By Chris Lang
Hundreds and even thousands of sensors are currently being used to monitor a wide variety of phenomena including greenhouse gases, air pollution, landslides, tsunamis, earthquakes, and the structural integrity of pipelines and other machinery [1] [2] [3]. With these networks comes the necessity to power each sensor. Traditional methods of power distribution fail when it comes to these systems. It is not feasible to run power lines to thousands of sensors distributed throughout the Pacific Ocean. Batteries have been the predominant source of energy for mobile devices in the past, but they are not an ideal solution, as they have a limited lifespan and must be replaced at considerable cost [4]. Despite an increase in battery energy density, figure 1 shows that they are not keeping pace with the required power consumption of other technological advancements. Again, batteries are unable to meet the demands of a large-scale sensor network. Additionally, batteries often contain contaminants that are harmful to the environment and can leak into the nearby ecosystems if simply left out [5]. What is needed is an autonomous and renewable energy source to power low-voltage, low-power sensor networks across the world.
Figure 1[4]
There is ambient energy around us in many forms. Solar radiation, vibration energy, radio frequency (RF) waves from radio and TV towers, and thermal gradients can all be harvested through various means to generate power [6]. Figure 2 shows the available energy in various sources of power. Previously, these power outputs were too low to be used effectively in any sensor network. However, recent advances in lower-power processors have made these processors viable for powering sensor networks [7].
Figure 2[6]
While outdoor light provides the greatest power, it is only available in the daytime, meaning that systems must either shut down, or store enough power to last the night; this storage is often too demanding. Additionally, many sensor networks are in places that are not exposed to light. Oil pipelines in Alaska, for example, are under the cover of darkness for multiple months every year, making sunlight harvesting impossible [8]. Many sensors are also inside other machines which block the available sunlight. Finally, the processes for making solar panels are often expensive and harmful to the environment. What is needed is a low-cost energy harvesting solution that can generate power in any environment, no matter the conditions. Vibration energy is the most viable candidate.
Vibration energy harvesting has the second highest production of power of any of the previously mentioned sources (for industrial settings), and is present in almost any situation [6]. Almost all objects around us have some harvestable vibration, even if it is too small to notice through touch or sight. The flight of a moth produces vertical oscillation, a macroscopic vibration; a window contains a microscopic vibration in its glass pane. [9] [10]. While both vibrations are small, they can still be harvested for energy on the scale of microwatts. The vibration spectrum for a window can be seen in Figure 3 [10]. Note that the vibration amplitude is very small, on the order of one micrometer. The frequency however is relatively high, at 70 hertz.
Figure 3[10]
The main disadvantage of vibration energy harvesting is the scale of power that it can produce [6]. Vibrations only produce a small amount of energy, so they cannot be used to fuel power-intensive devices. However, for remote sensors which use little power, vibrations provide an ideal energy source. Three major mechanical-to-electrical energy converters, or generators, are most often used when designing vibrational-energy harvesters: those based on variable capacitors, piezoelectricity, and moving magnets [11].
Variable Capacitor
Figure 4
A variable capacitor generator is simple in concept. It comprises two separated metal plates, one with a positive electrical charge, and one with a negative electrical charge. The voltage varies in proportion to the separation and charge of the plates. When the two capacitor plates are separated, the voltage and the potential energy held in the capacitor increase [11]. The greater the separation, the more work is done, and therefore the more energy is stored in the capacitor. In this generator, one of the plates remains fixed to the vibration source, while the other is attached to a spring and mass, as can be seen in Figure 4. As the source vibrates, the spring will move up and down, changing the distance between the plates and, therefore, the energy stored in the capacitor. In order to generate electrical energy, the charge must be released when the voltage is greatest. Once it has been released, the charge must then be reapplied to the plates in order to recreate the initial conditions, so the process can be repeated. This is done with a simple circuit, as seen in Figure 5. The variable capacitor is centered in the circuit, and is connected to a second, non-variable, capacitor, which serves as a temporary energy reservoir. Between these two is a diode that will only allow electricity to flow towards the second capacitor. This occurs as the energy in the capacitor increases. Additionally, a battery is connected to the variable capacitor through a diode that will only allow electricity to flow towards the variable capacitor. This occurs only when the capacitor energy decreases as its plates compress. Finally, a flyback mechanism is used to syphon the temporarily stored energy in the second capacitor back to the battery. When the voltage of the variable capacitor is greatest, charge and energy flow into the second capacitor, and are temporarily stored before being sent back to the battery through the flyback mechanism. The charge in the variable capacitor is then depleted, and it compresses again. When it gets back to its compressed state, charge and energy from the battery then flow to the variable capacitor, which reverts to its original state. This cycle provides a net increase in energy because each time the capacitor is refilled it requires less energy than is sent back to the battery [11]. More energy is sent due to work done by the vibration during the expansion of the capacitor. This process happens repeatedly, constantly providing net power to the battery, and subsequently to whatever is being powered.
