion(

g)

-X(acceleration)

demandcontrol(fixture)chamber

-0.04 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 0.04 0.05 0.06-6

-5

-4

-3

-2

-1

0

1

2

3

time(s)

disp

lace

men

t(mm

)

+X (displacement)


-40 -30 -20 -10 0 10 20 30 40 50 60-10

-5

0

5

10

15

20

25

time(ms)

acce

lera

tion(

g)

+X (acceleration)


Fig. 10. Acceleration and displacement time histories of the fixture, and the chamber, and the time history of the approximate isolator deflection when subjected to shock in positiveand negative X directions. The demand legend in the upper panels refers to the ideal sawtooth shock profile that the control acceleration should follow.

-0.04 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 0.04 0.05 0.06-4

-2

0

2

4

6

8

10

time(s)

disp

lace

men

t(mm

)

-Y (displacement)


-40 -30 -20 -10 0 10 20 30 40 50 60-25

-20

-15

-10

-5

0

5

10

time(ms)

acce

lera

tion(

g)

-Y (acceleration)


-0.04 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 0.04 0.05 0.06-10

-8

-6

-4

-2

0

2

4

time(s)

disp

lace

men

t(mm

)

+Y (displacement)


-40 -30 -20 -10 0 10 20 30 40 50 60-10

-5

0

5

10

15

20

25

time(ms)

acce

lera

tion(

g)

+Y (acceleration)


Fig. 11. Acceleration and displacement time histories of the fixture, and the chamber and the time history of the approximate isolator deflection when subjected to shock in positiveand negative Y directions. The demand legend in the upper panels refers to the ideal sawtooth shock profile that the control acceleration should follow.


-0.04 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 0.04 0.05 0.06-4

-2

0

2

4

6

8

time(s)

disp

lace

men

t(mm

)

-Z (displacement)


-40 -30 -20 -10 0 10 20 30 40 50 60-100

-80

-60

-40

-20

0

20

time(ms)

acce

lera

tion(

g)

-Z (acceleration)


-40 -30 -20 -10 0 10 20 30 40 50 60-30

-20

-10

0

10

20

30

time(ms)

acce

lera

tion(

g)

+Z (acceleration)


-0.04 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 0.04 0.05 0.06-8

-6

-4

-2

0

2

4

6

time(s)

disp

lace

men

t(mm

)

+Z (displacement)


Fig. 12. Acceleration and displacement time histories of the fixture, and the chamber and the time history of approximate isolator deflection when subjected to shock in positiveand negative Z directions. The demand legend in the upper panels refers to the ideal sawtooth shock profile that the control acceleration should follow.


4.2.1. Random vibration testsThe chassis is subjected to randomvibrations in all three X, Yand

Z directions in both hard-mounted and isolated conditions. Whenthe chassis is hard-mounted only steps 1 and 2 are applied to thesystem. This is because higher excitation levels in the hard-mounted condition resulted in mechanical failures in the systemsuch as water leakage from the cooling chamber and tubing, sealingfailure of the pumps, and tube breakage. In both cases, a static test(i.e. no shock and vibration input) is conducted for thermal per-formance reference. Excitation in each step is maintained for about15 min so that the thermal system reaches steady state.

Fig. 13 shows the difference between the average impingementsurface temperature and the inlet water temperature for hard-mounted and isolated conditions when the chassis is subjected tosteps 1 and 2 of the random vibration in X, Y and Z directions. The

Table 2Experimental and theoretical attenuation factor and approximate isolator deflectionin shock tests for the three directions.

X Y Z

Positive Negative Positive Negative Positive Negative

Attenuation factor(experimental)

2.9 3.4 1.9 1.8 0.78 0.24

Attenuation factor(theoretical)

2 2 1.9 1.9 1.3 1.3

Approximate isolatordeflection (mm)(experimental)

5.6 4.8 9.0 8.9 6.8 7.4

Approximate isolatordeflection (mm)(theoretical)

6.2 6.2 6.2 6.2 3.7 3.7

figure shows that there is not much difference in terms of tem-perature difference between the isolated and hard-mounted con-ditions and also between the static and dynamic cases.

