in partial fulfillment for the award of the degree of Bachelor of Technology in Electrical and Electronics Engineering of
COCHIN UNIVERSITY OF SCIENCE AND TECHNOLOGY
DEPARTMENT OF ELECTRICAL ENGINEERING GOVERNMENT MODEL ENGINEERING COLLEGE KOCHI 682 021 NOVEMBER 2013
CERTIFICATE
This is to certify that the seminar report entitled
Submitted by is a bonafide account of the work done by him/her under our supervision
ACKNOWLEDGEMENT
During the course of my seminar work, there were many people who were influential in helping to achieve the target. Without their guidance, assistance and patience, it would never have been able to accomplish this seminar presentation. I would like to take this opportunity to acknowledge some of them. Firstly, I would like to thank Prof.(Dr). V.P Devassia, Principal, for providing the opportunity to do this seminar and for the moral support in the seminar. I sincerely express my gratitude to Dr. Bindu V, HOD, the seminar coordinator and Mrs.Shilpi M, the guide who channeled me all through this seminar and provided valuable suggestions. I am also thankful to all the other staff, friends and my family for all the technical and moral support they have given throughout the course of this seminar work. Above all, God is the sole reason of my success.
ABSTRACT
Networks of low power wireless devices are increasingly used in applications ranging from environmental to factory automation monitoring. Most of these devices must be operative 24hrs a day and may be in locations where manual battery replacement is difficult or costly. It would be desirable if there exists a miniaturized device that can convert ambient mechanical energies such as vibrations, which are readily available 24hrs a day, to power wireless devices. Over the past decade, piezoelectric cantilever energy harvesters have been increasingly investigated for this application. The challenges are two folds: improving the voltage and power output within the constraints of size and weight.
To increase the vibration energy harvesting capability of the piezoelectric generator based on a cantilever beam, a piezoelectric generator that not only uses the strain change of piezoelectric components bonded on a cantilever beam, but also employs the weights at the tip of the cantilever beam to hit piezoelectric components located on the 2 sides of weights is designed . A prototype of the piezoelectric generator has been fabricated and its characteristics have been measured and analyzed.
CONTENTS
List of figures List of abbreviations Chapter 1: Introduction 1.1. Piezoelectric effect 1.2. Piezoelectric materials 1.3. Anisotropic effects and coupling modes 1.4. Device configuration 1.5. Modes of vibration 1.6. Vibration sources 1.7. Rectification and storage Chapter 2: Classification 2.1 Bulk power generation 2.2 Micro power generation Chapter 3: Experimental model of a piezo electric generator 3.1 Construction 3.2 Experimental method 3.3 Experimental Results 3.4 Experimental setup for microscale energy harvester Chapter 4: Applications Chapter 5: Conclusion REFERENCES i ii 1 1 3 4 6 6 7 8 9 9 9 10 10 12 15 21 23 24 26
LIST OF FIGURES
Figure 1: Energy generation using piezo effect Figure 2: Structure of PZT Figure 3: Quartz crystal Figure: 4 Various modes of piezoelectric materials Figure 5: Configuration of a piezoelectric material Figure 6: Modes of Vibration Figure 7: Structure and size of a piezoelectric material Figure 8: Electrical connection in generator Figure 9: (a) Energy harvested and voltage built across capacitor (b) Waveform output of voltage across capacitor Figure 10: (a) Experimental setup for energy harvesting (b) Energy harvested and voltage built across capacitor Figure 11: Voltages when rectified voltages are applied Figure 12: Energy harvested and voltage built across capacitor per excitation Figure 13: Simultaneously measured waveforms of the rectified voltages Figure 14: The average output power during the charging for different deformation Figure 15: Overhead view of one 400m beams Figure 16: Side view of beams showing the bending caused by the mass of the beam 2 3 4 5 6 7 11 13 16 16 17 17 18 19 20 21 21 22
LIST OF ABBREVIATIONS
PZT- Lead Zirconate Titanate IDE- Inter digitated Electrodes MEMS- Micro Electro Mechanical Systems
ii
CHAPTER 1 INTRODUCTION
The term energy harvesting is popularly used when electricity is generated from sources such as ambient temperature, vibrations or air flows. Since there are now electronic circuits whose power requirement is of the order of milliwatts, even though its energy yield is relatively low, energy harvesting with piezo-based solutions is always of great interest in situations where electricity cannot be supplied via power cables and one wants to avoid batteries and the maintenance effort required. Energy harvesting by piezoelectric devices has great potential applications in self-powered sensor networks, portable electronic devices, ubiquitous computing systems, and other areas. In recent years, energy harvesting using piezoelectric materials has become a very popular research topic.