Figure 5
Another factor that must be taken into consideration is the frequency of the vibration of the object. Ideally, the resonant frequency should be high, in order to maximize the number of energy-producing cycles of the circuit per second [11]. If the object vibrates only once per second, then the circuit will only power the battery once per second. However, it is also important for the amplitude of the vibration to be great, in order to increase the energy converted per cycle. The more the plates are separated, the more energy is added to the system, creating more power to the battery [11].
Figure 6[12]
When an object vibrates, it generally has more than one vibration [12]. It may move back and forth slowly across a long distance, while at the same time vibrating very quickly across a short distance. Below is Figure 6, showing the vibration spectrum of a gear pump. This figure shows that there are many frequencies at which the pump vibrates, some with greater amplitudes than others. When considering the design of the harvesting generator, choosing the correct spring is imperative. It and the additional mass must have a resonant frequency that matches a frequency present in the vibration spectrum. When choosing which frequency to match, one must take into consideration the magnitude of the frequency (higher is better as it leads to more cycles a second), as well as the amplitude (more energy per cycle). One common problem that accompanies the construction of variable capacitors, and other generators, is that the frequencies with the highest amplitude are usually the ones with the lowest frequency [11]. If you move your arm up and down, it may convert a lot of energy, but you cannot do it hundreds of times a second. Instead, vibrations of much lower amplitude, so small that they are undetectable to our sight, must be chosen in order to increase the frequency and generate the necessary power to fuel a remote device.
Piezoelectricity
Piezoelectricity is a separation of charge that occurs when a piezoelectric material is put under strain. Common piezoelectric materials include many crystals, ceramics, and bone [14]. Piezoelectric materials have a very specific lattice structure which when deformed separate positive and negative charges. The lattice for lead titanate, a common piezoelectric material can be seen in Figure 7. When no strain is applied, the lattice appears symmetric. There is an evenly spaced and alternating arrangement of positive and negative ions, with a highly charged center, Ti4+. When compressed, the Ti4+ ion gets pushed to one side by the O2- ions. This happens because the atomic radii define a volume in which the Ti4+ can exist. Because two atoms cannot overlap, the Ti4+ gets pushed to the side. This happens for every lattice unit in the entire material. The direction of displacement for all Ti4+ atoms is the same, because when one is pushed towards another, the like charges will repel, and the second Ti4+ will be pushed away from the first, in the same direction [11]. Once this happens, a charge separation is established in the piezoelectric material. One side is positive, the side to which the Ti4+ is pushed, and the other is negative, the side from which the positive ion was taken in which only negative ions remain. However, the inside of the crystal remains uncharged, because for every Ti4+ ion pushed away, another will move towards where it was taken from. Only the surfaces of the material have non-neutral charges.
Figure 7[14]
Once the piezoelectric charge difference has been established by a distortion of the piezoelectric material, it can be used to power a system, for example a remote sensor. The piezoelectric surface charge is used to induce currents through conductors connecting one charged side to the other. As the piezoelectric material vibrates, the charge separation oscillates. and this will produce current through the external conductors. The type of power produced is AC, not DC power which is more typically used for remote sensors. In order to use the AC power, switches alternate direction of flow through the external conductors [11].