Figs. 14e16 show the temperature difference for static and allrandom vibration steps in X, Y and Z directions, respectively. Asmentioned earlier, the shaker shown in Fig. 5 is incapable ofgenerating sufficient force in the lateral directions to reach step 3 ofthe randomvibration. To subject the system to step 3 of the randomvibration in the lateral directions i.e. X and Y directions, out-houseexperiments are conducted with a larger shaker. This is the reasonwhy there are two static test references in Figs. 14 and 15; onecorresponds to the static test conducted before the in-houserandom vibration tests (step 1 to reduced step 3) and the othercorresponds to the static test before the step 3 of the random vi-bration in the out-house experiments. Some conditions such as theambient temperature are different for the in-house and out-houseexperiments which explains the difference in the correspondingstatic tests' thermal performances. According to Figs. 14e16, thereis not much difference between the temperature difference in thestatic tests and that of the randomly excited chassis. This simplymeans that the thermal performance of the isolated chassis is notaffected much by the random vibration input (step 1 to step 3).

4.2.2. Shock testsSince the shock tests are so quick (with a duration of only 11ms),

the main purpose of the isolation system is protecting the me-chanical components of the system rather than maintaining thethermal performance of the system in that short duration. A fewrounds of the shock test on the hard-mounted chassis resulted inmechanical failure of the system like water leakage, and tubebending and misalignments. However, the isolation system well

Fig. 13. Comparison of temperature difference between the average impingementsurface temperature and the inlet water temperature for hard-mounted and isolatedchassis when subjected to the first and second steps of the random base vibration.

random vibration step

tem

pera

ture

diff

eren

ce (

o C)

0

2

4

6

8

10

12

14

16

18

20

for s

tep1

to re

duce

d st

ep 3

for s

tep

3

static step 1 step 2 reduced step 3 step 3

Fig. 15. Temperature difference between the average impingement surface tempera-ture and the inlet water temperature for the static and four random vibration stepswhen the chassis is excited in the Y direction.


protected the cooling system with no failure detected. The shockhas some effects on the thermal performance of the system whichis transient and dies out soon after the shock ends.

5. Conclusion

This paper reports extensive experimental results on isolationsystem design and protection of a double-chamber submerged jetimpingement cooling system against shock and random vibration.


tem

pera

ture

diff

eren

ce (

o C)

0

2

4

6

8

10

12

14

16

18

20


for s

tep

3

for s

tep1

to re

duce

d st

ep 3

Fig. 14. Temperature difference between the average impingement surface tempera-ture and the inlet water temperature for the static and four random vibration stepswhen the chassis is excited in the X direction.

A continuous model adopted from the authors' previous study [29]has been used for isolators selection. The criterion for isolator se-lection is minimizing the nozzle and impingement plates' vibrationsubject to a maximum allowed rattle space for the whole chassis.

First, the cooling system is subjected to different steps of randomvibration with overall input vibration level ranging from about4 grms to 31 grms in both hard-mounted and isolated conditions.Transmissibility curves obtained from sinesweep and random vi-bration tests prove that the isolation frequency and damping are inthe desired range. In both hard-mounted and isolated cases, thethermal performance of the system is assessed by measuring thedifference between the average impingement surface temperatureand the inlet water temperature. The results show that there is nosignificant difference between the thermal performance of the hard-mounted and isolated chassis in steps 1 and 2 of the random exci-tation. However, vibration levels higher than that of step 1 (8.8 grms)resulted in mechanical, and thus thermal failure of the systemwhenhard-mounted, like water leakage, tubing breakage, and pumpsealing damage. The results also show that the isolation systemwellprotected the cooling system in all steps i.e. steps 1 to 3 without anymechanical failure or significant change in the thermal performanceof the system. The isolation system provided an attenuation factor ofup to 16 at the cooling chamber.


tem

pera

ture

diff

eren

ce (

o C)

0

2

4

6

8

10

12

14

16

18


Fig. 16. Temperature difference between the average impingement surface tempera-ture and the inlet water temperature for the static and four random vibration stepswhen the chassis is excited in the Z direction.