Various device sizes and structures have been tested, but it is difficult to compare power measurements as device fabrication and experimental methods vary. It uses piezoelectric effect to convert mechanical vibration or the strain variation with time into electric energy and store it in energy storage devices such as super-capacitors and rechargeable batteries. In an effort to standardize comparisons in spite of these changing parameters, the dependence of generator power output on device dimensions has been investigated.
When a piezo crystal mechanically deforms as a result of a force applied with tension or pressure, charges are generated which can be measured as a voltage on the electrodes of the piezo element, a phenomenon known as the direct piezo effect.
This method of charge generation is familiar from gas ignition systems to generate the ignition voltage, for example. The charge generated (Q) can be described by the mathematical expression below: Q = d F
The charge constant d (ratio of charge generated to force applied) in this equation is a materialspecific constant of the order of 10-10 C/N.
It therefore quickly becomes apparent that the quantity of charge generated is relatively low. This aspect places high demands on mechanical systems and electronics in order to harvest the optimum amount of energy.
A majority of piezoelectric generators that have been fabricated and tested use some variation of lead zirconate titanate (PZT). Typically, PZT is used for piezoelectric energy harvesters because of its large piezoelectric coefficient and dielectric constant, allowing it to produce more power for a given input acceleration. Another less common material is aluminum nitride (AlN). Though it has a smaller piezoelectric coefficient and dielectric constant, aluminum nitride has advantages in material deposition and in compatibility with the standard CMOS processes used for fabrication of integrated circuits.
More recently, there is growing concern regarding the toxicity in lead-containing devices driven by the result of restriction of hazardous substances directive regulations. To address this concern, there has been a resurgence in the compositional development of lead-free piezoelectric materials.
While the piezoelectric coefficient is higher in the 3-3 mode for most materials, taking advantage of the larger coefficient requires a much more complex design. Instead of simple planar electrodes, a series of interdigitated electrodes (IDE) can be used to take advantage of the 3-3 coupling mode. However, this approach leads to a very small device capacitance and therefore a high output impedance, making load matching difficult. Another disadvantage is that the IDE approach only works for electrically-poled piezoelectrics such as PZT. In AlN, the direction of polarization of the material is set during deposition, so fabrication of a 3-3 mode device would be prohibitively complex. For this reason, the devices analyzed in this report will only utilize the 3-1 coupling mode.
lower frequencies. Therefore, energy harvesters are generally designed to operate in the first resonant mode.
Each mode of vibration has a characteristic mode shape. This describes the deflection of the beam along its length. Figure shows some examples of mode shapes for the first three vibrational modes of a beam. When a beam is vibrating in a particular mode, the deflection will vary sinusoidal with time, with the amplitude of the sine wave along the length of the beam given by the mode shape. The points where the mode shape is zero are stationary and are referred to as nodes. In general, the nth vibrational mode will have n nodes.
After rectification, the DC voltage is used to charge a capacitor or battery. This allows the device to draw more power over a short period than the harvester is able to provide. DC-DC conversion schemes have also been explored and have been shown to charge batteries far more efficiently.
CHAPTER 2 CLASSIFICATION
Based on scale of generating power piezoelectric power generation can be classified as bulk power generation and microscale power generation.
10
Vibration energy is harvested by the piezoelectric components distributed in 3 groups, i.e., groups PZT-T, PZT-M, and PZT-B. Group PZT-T is located at the end of the upper plate of holder, PZT-M at the root of the brass beam, and PZT-B at the end of the lower plate of holder. The piezoelectric plates in groups PZT-T, PZT-M, and PZT-B have rectangular shape. Material Fuji C203 is used for PZT-T, PZT-M, and PZT-B, with k31 of 0.35, k33 of 0.71, Qm of 2000, and tg of 0.3. PZT-T and PZT-B are bonded onto the end of holder and PZT-M onto the brass beam 3 mm away from the nut stacks. Two identical aluminum plates, Al1 and Al2, are bonded onto the lower surface of PZT-T and upper surface of PZT-B, respectively.
