Arabian Journal for Science and Engineering https://doi.org/10.1007/s13369-018-3395-8

RESEARCH ARTICLE - MECHANICAL ENGINEERING

Heat Transfer Performance of a Synthetic Jet at Various Driving Frequencies and Diaphragm Amplitude S. M. Firdaus1,2

· M. Z. Abdullah1 · M. K. Abdullah3 · Z. M. Fairuz1

Received: 18 January 2018 / Accepted: 29 May 2018 © King Fahd University of Petroleum & Minerals 2018

Abstract The miniaturization of electronic devices with high-speed processing components is aggravating the heat generation of devices/systems. Space constraint has become a major issue in electronic cooling as these system can no longer accommodate a fan and liquid piping. Synthetic jets are an alternative solution because of their low operating cost and low space requirement. In this work, we fabricated a synthetic jet and analyzed its amplitude motion at different frequencies to measure the enhancement of heat transfer. ANSYS FLUENT® 15 was used to identify the vortex formation related to the fluid velocity profile during the ejection and suction phases to substantiate heat transfer performance. The amplitude was determined by conducting laser Doppler experiments for each frequency applied. The experimental results were validated against numerical prediction using an appropriate turbulent model and a structured meshing grade. The conformity between the numerical and experimental results was found to be < 5%. The maximum velocity was observed at 500 Hz driving frequency, which agreed with the result that the resonance frequency at 500 Hz had the highest amplitude and sweep volume. A large vortex formation was also recorded during the ejection phase at 500 Hz, resulting in the maximum temperature drop and a higher heat transfer coefficient (h) than the nonresonance frequency. The synthetic jet operating at the resonance frequency produced the maximum amplitude, fluid velocity, and large vortex formation proportional to h. Keywords Synthetic jet · Driving frequency · Sinusoidal · Heat transfer · Amplitude

1 Introduction Effective cooling solutions are critical in current electronic devices for preventing overheating and extending the life of electronic components or semiconductor-based electronics. In electronic devices, the processing power increases with the reduction of the device dimension and space availability. The processing power is proportional to the heat generated in electronic devices with multi-processors [1–4]. Randy et al. [5] stated that existing conventional fans and natural flow (passive cooling) have reached their limit because of the min-

B

S. M. Firdaus [email protected]

1

School of Mechanical Engineering, Universiti Sains Malaysia, Engineering Campus, 14300 Nibong Tebal, Penang, Malaysia

2

Faculty of Mechanical Engineering, Universiti Teknologi MARA, Penang Campus, 13500 Permatang Pauh, Penang, Malaysia

3

School of Materials and Mineral Resources Engineering, Universiti Sains Malaysia, Engineering Campus, 14300 Nibong Tebal, Penang, Malaysia

imum space available inside the device assembled package or system. Synthetic jets are promising future cooling systems with their zero-net mass input capabilities. An induced fluid flow due to an upward diaphragm movement has entrained fluid motion into the cavity. Then, as the diaphragm moves downward, the fluid motion is expelled through one or more opening slots, which are called nozzles [6]. When applied to a surface, the air flow produces unique effects, such as fluid momentum, recirculation, and vortex formation. Figure 1 shows the synthetic jet working sequence for a complete cycle. During the cycle, the diaphragm imparts a high-speed flow, which creates a pair of vortex in the surrounding fluid. The formed vortices periodically have a momentum that quickly sweeps away the heat on the target surface. Smith and Glezer [7] presented the jet formation and evolution when a rectangular slot is used. The velocity profile, which was measured using a hot wire anemometry, described the flow feature. Comparison of the synthetic and continuous jets revealed that the synthetic jet lost momentum more rapidly.

123

Arabian Journal for Science and Engineering

Fig. 1 Synthetic jet working principle; a suction phase, b ejection phase [8]

faster rate than the vorticity diffused by the viscosity. The heat transfer coefficient (h) is a key parameter in electronic cooling. However, no study has attempted to determine the resonance driving frequency with a diaphragm amplitude for fabricated synthetic jets. Thus, in this paper, the diaphragm amplitude was measured at various driving frequencies, and experimental and numerical values of resonance frequency were compared. The effects of each driving frequency variation on the vortex formation and the heat transfer coefficient were analyzed.