Implementing a piezoelectric generator in a vibration energy harvester can be accomplished with a simple design. On one side, the piezoelectric material is attached to the vibrating object, a pipeline for example, and its other side is attached to a large mass. This design can be seen in Figure 8. When the vibration source vibrates, the piezoelectric material will act like a spring, and this spring and the attached mass resonate exactly as a spring and mass do in elementary physics. As the material bends back and forth, AC charge and voltage are constantly created. The voltage can be presented through the wires to a load where the energy is utilized; it may be changed to DC along the way.
Figure 8
The spring and mass system will oscillate at the same frequency as the vibration source. The spring and mass system will have a unique resonance frequency that will give the greatest amplitude of oscillation. Greater amplitude creates a greater strain on the piezoelectric material, which creates a greater voltage [11]. The resonance of the spring mass system can be altered by changing the mass at the end of the piezoelectric material. Increasing the mass will decrease the frequency but increase the amplitude [11]. In the same way that choosing a frequency for a variable capacitor generator is important, so it is for a piezoelectric generator. A vibration component should be chosen which has a large amplitude and a large frequency, to produce the most strain per oscillation and the most oscillations per second. This design is a commonly used and effective way of harvesting energy from vibrations that are almost unnoticeable. In addition to piezoelectric power generation, moving magnets can also be used to harvest energy, as will be described in the next section.
Moving Magnets
The technique of harvesting vibration energy through moving magnets utilizes the most common form of mechanical-to-electrical energy conversion [11]. This technique relies on Faraday’s Law of Induction, which says that a wire moving through a magnetic field will experience a separation in charge, and hence a voltage will be created along it [15]. There are countless generators that exploit this phenomenon, including all utility generators[11], and one of these can be seen in Figure 9. An alternating series of north-facing and south-facing magnets are laid next to each other, creating a magnetic field that extends from each north-facing magnet to the adjacent south-facing magnets. These magnets are connected to the vibration source through a spring. The wire in the figure is connected to the vibration source. When the source vibrates, the magnets will oscillate back and forth past the wire. Here, too, the spring and mass should resonate at the best frequency of the vibration source. Then, because of the previously mentioned Faraday Induction, a voltage will be created along the wire, pushing current through a load.
Figure 9
The voltage is proportional to the change over time of the magnetic field strength experienced by the wire. The greater the rate of change in magnetic field that the wire goes through, the more voltage is produced. The strength of the magnetic field follows that of a cosine curve. Because voltage is the derivative of the magnetic field strength over time, when graphed, the voltage follows a sine function [11]. Simply running the wire back and forth through the magnetic field creates AC power, rather than the DC required by most electronics. So, when the voltage is negative, switches must be flipped to reverse the direction of flow through the load [11]. This creates an output that is the absolute value of the voltage function.
Three design factors enhance the efficiency of moving magnet generators. Using stronger magnets with greater magnetic fields creates a larger voltage. In addition, changing the direction of the magnets a greater number of times over the distance the wire oscillates increases the change over time of the magnetic field applied to the wire, thereby increasing the voltage even more. Finally, wrapping the wire in a coil multiple times around will increase the number of wire segments that are acted on by the magnets. More turns in the coil mean more total voltage produced by the moving wire [11].
Applications
At this point, you may wonder what the uses are of small amounts of harvested energy. Until recently, there may have been none. However, with the development of low power processors, such as the Sensirion SHT11 and the Chipcon CC2420, these constant sources of low power have begun to find a use [7]. One of the most promising applications of vibration energy harvesting is in sensor networks. Energy harvesting techniques provide ideal power sources for these networks. They provide a number of advantages over batteries, most notably infinite lifespan, lower pollution, and increased cost effectiveness [4]. Building a simple circuit is much cheaper than providing temporary batteries that must be frequently replaced.