11
12
When the generator is subjected to external vibration or shock, flexural vibration of the brass beam can be excited. As the brass beam vibrates, it hits the piezoelectric plates PZT-T and PZT-B and causes the strain in PZT-M to change with time. Because of the piezoelectric effect, these physical processes can cause an output voltage across the piezoelectric plates. Connecting this voltage to a capacitor via a rectifier, a dc voltage is obtained across the capacitor and small amount of energy is harvested. In the operation of the generator, PZT-T and PZT-B mainly operate in the k33 mode and PZT-M operates in the k31 mode. The purpose of using aluminum plates on the surfaces of PZT-T and PZT-B is to avoid the breaking of piezoelectric plates in PZT-T and PZTB owing to the direct hit.
During the experiments, the piezoelectric generator was placed and fixed on the top of a spring, as shown in Figure 8. Pressing down the generator with the spring and then releasing, the generator would vibrate with the spring. As the generator-spring system vibrates, AC voltage was generated across the piezoelectric components, and energy can be harvested. AC voltage from each group of piezoelectric plate was applied to a full wave rectifier, the output port of which was connected to
13
a capacitor to store electric energy. Two methods were used in the management of rectified output voltages from the 3 groups of piezoelectric plates. In one of them, 3 DC voltages from rectifiers were connected in series and then the total DC voltage was applied to a 30-F capacitor, as shown in Figure 9.
In the connection, the positive output terminal of one rectifier was connected to the negative output terminal of next rectifier. In another one, DC voltage from each group was applied to different capacitors (30 F) individually, as shown in Figs. 9(b) and (c). In both of the methods, the 2 piezoelectric plates in group PZT-M were connected in parallel, based on their poling directions. The following experimental steps are defined as one excitation: 1) The generator is pressed down together with the spring so that the spring has a compression deformation x; 2) Then the compression force is released;
14
3) After the release, the generator does an attenuated vibration together with the spring until the vibration stops. Energy harvested per excitation Ec is used to express the generators capability of harvesting energy. When the initial voltage of the capacitor is zero, Ec can be calculated by
1 2 2
where C is the capacitance of the capacitor and Vc is the DC voltage across the capacitor after the charging caused by one excitation. To make the initial voltage of the capacitor be zero, the capacitor is completely discharged before each measurement. The AC voltage across piezoelectric components and dc voltage across capacitor were measured by an oscilloscope. Resonance frequency fs of the generator-spring system can be calculated by
1 2
where k is the force constant of spring and Ms the total mass of the generator and spring. In our experiments, k is 1192 N/m and Ms is 88.1103 kg. Therefore, the resonance frequency is calculated to be 18.5 Hz, which is also the vibration frequency of the generator after the compression force is removed. The lowest resonance frequency fb of the cantilever brass beam can be calculated by
1 3 2 3
where E is the Youngs modulus of the beam, I the inertial moment of cross-sectional area of the beam, Mw the mass of the iron weights at the tip, and Le the effective length of the beam. The inertial moment I can be calculated by
3 = 12
where W is the width of the beam and t the thickness of the beam.
15
16
Figure 10: (a) Energy harvested and voltage built across capacitor (b) Waveform output of voltage across capacitor
To investigate the harvested energy without using the hit mode, the brass beam was rotated 180 as shown in Fig. 11(a), and energy harvested per excitation and voltage built across the capacitor per excitation were measured for different initial compression deformation of spring x. Fig. 11(b) shows the result when the distance H between the upper and bottom plates of holder was 46.2 mm. With the comparison of Figs. 10(a) and 11(b), it is known that the piezoelectric components operating in hit mode can substantially enhance the energy harvesting of the piezoelectric generator based on a cantilever beam.
17
Figure 11: (a) Experimental setup (b) Energy harvested and voltage built across capacitor per excitation versus initial deformation of spring
To analyze the cause of the above-described phenomenon, PZT-T, PZT-B, and PZT-M are connected to their own capacitors, as shown in Figs. 9(b) and (c), and waveforms of the AC voltage across each group of piezoelectric component(s) and of the DC voltages across capacitors are recorded. Fig. 12. In Figs. 12(a), (b), and (c), the signals shown in the upper part of the figures are the DC voltage across capacitor and the ones shown in the lower part of the figures are the AC voltage across piezoelectric component(s). From the waveforms of the AC signals, it is seen that the amplitude of the AC voltage across PZTT (Vpzt-t) is 1.4 V and the one across PZT-M is about 0.78 V (Vpzt-m). The difference between them is not very large. However, from the waveforms of the DC signals, it is seen that the rectified voltage from PZT-T or PZT-B is much larger than that from PZT-M. This experimental result indicates that the large difference in harvested energy from PZT-T (or PZT-B) and PZT-M may be caused by the nonlinear I-V characteristic of diodes in the rectifier.