After the random vibration tests, the chassis is subjected tosawtooth base shock. A few rounds of shock excitation resulted inmechanical failure of the hard-mounted system; however, theisolated chassis withstood all the shock tests without any failure inthe cooling system or in the isolators. The isolation systemprovidedattenuation factor of up to 3.4 in the shock tests.

In brief, in view of the authors' recent study [29], an isolationsystem was selected and shock and random vibration tests wereconducted in hard-mounted and isolated conditions. It was shownthat the isolation system well protected the cooling system fromanymechanical failure and without any compromise in the thermalperformance.

Acknowledgements

This research was financially supported by Defense ScienceOrganization of Singapore and Temasek Laboratories @ NTU; theauthors would like to gratefully acknowledge their financial andtechnical support [grant number TL/DSOCL/0083/02]. The authorswould also like to sincerely thank Loh Chiah Vern, Poon Weng Waiand Lee Weng Kee for their unsparing help in the experiments.

References

[1] A. Agrawal, G. Verma, Similarity analysis of planar and axisymmetric turbulentsynthetic jets, Int. J. Heat Mass Transfer 51 (25e26) (2008) 6194e6198.

[2] R. Mahalingam, A. Glezer, Design and thermal characteristics of a synthetic jetejector heat sink, J. Electron. Packag. 127 (2) (2005) 172e177.

[3] A. Pavlova, M. Amitay, Electronic cooling using synthetic jet impingement,J. Heat Transfer 128 (9) (2006) 897e907.

[4] J. Garg, et al., Meso scale pulsating jets for electronics cooling, J. Electron.Packag. 127 (4) (2005) 503e511.

[5] M. Chaudhari, B. Puranik, A. Agrawal, Heat transfer characteristics of syntheticjet impingement cooling, Int. J. Heat Mass Transfer 53 (5e6) (2010)1057e1069.

[6] M. Chaudhari, et al., Near-field measurements and cavity design for syntheticjet, in: 38th AIAA Fluid Dynamics Conference and Exhibit, June 23, 2008-June26, 2008, American Institute of Aeronautics and Astronautics Inc, Seattle, WA,United States, 2008.

[7] M. Chaudhari, et al., Frequency response of a synthetic jet cavity, Exp. Therm.Fluid Sci. 33 (3) (2009) 439e448.

[8] K. Kataoka, et al., The effect of surface renewal due to large-scale eddies on jetimpingement heat transfer, Int. J. Heat Mass Transfer 30 (3) (1987) 559e567.

[9] F.C. Lai, P.J. McKinney, J.H. Davidson, Oscillatory electrohydrodynamic gasflows, J. Fluids Eng. 117 (3) (1995) 491e497.

[10] F.C. Lai, J. Mathew, Heat transfer enhancement by EHD-induced oscillatoryflows, J. Heat Transfer 128 (9) (2006) 861e869.

[11] M. Molki, et al., Heat Transfer enhancement of airflow in a channel usingcorona discharge, J. Enhanc. Heat Transf. 7 (6) (2000) 411e425.

[12] M.-Y. Wen, Flow structures and heat transfer of swirling jet impinging on aflat surface with micro-vibrations, Int. J. Heat Mass Transfer 48 (3e4) (2005)545e560.

[13] B.R. Allen, et al., Advanced isolation design for avionics on launch vehicles, in:Smart Structures and Materials 2000: Damping and Isolation, 6-8 March 2000,SPIE-Int. Soc. Opt. Eng, USA, 2000.

[14] M. Baril, et al., Passive vibration isolation for SITELLE's closed cycle cooledcryostats, in: Modern Technologies in Space- and Ground-Based Telescopesand Instrumentation II, July 1, 2012-July 6, 2012, SPIE, Amsterdam,Netherlands, 2012.

[15] H. Harmoko, et al., Design and analysis of shock and random vibrationisolation of operating hard disk drive in harsh environment, Shock Vib. 16(2009) 143e154 (Compendex).