Table 1 Summary of velocity and frequency References

Velocity (m/s)

Voltage supply (V)

Resonance frequency (Hz)

[14]

0–5

15

4500

[15]

6–8

10

100

[16]

1–10

12

4500

[17]

6.3

13

700

[18]

3–22

10

280

Pavlova and Amitay [6] stated that at a low z/d ratio (axial distance/nozzle diameter), the high frequency at 1200 Hz was more effective in removing the heat than the low frequency at 500 Hz. Chaudhari et al. [9] investigated the low frequency in the heat transfer enhancement with focus on the z/d ratio. Persoons [10] declared that the efficiency of a synthetic jet is typically equivalent to the ratio of the fluidic power to the electrical power, as in Eq. (1). ε=

Q force Pelectric power

(1)

However, in this study, the heat transfer was compared to the power input to quantify the efficiency, ε, for measuring the cooling capability. Table 1 summarizes the velocity ranges and frequencies used in previous studies. Researchers [11–13] agreed that the maximum amplitude was obtained at the resonance frequency, which must be determined because each design has its own resonance value. Each researcher used different values of resonance driving frequency for their fabricated synthetic jet device. Liu et al. [17] mentioned that few studies have explored the low frequency ranging from 200–800 Hz. Moreover, limited information is available on vortex formation related at the 200–800 Hz frequency and on jet fluid characteristics. The effect of resonance frequency in propagating heat transfer performance was experimentally determined and numerically substantiated in this paper by correlating it with the induced velocity and the vortex formation profile. Cater and Soria [19] suggested that a zero-net mass flux jet was formed when the vorticity was adverted away from a generator at a

123

2 Governing Equations The fluid motion, which was initiated by the diaphragm movement, can be described by the conservations equations. The conservations of mass, momentum, and energy were solved by ANSYS FLUENT® using Cartesian spatial coordinates. The air was assumed incompressible and laminar in the model. The governing equations employed in ANSYS FLUENT® to describe the transient fluid flow are as follows. The conservation of mass is ∂ ∂ρ + (ρu i ) = 0 ∂t ∂ xi

(2)

Equation (1) is in the general form, which is valid for incompressible and compressible flows. The momentum (non-accelerating reference frame) is expressed as  ∂τi j ∂P ∂  ∂ ρu i u j = − + + ρgi + Fi (ρu i ) + ∂t ∂x j ∂x j ∂x j

(3)

where ρ is the fluid density; P is the pressure in the fluid; τi j is the viscous stress tensor; and gi and Fi are the gravitational acceleration and external body force in thei-direction, respectively. ANSYS FLUENT® allows the user to simulate moving and deforming domains (i.e., diaphragm movement) through the User-Defined Function (UDF) [18,19]. Dynamic meshes can be used to model the flow where the shape of the domain changes due to motion. The integral form of the transport equation for a general scalar () with an arbitrary control volume (V) on a moving mesh is written as follows: d dt





  ρ u − u g dA dV  ∇dA + S dV

ρdV + 

V

= dv

(4)

V

where u is the flow velocity vector, and ug is the grid velocity of the moving meshes. The first and second terms on the

Arabian Journal for Science and Engineering

Fig. 2 Detailed dimensions of the synthetic jet model

left-hand side are the time derivative and convective terms, respectively. The terms on the right-hand side are the diffusive and the source terms.  is the diffusion coefficient, and S is the source term of . ∂ V represents the boundary of the control volume V , and dA is the area movement. In this study, the UDF was used to simulate the diaphragm movement using a dynamic layering meshing technique. The UDF contains the periodic diaphragm movement by the sinusoidal equation as follows: y = A sin (ωt)

ω ADdiap ν

(6)

where is Ddiap is the hydraulic diameter of the vibration envelope and is expressed as Ddiap =

4 Aw 2A + w

where A is the amplitude; and w is the width.

ANSYS Mechanical ® is a popular finite element method (FEM) software for structural analysis. In this study, the modal analysis in ANSYS Mechanical ® was used to determine the resonance driving frequency for the synthetic jet design. The computational fluid dynamic package in ANSYS FLUENT® was then used to model the fluid flow and the vortex formation. The synthetic jet design was fabricated using rapid prototyping machine for characterization.

(5)

where A is the diaphragm amplitude; ω is the angular frequency; and t is time. The amplitude motions were the highest at the resonance frequency and created a high-velocity fluid output at the nozzle. The Reynolds number for the diaphragm (Rediap ) was defined using the maximum deflection of the diaphragm in the center [or the product of frequency and amplitude (ω A)], the kinematic viscosity (v) of the working fluid, and the same length scale chosen for Nu [20,21]: Re =

3 Methodologies

(7)