Robert Xia and Christopher Farm proposed a sensor device based on piezoelectricity that is capable of harvesting vibration energy from pipelines in Alaska [16]. This sensor will measure the temperature of the pipeline and then transmit it back to a base station to aggregate the data. The temperature of a pipeline is a crucial piece of data that must be constantly monitored. Recently, in Anchorage Alaska, ice built up and interfered with the trans-Alaska pipeline, causing a major leak [1]. In the time that the pipeline was non-operational, the amount of oil being transported from the North Slope was cut to just 5% of its previous output. Because of this leak, the pipeline was shut down for over 50 hours. In addition to losing potential profits from the lost time and oil, the company also had to pay for costly repairs to the pipeline. In response, Katie Pesznecker, a spokeswoman for Alyeska Pipeline Service Co., told reporters that workers have begun to place temperature sensors, similar to the ones proposed by Robert Xia and Christopher Farm, along the pipeline [1]. For a fraction of the cost, sensors powered by vibration energy harvesting could have been added that would have alerted the company to an ice buildup, allowing it to act before it was too late.
These sensors can also be used to monitor machinery in industrial settings. Because these machines often incorporate moving parts, the vibration energy available is often much greater than in other settings. Motor-operated valves, MOVs, are often used to control the fluid flow through a system. One common use is in the control of water used for the cooling of nuclear reactors [2]. As can be imagined, if there was a malfunction in the valve, and the proper amount of water necessary to cool a nuclear reactor was not delivered, the results could be devastating. For this reason, it is imperative that the valves never malfunction. In “Non-Invasive Diagnostics of Motor-Operated Valves,” Jangbom Chai et al propose that a sensor network, powered by vibration energy harvesting, be used to monitor the valve vibrations. In this example, the vibrations are both the power source and what is being monitored by the sensor. A valve under a particular strain will have a unique frequency spectrum that goes along with it [2]. If a valve is improperly lubricated, for example, then it will grind against itself, increasing certain vibrations. By constantly measuring valve resonance frequencies, and then analyzing them, it is possible to predict failures in MOVs before they happen, preventing costly and potentially life-threatening accidents.
There are other uses for vibrational energy harvesting besides the powering large scale sensor networks. Recently, the Defense Advanced Research Projects Agency (DARPA) issued a challenge to remotely control the flight of a moth. Due to the ability for a moth in pupa stage to heal itself, researchers have been able to directly wire into a moth’s brain during that stage and gain control of it, an impressive task by itself [9]. By sending RF signals to a computer onboard the moth, a remote operator is able to send signals to the brain and control the flight path of the moth. Vibration energy harvesting was used to power this crucial onboard computer. The batteries required to power such a computer and receiver were much too heavy for the moth to carry [11], so an alternative source of energy was required. By utilizing the “moving magnets” design, engineers at MIT were able to harvest 1 milliwatt of electrical power from the flight of the moth [9], enough to power the onboard computer and receiver. A camera attached to the moth can be used for surveillance [9]. The military advantages of this are extraordinary. Such moths could be used to spy on enemy troops, look for weapons of mass destruction, or discover enemy plans before they are executed. Without the vibration energy harvesting, this type of system could never have been built.
In addition to government and industrial uses of vibration energy, there have been some commercial products that utilize vibration energy harvesting. Many flashlights powered by moving magnets are on the market today. These flashlights require no batteries and will never run out of power [17]. If the batteries in a hiker’s flashlight were to die at night, he would be stranded without light in the middle of the wilderness. However, with these new batteries, all that is needed to recharge them is to shake them back and forth, a macroscopic vibration. Additionally, watches may be powered by the constant movement of our arms. Seiko commercially released a kinetic watch in 1998 that relies on harvesting the energy available when we move our arms [18].
Conclusion
Vibrations are all around us. Almost every object offers some amount of harvestable vibration. This energy is typically very small, and while it will not replace fossil fuels anytime soon, it does offer a constant and reliable power source to low-power sensor networks. Because it is infeasible to run wires to thousands of sensors, an alternative energy source is necessary. Due to the recent advancements in low-power processors, vibration energy has filled this requirement. By using piezoelectricity, moving magnets, and variable capacitors, large-scale, autonomous sensor networks are now becoming viable. They can be used to collect countless data and then transmit them to a data center where they can be organized and analyzed as, for example, with the network of temperature sensors on a trans-Alaskan pipeline. By sensing temperature, the data center can determine the location of the ice buildup and then notify a worker to clean it before it becomes a major problem. Sensor networks such as these will become even more common in the future.