[16] L.K. Liu, G.T. Zheng, Parameter analysis of PAF for whole-spacecraft vibrationisolation, Aerosp. Sci. Technol. 11 (6) (2007) 464e472.

[17] H.-U. Oh, K.-J. Lee, M.-S. Jo, A passive launch and on-orbit vibration isolationsystem for the spaceborne cryocooler, Aerosp. Sci. Technol. 28 (1) (July 2013)324e331.

[18] F.F. Yap, N. Vahdati, H. Harmoko, Design and analysis of vibration isolationsystems for hard disk drives, J. Magn. Magn. Mater. 303 (2006) 52e56(Copyright 2006, The Institution of Engineering and Technology).

[19] A.H. Hosseinloo, F.F. Yap, A new passive vibration isolator design for randombase excitations in zero and non-zero G-loading situations, in: 2011 DefenseScience Research Conference and Expo, DSR 2011, August 3, 2011-August 5,2011, IEEE Computer Society, Singapore, Singapore, 2011.

[20] A.M. Veprik, Vibration protection of critical components of electronic equip-ment in harsh environmental conditions, J. Sound. Vib. 259 (2003) 161e175(Copyright 2003, IEE).

[21] A.M. Veprik, V.I. Babitsky, Vibration protection of sensitive electronic equip-ment from harsh harmonic vibration, J. Sound. Vib. 238 (2000) 19e30(Copyright 2001, IEE).

[22] A. Veprik, S. Djerassy, V. Babitsky, Synthesis of snubber's spectral character-istics for vibration protection of sensitive components in airborne gyrostabi-lized electro-optic payloads, Proc. Inst. Mech. Eng. Part C J. Mech. Eng. Sci. 222(2008) 1885e1897 (Compendex).

[23] A.M. Veprik, et al., Vibration protection of sensitive components of infraredequipment in harsh environments, Shock Vib. 8 (2001) 55e69 (Compendex).

[24] E.V. Golycheva, V.I. Babitsky, A.M. Veprik, Dynamic correction of excitation inhand-held electro-pneumatic percussion machines, J. Sound. Vib. 259 (4)(2003) 829e843.

[25] E.V. Golysheva, V.I. Babitsky, A.M. Veprik, Vibration protection for an operatorof a hand-held percussion machine, J. Sound. Vib. 274 (2004) 351e367(Copyright 2004, IEE).

[26] I.J. Sokolov, V.I. Babitsky, N.A. Halliwell, Hand-held percussion machines withlow emission of hazardous vibration, J. Sound. Vib. 306 (1e2) (2007) 59e73.

[27] A.H. Hosseinloo, F.F. Yap, L.Y. Lim, Design and analysis of shock and randomvibration isolation system for a discrete model of submerged jet impingementcooling system, J. Vib. Control (2013), http://dx.doi.org/10.1177/1077546313490186.

[28] A.H. Hosseinloo, F.F. Yap, N. Vahdati, Analytical random vibration analysis ofboundary-excited thin rectangular plates, Int. J. Struct. Stab. Dyn. 13 (3)(2013).

[29] A.H. Hosseinloo, F.F. Yap, E.T. Chua, Random vibration protection of a double-chamber submerged jet impingement cooling system: a continuous model,Aerosp. Sci. Technol. 35 (2014) 29e38.

[30] MIL-STD-810G, Department of Defense Test Method Standard: EnvironmentalEngineering Considerations and Laboratory Tests, United States Departmentof Defense, 2008.