3.1 Computational Domain and Boundary Conditions Figure 2 shows the detailed dimensions of the synthetic jet model with a 5 mm-deep volume chamber containing a 2 mm nozzle, as illustrated in Fig. 2(iii). Figure 2 provides the crosssectional views of the center of the synthetic jet. Structured meshing was applied to save computing time and observe the uniformity of the fluid characteristics. The computational domain was built based on the dimension of the synthetic jet by using a preprocessing software. Table 2 presents the model dimensions and operating range. The volume chamber was fixed at 6.1 mm3 , whereas the frequency was varied from 300 to 700 Hz at 100 Hz increments. The simulation was conducted for 1000 cycles, and each cycle required 100 time steps. A total of 10,000 time steps were needed to obtain a steady velocity output. The internal iteration continued until the residues of mass and momentum were reduced to below 10−3 and 10−6 , respectively,

123

Arabian Journal for Science and Engineering

(a) Wall: u = v = w = 0 (b) Outlet pressure; P = 0 Pa (c) Diaphragm (UDF) sinusoidal movement; y = A sin (ωt)

Table 2 Parametric range for the numerical setup Parameter

Unit

Diaphragm frequency

300–700 Hz

Diaphragm diameter

50 mm

Nozzle diameter

2 mm

300 Hz 400 Hz 500 Hz 600 Hz 700 Hz

0.00004

Amplitude (m)

0.00002

0.00000

-0.00002

-0.00004 10100

10150

10200

10250

10300

Time (s)

Fig. 3 Sinusoidal motion of the diaphragm surface monitoring

which were the convergence criteria for the computation. The data were extracted at every 45th and 85th time steps, at which the diaphragm positions were at the ejection and suction phases, respectively, for one cycle. Figure 3 shows the diaphragm surface monitoring for the sinusoidal motion that was compiled using the UDF in ANSYS FLUENT® . The diaphragm at the top layer of the synthetic jet with sinusoidal motion that created the ejection and suction phases was defined in the UDF. The detailed boundaries in the computational domain are shown in Fig. 4. The boundaries and initial conditions were as follows: Fig. 4 Computational domain and its boundary conditions

123

3.2 Turbulent Model The shear-stress transport (SST) K –ω. model was used as the model well blend the robust and accurate representation of the K –ω. model in the near-wall region flow with the free-stream independence of the K –ε model in the far field [22,23]. Semi-Implicit Method for Pressure Linked Equation (SIMPLE) solution method with an implicit solver construction was used as the numerical algorithm meant for computational efficiency suitability of unsteady flow with small time step [24]. The first-order scheme was utilized for pressure, kinetic energy, specific dissipation rate, and energy, whereas the second-order scheme was applied for density and momentum.

3.3 Grid Independence Test In the CFD analysis, another factor that contributed to the over/underestimation of the numerical prediction was number of grids of the computational domain that needed prudent handling. Four different numbers of structured meshes within the range of 30 × 103 to 100 × 103 elements were investigated using grid independence test. Table 3 summarizes the Table 3 Summary of meshing element

Meshing type

Element count

Coarse

25,000

Medium

45,000

Fine

60,000

Ultra-fine

90,000

Arabian Journal for Science and Engineering Table 4 Fluid velocity percentage difference between the numerical and experimental values for different meshing types Meshing type

Velocity (m/s)

Percentage difference

Experimental

1.91

0

Coarse

1.68

12

Medium

1.75

8

Fine

1.88

1.57

Ultra-fine

1.88

1.57

The domain with fine elements displayed a better settlement after 20,000 time steps with the smallest difference among the number of elements. Therefore, the fine elements were further investigated using K –ω SST and SIMPLE at different frequencies applied, and the velocity output was measured.

3.4 Experimental Setup This study conducted three types of experiments, namely diaphragm movement, frequency, and heat transfer performance. First, Keyence® Laser Doppler experiment was used to determine the exact sinusoidal movement of the diaphragm of the synthetic jet device. Then, Impact Hammer LMS frequency experiment was adopted to validate the resonance frequency of the synthetic jet device. Lastly, the effects of various driving frequencies of the synthetic jet on the heat transfer performance were experimentally investigated under the condition that the heater was supplied with 30 W power.

3.5 Laser Doppler Experiment

Fig. 5 Grid independent tests

element count for coarse, medium, fine, and ultra-fine meshing types, which were further investigated using K –ω SST and SIMPLE. The results were verified by using the experimental output velocity. Table 4 presents the effects of meshing element counted from the velocity output in the experiment and the percentage difference. Among them, the fine meshing with 60,000 counted elements had a closer value with a 1.57% difference.

