conference slides_Final

Editor's Notes

• #2: Hello everyone. I hope you’ve been having a pleasant day so far. In this presentation I’m going to introduce our recent project concerning the “power efficiency of piezoelectric fans”.
• #3: First, I’m going to introduce the concept of piezoelectric fans, their important merits over conventional rotary fans, and their principal of working. Secondly, we will review some backgrounds related to this work and some previous works that have been done, their contributions and the motivation of our project. Next, comes the measurement instruments and techniques that we used. After that, we will go over the results of a portion of our experiments that lead us to the next part, which is conclusion. Finally, ongoing and future works.
• #4: The piezoelectric fan is composed of 2 parts: piezoelectric actuator and a blade (usually made of polymer). The piezo actuator consists of two layers of PZT ceramics on top and bottom with a brass shim sandwiched at the middle. By applying an alternating electric field, the piezo-actuator starts to vibrate. When the driving frequency equals the resonance frequency of the fan blade, the maximum vibration amplitude can be obtained, as well as the maximum induced flow that can be employed to create forced convection. Compared to traditional rotary fans, the piezoelectric fan ****consumes much less power, around 15mW compared to 130mW for an equivalent rotary fan ****introduces less noise in delicate and noise sensitive electronics. ****can be easily scaled down due to its super-simple mechanism.
• #5: The above picture is an example of substituting traditional rotary fan by an array of piezo fan in chip cooling application. Some great works studying the piezoelectric fan cooling have been done by others. For instance, Acikalin and others investigated the effect of various factors on the COOLING performance of the piezoelectric fan like relative position from the heat source, amplitude, frequency etc. Wait and other coworkers ... Kimber derived correlations for forced heat transfer coefficient for the piezo electric fan vibration. In electronics cooling, the power consumption can be a great concern due to limited battery capacity. However, there isn’t much work done in systematic study of the cooling efficiency in terms of power consumption. Therefore, we took a chance to dig a little deeper in the power consumption of piezoelectric fan. As mentioned in the literature, the power consumption of piezoelectric fan can be obtained by measuring its alternating voltage and current and integrate through the time of one period.
• #6: In doing so, we have a function generator and an amplifier to drive the piezoelectric fan. The amplifier can output monitoring signal for both voltage and current. The Labview Vi is created to display and record the voltage and current values. Finally, a MATLAB script is created to fit the signal with two sine functions, which eliminates noise, and to do the integration in getting power.
• #8: According to previous studies, the power consumption of PZT actuator concentrates on three categories, internal friction, dielectric loss and heat generation. As shown by the plot on the left, the power consumption of solo PZT actuator without the fan blade is linearly proportional to the frequency at a given input voltage. However, the total power consumption of piezoelectric fan shows a peak at the fan blade’s resonance frequency. Clearly, a portion of total power consumption comes from the fan blade, and we believe it is the flow power that the fan blade provides to surrounding air via vibration. To estimate the flow power, we need to create a dynamic model for piezoelectric fan.
• #9: The piezoelectric fan is modeled as a 2 DOF mass-spring-damper system with effective mass 𝑚, stiffness 𝐾, linear damping coefficient 𝐶, aerodynamics damping coefficient 𝐶 𝑎 and an excitation force 𝐹 0 . Previous studies only include linear damping in the governing equation. Therefore, we did a comparison between curve fitting the blade’s amplitude by having only linear damping coefficient and aerodynamic damping coefficients. Plot on the left Includes aerodynamic damping coefficient. the aerodynamic damping of PZT actuator, 𝐶 1𝑎 , is assumed to be zero. Aerodynamic damping of blade, 𝐶 2𝑎 , can then be determined by fitting the measured amplitude. As shown above, the amplitude obtained from including aerodynamic damping fits the measured amplitude better.
• #10: The flow power 𝑃 𝑓 can then be determined based on blade’s vibrating amplitude 𝑦 2 , frequency 𝑓 and aerodynamic damping coefficient 𝐶 2𝑎 As expected, the flow power peaks at the fan blade’s resonance frequency. And if we compare the total power and flow power, shown by the plot on the right, there is a finite difference between them. And this difference represents the parasitic power dissipation in the PZT actuator. We postulated that the flow power is the portion of power related to heat transfer performance.
• #11: In this experiment, we change the configuration of the piezoelectric fan by changing the thickness of the blade by adding layers of polyester, and as a result, the resonance frequency also changes. A maximum number of 4 layers was used by which we could obtain a range of natural frequencies from 35Hz to 119Hz. We varied the voltage for each configuration from 70V to 140V.
• #12: In the left hand side, the power normalized heat transfer performance (as a representative for power efficiency) is shown vs. blade thickness. On the right hand side you see the same curve, this time normalized by flow power. As it is clear, normalizing by flow power leads to a convergent curve.