Sources
[1] Alaska Oil Pipeline Operator Works on Bypass – ABC News. ABCNews.com: Daily News, Breaking News and Video Broadcasts – ABC News. Web. 07 Nov. 2011. <http://abcnews.go.com/Business/wireStory?id=12585733>.
[2] Chai J, Lyon R.H., Lang J.H. Non-invasive diagnostics of motor-operated valves. American Control Conference, 1994 , vol.2, no., pp. 2006- 2012 vol.2, 29 June-1 July 1994
doi: 10.1109/ACC.1994.752427
[3] Wireless Sensor Network. Wikipedia, the Free Encyclopedia. Web. 07 Nov. 2011. <http://en.wikipedia.org/wiki/Wireless_sensor_network>.
[4] Paulo J, Gaspar P.D. Review and Future Trend of Energy Harvesting Methods for Portable Medical Device. Proceedings of the World Congress on Engineering 2010 Vol II
WCE 2010, June 30 – July 2, 2010, London, U.K. pp.909-914.
[5] Batteries. Washington State Department of Ecology | Home Page | ECY WA DOE. Web. 07 Nov. 2011.
[6] Vullers RJM, Van Schaijk R, Doms I, Van Hoof C, Mertens R. Micropower energy harvesting. Solid-State Electronics. Volume 53, Issue 7, July 2009, Pages 684-693, ISSN 0038-1101, 10.1016/j.sse.2008.12.011.
[7] Demirbas M, Chow KY, Wan CS. INSIGHT: Internet-Sensor Integration for Habitat Monitoring. Proceedings of the 2006 International Symposium on World of Wireless, Mobile and Multimedia Networks. IEEE Computer Society, Washington, DC pp. 553-558.
[8] Alaska Facts. John’s Alaska Railroad Page. Web. 07 Nov. 2011. <http://www.alaskarails.org/AK-facts.html>.
[9] Adee S. BBC NEWS | Americas | Pentagon Plans Cyber-insect Army. Biotele. Web. 07 Nov. 2011. <http://www.biotele.com/robomoth.html>.
[10] Lang JH. MEMS EnergyHarvesting. 19 Sept. 2011. PowerPoint.
[11] Lang JH. Vibration Energy Harvesting. Personal interview. 1 Nov. 2011.
[12] Gear Pumps. Condition Monitoring Experts – Azima DLI. Web. 07 Nov. 2011. <http://www.dliengineering.com/vibman/gearpumps.htm>.
[13] Piezoelectricity. Hyperphysics. Web. 07 Nov. 2011. <http://hyperphysics.phy-astr.gsu.edu/hbase/solids/piezo.html>.
[14] Piezoelectricity. Wikipedia, the Free Encyclopedia. Web. 07 Nov. 2011. <http://en.wikipedia.org/wiki/Piezoelectricity>.
[15] Faraday’s Law. Hyperphysics. Web. 07 Nov. 2011. <http://hyperphysics.phy-astr.gsu.edu/hbase/electric/farlaw.html> .
[16] Xia R, Farm C, Choi W; Kim S-G. Self-Powered Wireless Sensor System using MEMS Piezoelectric Micro Power Generator. Thesis (M. Eng.)–Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2006.
[17] How Faraday Flashlights Work. Shake Flashlights Info. Web. 07 Nov. 2011. <http://www.shake-flashlights.com/how-they-work.html>.
[18] Seiko. Wikipedia, the Free Encyclopedia. Web. 07 Nov. 2011. <http://en.wikipedia.org/wiki/Seiko>.
Chris Lang is a member of the class of 2015, majoring in electrical engineering. He was born in Massachusetts and has lived there all his life. He enjoys math and all sciences, especially those with a heavy physics base.
He was inspired to write this essay for his 21W.031 class, “Explorations in Communicating About Science and Technology”. It was meant to inform the reader about an emerging technology that was invented at MIT. Additionally, Chris was inspired to write this piece by his father, a microelectronics researcher.