Figure 6 illustrates the experimental setup of the Keyence® Laser Doppler experiment. The synthetic jet device (vii) with a reflector was fixed at the transverse axis (viii) with antivibration characteristics. This anti-vibration characteristic was vital for this experiment, because the laser sensor (vi) sensitivity captured 1 µm of deflection or movement. Any nearby vibration affected the obtained reading. The synthetic jet device applied 10 V peak to peak, which was amplified (v) by the sinusoidal function by using an amplified function generator (iv). The sinusoidal function was applied with various driving frequencies from 300 to 700 Hz at increments of 100 Hz. This setup was similar to the simulation setup. The synthetic jet device was stabilized after 5 min. The deflection data for each driving frequency were collected at 20,000

Fig. 6 Keyence® Laser Doppler experimental setup. (i) Monitor, (ii) central processing unit (CPU), (iii) data acquisition, (iv) function generator, (v) amplifier, (vi) laser sensor, (vii) synthetic jet device, (viii) transverse axis

123

Arabian Journal for Science and Engineering Fig. 7 Sinusoidal movement of the surface monitoring

4.00e-5

Amplitude (m)

2.00e-5

0.00

-2.00e-5 300 Hz 400 Hz 500 Hz 600 Hz 700 Hz

-4.00e-5

0

100

200

300

400

500

600

Time (s)

Fig. 8 Modal analysis experimental setup. (i) Monitor, (ii) central processing unit (CPU), (iii) LMS system, (iv) report (anti-vibration table), (v) synthetic jet device, (vi) accelerometer axis

data/s for 15 min. The Laser Doppler results were plotted and recorded in the computer system (i and ii) via DAQ (iii). Figure 7 shows that the synthetic jet device had different amplitudes for each driving frequency, such that each cycle was completed differently. The lowest driving frequency took a long time to finish, whereas the highest driving frequency finished quickly. The maximum amplitude was recorded at 0.00004 m at a 500 Hz driving frequency, indicating that the resonance frequency for the fabricated synthetic jet device was at the 500 Hz driving frequency. For this reason, the resonance frequency experiment will be expanded further.

3.6 Resonance Frequency Experiment Arik [14], Bhapkar et al. [25], and Chaudari et al. [26] posited that a high h can be obtained at the resonance frequency of the fabricated device. However, the resonance frequency for

123

the sinusoidal diaphragm was disproportionate because of the varying diaphragm piezo material abilities and sweep volume dimensions. Therefore, a modal analysis experiment was conducted to determine the exact resonance frequency of the fabricated synthetic jet. Figure 8 shows the diagram of the experimental setup of the modal analysis experiment. The modal analysis experiment was selected because it is ideal for small lightweight structures, and the synthetic jet device was only 50 g. The synthetic jet device (v) was hanging freely at the retort stand (iv) attached on an anti-vibration table. First, the accelerometer (vi) was calibrated with a LMS© calibrator vibrator to check and validate its sensor data output. Then, the accelerometer (vi) was attached to the synthetic jet diaphragm to capture the device movement during testing. An impact hammer was used to initiate an impulse, which caused a constant amplitude in the frequency domain. As the hammer knocked, the

Arabian Journal for Science and Engineering

8000

2

Amplitude ((m/s )/N)

10000

6000

4000 ANSYS Mechanical Exp.

2000

0 0

200

400 Frequency (Hz)

600

800

Fig. 9 Modal analysis mode shape result

3.7 Heat Transfer Experiment

Fig. 10 Deformation at the first mode shape

accelerometer transferred the movement signal data to the LMS© system (iii), which converted the analog to digital data. A software in the host computer (ii) analyzed the given data, amplitude, and frequency. The gathered data were displayed on the monitor (i) and recorded in the computer’s data storage. Figure 9 shows the effects of the amplitude results on the driving frequency under the first mode shape for the experiment and numerical values. The highest amplitudes obtained experimentally and numerically were at 9303.34 and 8620 (m/s2 )N−1 , respectively, in the driving frequency

Fig. 11 Synthetic jet heat transfer experimental setup

range of 300 to 800 Hz. Both experimental and numerical values agreed that the resonance driving frequency for the designed synthetic jet was at 500 Hz, and that the amplitude percentage difference was 7.35%. Figure 10 illustrates the deformation that occurred on the synthetic jet diaphragm under the first mode shape. The contour’s color represents the intensity of the deformation that occurred on the diaphragm of the synthetic jet from blue (lowest deformation) to red (highest deformation). The highest deformation occurred at the center of the diaphragm with 4.8 × 10−4 m under the first mode shape because of the resonance frequency. This first mode shape has a similar sinusoidal motion to that shown in Fig. 9.