18
Figure 12: Voltages across the piezoelectric components and capacitors when the rectified voltages from PZT-T, PZT-M, and PZTB are applied
The forward bias of each rectifier is about 1.0 V, which is less than Vpzt-t and larger than Vpztm.
For this reason, the charging current from PZT-T is much larger than that from PZT-M. This
causes the large difference in harvested energy from PZT-T (or PZT-B) and PZTM. Based on this analysis, it is known that if we increase Vpzt-m, the difference in harvested energy from PZT-T (or PZT-B) and PZT-M may be reduced, and they may have a similar order of magnitude because it is not difficult to increase Vpzt-m from 0.78 V to 1.4 V by optimizing the location and size of PZTM. From Fig. 12, it is also seen that before the charging, there are some small AC voltages across the capacitors. These voltages are caused by the structure vibration of the holder, excited by the impact at the release. For the electric connection shown in Figs. 9(b) and (c), energy harvested per excitation and voltage built across the capacitor per excitation were measured for different H, and the results for PZT-T and PZT-B are shown in Fig. 13(a) and (b), respectively.
19
Figure 13: Energy harvested and voltage built across capacitor per excitation versus the distance H between the upper and lower plates of the holder when the rectified voltages from PZT-T, PZT-M, and PZT-B are applied to individual capacitors
From Fig. 13, it is seen that the energy harvested per excitation increases as the distance H between the upper and bottom plates of holder increases in the range H < 50 mm. This phenomenon occurs because, for H <Hc, there is not enough time for the iron weights at the tip to develop enough speed before hitting. Phase differences between the rectified voltages from PZT-T and PZT-B were compared, as shown in Fig. 14.
It is known that the time of charging the capacitor connected to PZT-B lags behind that of charging the capacitor connected to PZT-T, which is the time it takes for the weights to swing from aluminum plate Al1 to Al2
20
.
Figure 14: Simultaneously measured waveforms of the rectified voltages from PZT-T and PZT-B, caused by the same excitation.
And 22.3% of the harvested energy is wasted due to the series connection of full wave rectifiers. The loss mechanism is analyzed as follows: 1) When PZT-T is hit by the weights, PZT-T turns on its full wave rectifier partially and charges the capacitor C through the other 2 full wave rectifiers. Hence, besides the 2 diodes in the rectifier connected to PZT-T, there are 8 more diodes in the charging loop. 2) After the charging by PZT-T, PZT-B supplies an AC voltage. It turns on the full wave rectifier of PZT-B partially and charges the capacitor C through the other 2 full wave rectifiers. Hence, besides the 2 diodes in the rectifier connected to PZTB, there are 8 more diodes in the charging loop. Comparing the above operation with the one charging the 3 capacitors separately, it is known that the energy harvesting based on Fig. 9(a) has 8 more diodes in the charging loop, which should result in more electric loss. Hence, choosing the full wave rectifiers with less energy dissipation and lower forward voltage drop may decrease this electric loss and increase the charging current. Fig. 15 shows the average charging power for the connection shown in Fig. 9(a) at different initial spring deformation. The average charging power can be estimated by the measured harvested energy per excitation and charging time.
21
Figure 15: The average output power during the charging at different initial spring deformation when the rectified voltages from PZT-T, PZT-M, and PZT-B are connected in series.
Due to the time constraints, the design and fabrication of a generator was not feasible. Instead, a cantilever on an older wafer was chosen for analysis. The cantilever to be tested is approximately 400m long and 100m wide, with tPt = 0.2m and tAlN = 1m. There is actually a series of
22
cantilevers, with lengths of 400, 300, 200, and 100 m. However, the longest cantilever would exhibit the greatest response, so the 400m was chosen for testing. Since the cantilevers were not designed with energy harvesting in mind, they lacks proof masses. This design is not optimal, but it does serve to illustrate many of the characteristics that would need to be taken into account.