http://refhub.elsevier.com/S1359-4311(14)00755-8/sref1http://refhub.elsevier.com/S1359-4311(14)00755-8/sref1http://refhub.elsevier.com/S1359-4311(14)00755-8/sref1http://refhub.elsevier.com/S1359-4311(14)00755-8/sref1http://refhub.elsevier.com/S1359-4311(14)00755-8/sref2http://refhub.elsevier.com/S1359-4311(14)00755-8/sref2http://refhub.elsevier.com/S1359-4311(14)00755-8/sref2http://refhub.elsevier.com/S1359-4311(14)00755-8/sref3http://refhub.elsevier.com/S1359-4311(14)00755-8/sref3http://refhub.elsevier.com/S1359-4311(14)00755-8/sref3http://refhub.elsevier.com/S1359-4311(14)00755-8/sref4http://refhub.elsevier.com/S1359-4311(14)00755-8/sref4http://refhub.elsevier.com/S1359-4311(14)00755-8/sref4http://refhub.elsevier.com/S1359-4311(14)00755-8/sref5http://refhub.elsevier.com/S1359-4311(14)00755-8/sref5http://refhub.elsevier.com/S1359-4311(14)00755-8/sref5http://refhub.elsevier.com/S1359-4311(14)00755-8/sref5http://refhub.elsevier.com/S1359-4311(14)00755-8/sref5http://refhub.elsevier.com/S1359-4311(14)00755-8/sref6http://refhub.elsevier.com/S1359-4311(14)00755-8/sref6http://refhub.elsevier.com/S1359-4311(14)00755-8/sref6http://refhub.elsevier.com/S1359-4311(14)00755-8/sref6http://refhub.elsevier.com/S1359-4311(14)00755-8/sref7http://refhub.elsevier.com/S1359-4311(14)00755-8/sref7http://refhub.elsevier.com/S1359-4311(14)00755-8/sref7http://refhub.elsevier.com/S1359-4311(14)00755-8/sref8http://refhub.elsevier.com/S1359-4311(14)00755-8/sref8http://refhub.elsevier.com/S1359-4311(14)00755-8/sref8http://refhub.elsevier.com/S1359-4311(14)00755-8/sref9http://refhub.elsevier.com/S1359-4311(14)00755-8/sref9http://refhub.elsevier.com/S1359-4311(14)00755-8/sref9http://refhub.elsevier.com/S1359-4311(14)00755-8/sref10http://refhub.elsevier.com/S1359-4311(14)00755-8/sref10http://refhub.elsevier.com/S1359-4311(14)00755-8/sref10http://refhub.elsevier.com/S1359-4311(14)00755-8/sref11http://refhub.elsevier.com/S1359-4311(14)00755-8/sref11http://refhub.elsevier.com/S1359-4311(14)00755-8/sref11http://refhub.elsevier.com/S1359-4311(14)00755-8/sref12http://refhub.elsevier.com/S1359-4311(14)00755-8/sref12http://refhub.elsevier.com/S1359-4311(14)00755-8/sref12http://refhub.elsevier.com/S1359-4311(14)00755-8/sref12http://refhub.elsevier.com/S1359-4311(14)00755-8/sref12http://refhub.elsevier.com/S1359-4311(14)00755-8/sref13http://refhub.elsevier.com/S1359-4311(14)00755-8/sref13http://refhub.elsevier.com/S1359-4311(14)00755-8/sref13http://refhub.elsevier.com/S1359-4311(14)00755-8/sref14http://refhub.elsevier.com/S1359-4311(14)00755-8/sref14http://refhub.elsevier.com/S1359-4311(14)00755-8/sref14http://refhub.elsevier.com/S1359-4311(14)00755-8/sref14http://refhub.elsevier.com/S1359-4311(14)00755-8/sref15http://refhub.elsevier.com/S1359-4311(14)00755-8/sref15http://refhub.elsevier.com/S1359-4311(14)00755-8/sref15http://refhub.elsevier.com/S1359-4311(14)00755-8/sref15http://refhub.elsevier.com/S1359-4311(14)00755-8/sref16http://refhub.elsevier.com/S1359-4311(14)00755-8/sref16http://refhub.elsevier.com/S1359-4311(14)00755-8/sref16http://refhub.elsevier.com/S1359-4311(14)00755-8/sref17http://refhub.elsevier.com/S1359-4311(14)00755-8/sref17http://refhub.elsevier.com/S1359-4311(14)00755-8/sref17http://refhub.elsevier.com/S1359-4311(14)00755-8/sref17http://refhub.elsevier.com/S1359-4311(14)00755-8/sref18http://refhub.elsevier.com/S1359-4311(14)00755-8/sref18http://refhub.elsevier.com/S1359-4311(14)00755-8/sref18http://refhub.elsevier.com/S1359-4311(14)00755-8/sref18http://refhub.elsevier.com/S1359-4311(14)00755-8/sref19http://refhub.elsevier.com/S1359-4311(14)00755-8/sref19http://refhub.elsevier.com/S1359-4311(14)00755-8/sref19http://refhub.elsevier.com/S1359-4311(14)00755-8/sref19http://refhub.elsevier.com/S1359-4311(14)00755-8/sref20http://refhub.elsevier.com/S1359-4311(14)00755-8/sref20http://refhub.elsevier.com/S1359-4311(14)00755-8/sref20http://refhub.elsevier.com/S1359-4311(14)00755-8/sref20http://refhub.elsevier.com/S1359-4311(14)00755-8/sref21http://refhub.elsevier.com/S1359-4311(14)00755-8/sref21http://refhub.elsevier.com/S1359-4311(14)00755-8/sref21http://refhub.elsevier.com/S1359-4311(14)00755-8/sref21http://refhub.elsevier.com/S1359-4311(14)00755-8/sref22http://refhub.elsevier.com/S1359-4311(14)00755-8/sref22http://refhub.elsevier.com/S1359-4311(14)00755-8/sref22http://refhub.elsevier.com/S1359-4311(14)00755-8/sref22http://refhub.elsevier.com/S1359-4311(14)00755-8/sref22http://refhub.elsevier.com/S1359-4311(14)00755-8/sref23http://refhub.elsevier.com/S1359-4311(14)00755-8/sref23http://refhub.elsevier.com/S1359-4311(14)00755-8/sref23http://refhub.elsevier.com/S1359-4311(14)00755-8/sref24http://refhub.elsevier.com/S1359-4311(14)00755-8/sref24http://refhub.elsevier.com/S1359-4311(14)00755-8/sref24http://refhub.elsevier.com/S1359-4311(14)00755-8/sref24http://refhub.elsevier.com/S1359-4311(14)00755-8/sref25http://refhub.elsevier.com/S1359-4311(14)00755-8/sref25http://refhub.elsevier.com/S1359-4311(14)00755-8/sref25http://refhub.elsevier.com/S1359-4311(14)00755-8/sref25http://refhub.elsevier.com/S1359-4311(14)00755-8/sref26http://refhub.elsevier.com/S1359-4311(14)00755-8/sref26http://refhub.elsevier.com/S1359-4311(14)00755-8/sref26http://refhub.elsevier.com/S1359-4311(14)00755-8/sref26http://dx.doi.org/10.1177/1077546313490186http://dx.doi.org/10.1177/1077546313490186http://refhub.elsevier.com/S1359-4311(14)00755-8/sref28http://refhub.elsevier.com/S1359-4311(14)00755-8/sref28http://refhub.elsevier.com/S1359-4311(14)00755-8/sref28http://refhub.elsevier.com/S1359-4311(14)00755-8/sref29http://refhub.elsevier.com/S1359-4311(14)00755-8/sref29http://refhub.elsevier.com/S1359-4311(14)00755-8/sref29http://refhub.elsevier.com/S1359-4311(14)00755-8/sref29http://refhub.elsevier.com/S1359-4311(14)00755-8/sref30http://refhub.elsevier.com/S1359-4311(14)00755-8/sref30http://refhub.elsevier.com/S1359-4311(14)00755-8/sref30

Shock and vibration protection of submerged jet impingement cooling systems: Theory and experiment1. Introduction2. Mathematical modeling and theory2.1. Random base excitation2.2. Shock excitation2.3. Mathematical modeling

3. Experimental set-up and tests4. Results and discussion4.1. Isolation system performance4.1.1. Random vibration tests4.1.2. Shock tests

4.2. Thermal system performance4.2.1. Random vibration tests4.2.2. Shock tests

5. ConclusionAcknowledgementsReferences

Documents

Applied Thermal Engineeringashkanhh.mit.edu/sites/default/files/images/ATE1.pdf · 2016. 5. 23. · Shock and vibration protection of submerged jet impingement cooling systems: Theory