Signal Generator

Figure 11 shows the experimental setup for testing the capability of the fabricated synthetic jet device at various driving frequencies when heat flux was applied to the heated surface. A driving frequency range of 300–700 Hz was applied to the synthetic jet device. Rapid prototyping was used to fabricate the synthetic jet volume (casing) using ABS material, and the piezoresistive diaphragm was attached on the top of the synthetic jet casing. The GW instek function generator was used to drive the connected diaphragm movement. The supply voltage was amplified to increase the peak to peak 5 V as the frequency was varied from 300 to 700 Hz at 100 Hz increments. A Tektronix oscilloscope was attached to the function generator to monitor and verify the supplied signal pattern to avoid signal drift or distort. A 40 mm square heat source with 20 W was used to generate heat flux to obtain a 343 K heated surface temperature. The K-type thermocouple was suitable for a wide range of operating temperature up to 1200 ◦ C and was located at the center of the area below the heated surface. The temperature data on the heat source surface and ambient were recorded using the data acquisition system. This system operated by converting the analog data of the resistance value obtained at the welded tip of

Oscilloscope

Synthec Jet Data Acquision (DAQ)

Amplifier Heater Heat Insulator

Computer

123

Arabian Journal for Science and Engineering

the thermocouple based on the Wheatstone bridge circuit to temperature. A hot wire type anemometer was used for the velocity output from the nozzle and was located at 20 mm from the nozzle, which was the same distance for monitoring in ANSYS FLUENT® analysis. The temperature on the heated surface was monitored and recorded for 10 min to obtain heat and temperature stability. The synthetic jet device must be stabilized to avoid errors from the signals supplied or piezoresistive material readiness. The velocity output from the nozzle was monitored for 5–10 min until it produced a constant sinusoidal velocity pattern, which indicated that the diaphragm was soothed. The changes in the temperature on the heated surface were recorded in a DAQ and plotted on a computer.

ity output from the nozzle. When the diaphragm approached its resonance frequency, a pair of single vortex flows formed in the cross-flow. However, the number of vortex flows increased when the frequency of the diaphragm was slightly away from its resonance frequency. A new vortex pair was developed once the first vortex pair reached the surface. This secondary vortex pair disturbed the first vortex, resulting in a low fluid momentum on the heated surface and in less heat being swept. This phenomenon occurred because of the rate of one cycle completion for each driving frequency, as previously shown in Fig. 9. This trend was consistent with the findings of Pavlova and Amitay [6] and Mahalingam et al. [29].

4.2 Suction Phase

4 Results and Discussions Simulations were conducted using ANSYS FLUENT® 15 software with the setup details mentioned in Sect. 3.1. The geometry model was fabricated for the synthetic jet used in the experiment. Then, the results were validated against the experiment ones in terms of diaphragm deformation, vortex formation at different frequencies, velocity output, decreased temperature, and amplitude–temperature effect.

4.1 Numerical Simulation (Ejection phase) Figure 12 shows the velocity contour during the ejection phase at different frequencies from 300 to 700 Hz. All contours were obtained at 10,045 time steps, indicating that the simulation was stable before 1000 time steps. The velocity output was unstable at the beginning. Vortex pair formations were observed at the cross-flow area for 300 Hz, and the velocity was 0.34 m/s at 10 mm from the nozzle output (Fig. 14a). By contrast, only a pair of vortex formations was observed for 400 Hz (Fig. 14b), and their velocity was approximately 0.70 m/s. The velocity outputs were 0.5 and 0.35 m/s at 600 (Fig. 14d) and 700 Hz (Fig. 14e). Both frequencies produced a small vortex pair regardless of the vortex frequency. Figure 12c shows the vortex formation and velocity contour for 500 Hz as the diaphragm achieved its resonance frequency with a large diaphragm amplitude (Sect. 3.2). The highest velocity of 1.52 m/s with a large-sized vortex pair formation was recorded. The 500 Hz diaphragm produced a large-sized vortex pair formation with a maximum velocity output. The vortex formation during the ejection phase was important because it carried a high momentum, which swept away the heat on the heated surface, as mentioned in previous studies [27–29]. These phenomena demonstrated that the driving frequency significantly affected the vortex formation and veloc-

123

Figure 13 shows the velocity contour during the suction phase at different frequencies ranging from 300 to 700 Hz. All contours were obtained at 10,125 time steps, indicating that the simulation was stable before 1000 time steps. During the suction phase, the air was sucked into the volume chamber. At low driving frequencies of 300 (Fig. 13a) and 400 Hz (Fig. 13b), the velocity output flow was slow. The moving superficial and small vortex formed because the low fluid momentum was sucked into the volume chamber. At the resonance driving frequency, which had a high amplitude and fluid momentum, as shown in Fig. 13c, the previous vortex pair was fully discharged to flow outward. Thus, more air fluid was sucked into the volume chamber because it had the highest diaphragm amplitude. During the ejection phase, the fluid was impinged on the heated surface, whereas at the suction phase, the fluid was sucked into the volume chamber because of the low velocity. Whenever the frequencies increased, as shown in Fig. 13d, e, the flow was not fully embarked. At 600 and 700 Hz, the vortex had a lower velocity because both had a smaller diaphragm amplitude and a faster rate of one cycle completion.