Figure 17: A side view of two beams showing the bending caused by the mass of the beam
23
CHAPTER 4 APPLICATIONS
The best-known application is the electric cigarette lighter: pressing the button causes a springloaded hammer to hit a piezoelectric crystal, producing a sufficiently high voltage electric current that flows across a small spark-gap, thus heating and igniting the gas. Gas burners now have builtin piezo-based ignition systems. Battery-less wireless doorbell push button. The armed forces toyed with the idea of putting piezoelectric materials in soldiers boots to power radios and other portable electronic gear
Several nightclubs, mostly in Europe have already begun to power their strobes and stereos using the force of hundreds of people pounding on piezoelectric lined dance floors. Capitalizing on the friction and heat created by walking, running and even just wearing jeans, engineers from Michigan Technological University, Arizona State University devised a way to use this type of generated energy to charge portable electronic devices, like iPods and mobile phones. Biomechanical Energy Harvester. Energy harvesting by Piezoelectric windmills
24
CHAPTER 5 CONCLUSION
It has been shown that the current device structure does not have scaling advantages in power per unit area or volume. It is difficult to produce sufficient displacement at small scales to generate a considerable voltage. At the microscale, resonance frequencies are too low to effectively convert ambient frequencies as found in nature. As an illustration of the challenge of reaching lower frequencies with MEMS devices, consider a beam of equal width and thickness as the one tested, but a significantly larger length and an added proof mass. Using the frequency estimate in section 3.1.2, a 2000m by 100m by 2.6m beam would need to have a 0.11mg proof mass to reach even 200 Hz. Using a relatively dense metal such as gold would still require a cubic proof mass of approximately 148m on each side. Due to the difficulty in reaching low frequencies with MEMS scale devices, these types of energy harvesters would be limited to applications with very high frequency vibrations. However, for compact systems with very low power requirements, MEMS micro-generators are a very attractive means of powering devices indefinitely. Recommendations are to build devices of this form, with a proof mass added, while targeting lower resonant frequencies. Alternate geometries may help in lowering the resonant frequency, and gaining more power output. More effective solutions include designing a structure that is either not dependent on resonance, or has a means of tuning its resonant frequency. Examples of such devices have already been demonstrated by other researchers. To take advantage of the large deflections and strains that go with a beam oscillating at resonance, the tuning approach is recommended as the most useful for power output. The challenge will be in adapting existing tuning approaches to the MEMS scale, or in devising a new means to tune the beams resonant frequency.
In this paper, we report a piezoelectric generator that not only uses the strain change of piezoelectric components bonded on a cantilever beam, but also employs the weights at the tip of the beam to hit piezoelectric components located on the 2 sides of weights. It shows that the piezoelectric components operating in the hit mode can substantially enhance the energy
25
harvesting of the piezoelectric generator based on a cantilever beam. The rectified voltages from different groups of piezoelectric components can also be connected in series and then be applied to a capacitor to store energy. However, in this case, there is more electric energy loss in the charging loop. The motion and impact of the weights at the tip of the cantilever beam have been theoretically analyzed, which well explains the experimental phenomena and suggests the measures to improve the generator.
26
REFERENCES
[1]. Vibration Energy Harvesting Based on Integrated Piezoelectric Components Operating in Different Modes, Junhui Hu, Januar Jong, and Chunsheng Zhao, IEEE transactions on ultrasonics, ferroelectrics, and frequency control, vol. 57, no. 2, February 2010 [2]. Vibrational energy harvesting using mems piezoelectric generators, Andrew Townley [3]. Vibration and flow energy harvesting using piezoelectric, Xiaotong Gao [4]. Energy harvesting using piezoelectric material, Rajeev Kumar, School Of Engineering, IIT Mandi [5]. Energy harvesting from vibration and walking with piezoelectric materials, Mikko Leinonen, Jaakko Palosaari, Maciej Sobocinski, Jari Juuti and Heli Jantunen University of Oulu, Microelectronics and Material Physics Laboratories, Linnanmaa [6]. Performance evaluation of vibration-based piezoelectric energy scavengers, Yi-Chung Shu [7]. http://www.wikipedia.com [8]. http://www.iopscience.com [9]. http://www.greenoptimistic.com