4.3 Velocity Profile Distribution Figure 14 shows the velocity scattering along the x-position at 10 mm from the heated surface at the cross-flow region to avoid the flow reflection when the ejection flow affects the heated surface for various frequencies ranging from 300 to 700 Hz during a normalized velocity output. The highest and lowest velocities were found at 500 and 300 Hz, respectively. The maximum velocity was 1.08 m/s, and the minimum velocity was 0.3 m/s. The remaining peak velocities were 0.65, 0.37, and 0.32 m/s at 400, 600, and 700 Hz, respectively. Therefore, the normalized velocity depended on the amplitude of the diaphragm and was associated with the driving frequency. A high diaphragm amplitude generally produced a high velocity.

Arabian Journal for Science and Engineering Fig. 12 Vortex formation and velocity contour during the ejection phase at various frequencies (all unit in m)

Frequency (Hz)

Vortex Formation (Ejection Phase)

(a) 300

(b) 400 2.23 m/s

(c) 500

(d) 600

0 m/s

(e) 700

123

Arabian Journal for Science and Engineering Fig. 13 Vortex formation and velocity contour during suction phase at various frequencies (all unit in m)

Frequency

Vortex Formation

(Hz)

(Suction Phase)

(a) 300

2.63m/s

(b) 400

(c) 500 0 m/s

(d) 600

(e) 700

123

Arabian Journal for Science and Engineering Fig. 14 Velocity scattering at x-position for various driving frequencies

0.0

2

Velocity (m/s )

-0.2

-0.4

-0.6

-0.8

300 Hz 400 Hz 500 Hz 600 Hz 700 Hz

-1.0

-1.2 -0.006

-0.004

-0.002

0.000

0.002

0.004

0.006

X-Position (m)

Table 5 Average temperature at various frequencies

Fig. 15 Temperature reduction by the synthetic jet at various driving frequencies

4.4 Heat Transfer Performance One of the main focuses of this study was to assess the extent of cooling achieved by the heat source on the basis of the changes in the driving frequency and its amplitude. Figure 15 shows the effects of the driving frequencies of the diaphragm on the transient temperature distributions for the synthetic jet. At first, the heater was turned on, and the temperature was observed under natural convection condition until a steady state was reached. A heater temperature of approximately 343 K was observed 15 min after operation. As the driving frequencies were applied to the diaphragm, the temperature started to decline within the first 2 min. The device cannot sustain the reduction of the heated surface temperature at driving frequencies of 300, 600, and 700 Hz. The heated surface temperature slightly increased because of the

Frequency (Hz)

Temperature (K)

300

343.0

400

338.6

500

332.6

600

343.0

700

343.0

low velocity, which cannot overcome the applied heat flux. However, the heated surface temperature can yield a temperature drop of 5 K and remain at 339 K at 400 Hz. The best cooling performance was at 500 Hz, at which the temperature of the heated surface was reduced from 343 K to 332 K. The high velocity at the resonance driving frequency of 500 Hz produced a high momentum, as mentioned by [1,15,30], and it reduced the maximum temperature. Table 5 shows the average absolute temperature from 600 s to 900 s, during which the heated surface was soothed at different frequencies. The initial average h under natural convection when the synthetic jet was at 0 Hz was 26.8 W/m2 K. Once the synthetic jet was operated, h increased at various driving frequencies, as shown in Fig. 16. Given that 300, 400, 600, and 700 Hz had lower amplitudes than that of 500 Hz, their h values were approximately 27.40, 29.60, 26.80, and 26.80 W/m2 K, respectively. The maximum h value was 36.80 W/m2 K at 500 Hz, which had the highest amplitude. The maximum heat removal occurred at the resonance driving frequency, which had highest amplitude of the synthetic jet diaphragm. A high velocity was produced during the ejection phase as a large vortex formation occurred on the heated surface. This vortex phenomenon aided the heat dissi-

123

Arabian Journal for Science and Engineering

2

Heat Transfer Coefficient (W/m K)

50

40

30

36.80 29.60 28.60

28.60

28.60

28.60

20

Acknowledgements The authors wish to thank University Teknologi MARA, Universiti Sains Malaysia, and the Government of Malaysia for offering the Research Acculturation Grant Scheme Fund (600RMI/RAGS 5/3 (158/2014), which has allowed us to conduct this work.

10

0

the heat transfer through force convection. The best cooling was achieved on the heated surface when the synthetic jet vibrated at a driving frequency of 500 Hz and a 0.04 mm amplitude. These findings may be valuable in the application of synthetic jets in cooling applications and can be extended for jets with various volume chambers and nozzle diameters.

0

300

400

500

600

700

Frequency (Hz)

Fig. 16 Heat transfer coefficients, (h) at various driving frequencies Table 6 Energy efficiency for distinct frequencies Frequency (Hz)

Energy efficiency ratio

Percentage

0

1.000

300

1.022

2.0

400

1.104

10.4

500

1.373

37.3

600

1.000

0.0

700

1.000

0.0

0

pation by reducing the heated surface, thereby increasing the heat transfer through force convection. A similar pattern had been observed by Deng et al. [3] and van Buren et al. [31]. Table 6 shows the energy efficiency of the system. The maximum energy efficiency occurred at the resonance frequency, and its value was 37.3% better than those at the other driving frequencies. The data proved that applying a driving frequency other than the resonance value was less efficient. This hypothesis was consisted with Lee et al. [32] and Jeng and Hsu [33].

5 Conclusion The effects of the diaphragm amplitude of a synthetic jet were studied experimentally and numerically. The synthetic jet had the highest sinusoidal amplitude when operated at the resonance driving frequency. The maximum sweep volume due to amplitude movement created a high volume of air, which was subsequently jetted onto the heated surface area. In addition, the maximum amplitude produced a high Reynolds number and a high velocity. The air jet with a high velocity created a vortex pair upon hitting the surface area and freely moved to the ambient. This phenomenon aided the heat dissipation by reducing the temperature, thereby increasing

123

References 1. Holman, R.; Utturkar, Y.; Mittal, R.; Smith, B.L.; Cattafesta, L.: Formation criterion for synthetic jets. AIAA J. 43, 2110–2116 (2005) 2. Silva-Llanca, L.; Ortega, A.: Vortex dynamics and mechanisms of heat transfer enhancement in synthetic jet impingement. Int. J. Therm. Sci. 112, 153–164 (2017) 3. Deng, X.; Luo, Z.; Xia, Z.; Gong, W.; Wang, L.: Active-passive combined and closed-loop control for the thermal management of high-power LED based on a dual synthetic jet actuator. Energy Convers. Manag. 132, 207–212 (2017) 4. Firdaus, S.M.; Omar, H.; Azid, I.A.: High sensitive piezoresistive cantilever MEMS based sensor by introducing stress concentration region (SCR). In: Finite Element Analysis-New Trends and Developments, pp. 1–24. InTech (2012) 5. Randy, A.; Andika, A.; Rhakasywi, D.: The characteristics of cooling on heat sink using a cross flow synthetic jet actuated by variation of wave function. WSEAS Trans. Heat Mass Transf. 8(3), 83–90 (2013) 6. Pavlova, A.; Amitay, M.: Electronic cooling using synthetic jet impingement. J. Heat Transf. 128(9), 897 (2006) 7. Smith, B.L.; Glezer, A.: The formation and evolution of synthetic jets. Phys. Fluids 10(9), 2281–2297 (1998) 8. Krishnan, G.; Mohseni, K.: An experimental study of a radial wall jet formed by the normal impingement of a round synthetic jet. Eur. J. Mech. B/Fluids 29(4), 269–277 (2010) 9. Chaudhari, M.; Puranik, B.; Agrawal, A.: Effect of orifice shape in synthetic jet based impingement cooling. Exp. Therm. Fluid Sci. 34(2), 246–256 (2010) 10. Persoons, T.: General reduced-order model to design and operate synthetic jet actuators. AIAA J. 50(4), 916–927 (2012) 11. Qayoum, A.; Gupta, V.; Panigrahi, P.K.; Muralidhar, K.: Perturbation of a laminar boundary layer by a synthetic jet for heat transfer enhancement. Int. J. Heat Mass Transf. 53(23–24), 5035–5057 (2010) 12. Lievano, C.F.P.: Experimental and Computational Study of a Zero Net Mass Flux Synthetic Jet. Washington University, St. Louis (2008) 13. Ben Chiekh, M.; Ferchichi, M.; Béra, J.C.: Modified flapping jet for increased jet spreading using synthetic jets. Int. J. Heat Fluid Flow 32(5), 865–875 (2011) 14. Arik, M.: Local heat transfer coefficients of a high-frequency synthetic jet during impingement cooling over flat surfaces. Heat Transf. Eng. 29(9), 763–773 (2008) 15. Bhapkar, U.S.; Srivastava, A.; Agrawal, A.: Acoustic and heat transfer characteristics of an impinging elliptical synthetic jet generated by acoustic actuator. Int. J. Heat Mass Transf. 79, 12–23 (2014)

Arabian Journal for Science and Engineering 16. Chaudhari, M.; Puranik, B.; Agrawal, A.: Multiple orifice synthetic jet for improvement in impingement heat transfer. Int. J. Heat Mass Transf. 54(9–10), 2056–2065 (2011) 17. Liu, Y.-H.; Tsai, S.-Y.; Wang, C.-C.: Effect of driven frequency on flow and heat transfer of an impinging synthetic air jet. Appl. Therm. Eng. 75, 289–297 (2015) 18. Lin, C.; Bai, C.; Hsiao, F.: An investigation on fundamental characteristics of excited synthetic jet actuator under cavity and diaphragm resonances. Procedia Eng. 79, 35–44 (2014) 19. Cater, J.E.; Soria, J.: The evolution of round zero-net-mass-flux jets. J. Fluid Mech. 472, 167–200 (2002) 20. Kimber, M.; Garimella, S.V.; Raman, A.: Local heat transfer coefficients induced by piezoelectrically actuated vibrating cantilevers. J. Heat Transf. 129(9), 1168 (2007) 21. Abdullah, M.K.; Ismail, N.C.; Abdul Mujeebu, M.; Abdullah, M.Z.; Ahmad, K.A.; Husaini, M.; Hamid, M.N.A.: Optimum tip gap and orientation of multi-piezofan for heat transfer enhancement of finned heat sink in microelectronic cooling. Int. J. Heat Mass Transf. 55(21–22), 5514–5525 (2012) 22. Chandratilleke, T.T.; Jagannatha, D.; Narayanaswamy, R.: Heat transfer enhancement in microchannels with cross-flow synthetic jets. Int. J. Therm. Sci. 49, 504–513 (2010) 23. Argyropoulos, C.D.; Markatos, N.C.: Recent advances on the numerical modelling of turbulent flows. Appl. Math. Model. 39, 693–732 (2015) 24. Silva-Llanca, L.; Ortega, A.: Vortex dynamics and mechanisms of heat transfer enhancement in synthetic jet impingement. Int. J. Therm. Sci. 112, 153–164 (2017) 25. Bhapkar, U.S.; Mohanan, S.; Agrawal, A.; Srivastava, A.: Interferometry based whole-field heat transfer measurements of an impinging turbulent synthetic jet. Int. Commun. Heat Mass Transf. 58, 118–124 (2014)

26. Chaudhari, M.; Verma, G.; Puranik, B.; Agrawal, A.: Frequency response of a synthetic jet cavity. Exp. Therm. Fluid Sci. 33(3), 439–448 (2009) 27. McGuinn, A.; Farrelly, R.; Persoons, T.; Murray, D.B.: Flow regime characterisation of an impinging axisymmetric synthetic jet. Exp. Therm. Fluid Sci. 47, 241–251 (2013) 28. Ikhlaq, M.; Ghaffari, O.; Arik, M.: Effect of actuator deflection on heat transfer for low and high frequency synthetic jets. In: IEEE Intersociety Conference on Thermal and Thermomechanical Phenomena in Electronic Systems, pp. 882–888 (2014) 29. Mahalingam, R.; Heffington, S.; Jones, L.; Schwickert, M.: Newisys server processor cooling augmentation using synthetic jet ejectors. In: IEEE Intersociety Conference on Thermal and Thermomechanical Phenomena in Electronic Systems, pp. 705– 709 (2006) 30. Mahalingam, R.: Modeling of synthetic jet ejectors for electronics cooling. In: Annual IEEE Semiconductor Thermal Measurement and Management Symposium, pp. 196–199 (2007) 31. van Buren, T.; Whalen, E.; Amitay, M.: Interaction between a vortex generator and a synthetic jet in a crossflow. Phys. Fluids 27(10), 107101 (2015) 32. Lee, A.; Yeoh, G.H.; Timchenko, V.; Reizes, J.A.: Flow structure generated by two synthetic jets in a channel: effect of phase and frequency. Sens. Actuat. A Phys. 184, 98–111 (2012) 33. Jeng, T.-M.; Hsu, W.-T.: Experimental study of mixed convection heat transfer on the heated plate with the circular-nozzle synthetic jet. Int. J. Heat Mass Transf. 97, 559–568 (2016)

123