WelcomeGeneral InfoProgramTutorialsAbstractsAuthors/AwardsExhibitionRegistrationAbout HawaiiContact

Wireless Networked Sensing

Fiber Optic Sensor Technology

Analog Portable Sensors

Phosphor-Based Sensors

Energy Scavenging Storage

Neutron Imaging Sensors

Transducers Testing by Optical

Optical Fibre Nanowire Sensors

Tutorial 1c

Paul Wright
University of California, Berkeley, USA

Tutorial Description:

Energy scavenging, harvesting, and storage methods will be presented and analyzed in this workshop. Powering sensor nodes without the use of replaceable batteries is the main goal of the research work behind the workshop. In most applications the sensor nodes are part of a wireless communication network. First, we will consider the wide spectrum of scavenging techniques, including: mechanical to electrical, light, RF, thermal, wind, etc. and how they might meet the energy needs of "smart devices." Second, we will consider energy storage issues and the types of rechargeable batteries and super capacitors that can meet storage and form factor constraints. Third, we will discuss information on the spectrum of sensors that are used in various application settings (commercial buildings, automobiles, industrial equipment, and airframes etc.) and the energy needs that consequently arise from sensing, communications, and overall functional operation. Finally we will analyze the "Demand and Supply" design space. Here, it will be shown that - given certain conditions - self-powered nodes are already commercially viable if the energy source is plentiful and the data-rate and duty-cycle of a wireless sensor node are modest. Future applications will depend on lower powered radios (the demand side), improved power conditioning (the efficiency factor), and, of course, designs and materials that can scavenge greater amounts of energy from ambient sources (the supply side).

1. Energy Scavenging
Energy scavenging offers the possibility of powering sensor nodes without replaceable batteries and a number of approaches have been studied over the past years. Since the energy consumption of a node is limited by its size and efficiency, optimization of the complete power train is essential. This includes energy generation, storage and conversion. These future, self-powered nodes must have ultra-long life. (e.g. our research with the building industry indicates a need for >10 year life). Photovoltaics are a mature technology, and a solar cell based power source can be implemented using commercial off-the-shelf technology. However, photovoltaics obviously are less useful inside buildings, factories, and for many medical applications. Therefore researchers have pursued naturally occurring temperature variations; shoe inserts capable of generating 8.4 mW of power under normal walking conditions; and airflow turbines. Vibration energy scavengers based on electromagnetic, electrostatic, and piezoelectric conversion have been suggested in the literature. It is likely that no single energy scavenging solution will fit all environments and application spaces. Rather, unique solutions need to be considered on an application by application basis.

Case Study: Energy Scavenging from Ambient Vibrations
Meso-Scale devices made from Lead Zirconate Titanate (PZT) will be discussed as a first case study focusing on a two-layer PZT cantilever which, when driven by vibrations, provides an AC voltage. A resistor connected across the electrodes creates a simple RC circuit. The overall application uses the assumption that any device reduced to practice would be designed so that its natural frequency would match the frequency of the surface upon which is would be mounted - thus creating the maximum possible power output. Experimental results demonstrate a power transfer to a resistive load of 375 mW/cm3 from driving vibrations of 2.5 m/s2 at 120 Hz. We have shown that a custom designed radio transceiver that consumed 12 mW when transmitting could be powered at a duty cycle of 1.6% by a 1 cm3 generator. It was also possible to drive a complete system -- consisting of a vibration-based scavenger, power conditioning circuit, capacitors for storage, a wireless beacon, and a temperature sensor -- from the vibrations of foot-traffic on a building stairwell. A semi-active RFID tag was also modeled, prototyped and thus permanently powered from vibrations.

Case Study: Energy Scavenging from Ambient Heat Sources
Well-known as the Seebeck effect, an electrical voltage can be produced when two dissimilar metals in a closed circuit have their junctions heated to different temperatures. Using this effect we have powered wireless sensors in environments where waste heat is available. Industrial metallurgical operations were thought to be ripe for such wireless monitoring, with aluminum smelting being an obvious choice given its inherent thermodynamic inefficiency and fluoride pollutants. Given the desired duty cycle of the sensor measurements it will be shown that the waste heat provides the necessary power to continuously operate the wireless system. The erratic air flow experienced in the smelting plants, required modifications to be made to standard heat sinks and thermoelectric generating modules (TEGs) to improve their thermal performance. Also it was necessary to model, design and then prototype a passive power conditioning circuit system to increase and stabilize the TEG's output voltage despite electrical load and temperature gradient fluctuations, while still supplying sufficient current.

2. Energy Storage
Micro-fabrication of 'on-PCB' solid state electrochemical cells (for storing the above scavenged energy) will also be described. Low cost pneumatic dispenser-printer will be described, consisting of an electronically controlled variable pressure dispenser unit, a micron resolution three-axis stepper motor controlled stage and a standard PC. Common lithium-ion battery components can be printed including slurries such as LiCoO2, graphite, Li4Ti5O12, PVDF and BMIM-TFSI. In addition, metal nano-pastes of silver can be printed. Gold and copper formulations are also in preparation. Numerical simulations indicate that cells under 1 cm2 in area require thicknesses in excess of 50 µm per electrode to provide adequate capacity for most applications, including wireless transmission, and long term logging. The printer setup produces electrodes of micron level roughness in thicknesses of 30 µm to 100 µm. In addition, PVDF can be printed with thicknesses on the order of 20 µm. Since PVDF is an excellent host material for ionic liquids, we expect this ability will provide the link to fabricating a fully printed lithium-ion polymer cell. Preliminary capacities of ~300 µAh have been realized on cells with total electrode thickness of 150 microns and a footprint of 25 mm2.

3. The "supply and demand" trade off
Considering the "supply-side," described above, the experiments and vibration model indicate that a 1cm3 prototype can generate 375 mW from a vibration source of 2.5 m/s2 at 120 Hz. It is however important to balance this against the demand of a node. Assuming an average distance between nodes of approximately 10m (which means that the radio transmitter should operate at approximately 0 dBm), the peak power consumption of the radio transmitter will be around 2-3 mW depending upon its efficiency. Using ultra-low power techniques, the receiver should not consume more than 1-2 mW. Including the dissipation of the sensors and peripheral circuitry, a maximum peak power of 5 mW is quite reasonable. For a maximum data-rate for the radio of 100 Kbit/sec, and an average traffic load per node of 1 Kbit/sec (these numbers are based on real radio prototypes and a realistic smart home scenario), every node operates at a duty cycle of approximately 1%. During the remaining 99%, the only activities taking place in a node are a number of background tasks: low-speed timers, channel monitoring and node synchronization. Combining peak and standby power dissipation leads to an average power dissipation of approximately 100 mW. Thus considering the "demand-side" at an average power consumption of 100mW (which is still an order of magnitude smaller than any node that is currently available) a sensor node needs slightly more than 1 cm3 of lithium battery volume for 1 year of operation, assuming that 100% of the charge in the battery is used. So, given a 1cm3 size constraint, it would be necessary to update standard replaceable batteries in a node every ~9 months. (Another option is to increase the size of the node, which is also not acceptable in many scenarios). This example will be described in detail to show that a constant delivery of 375 mW from the vibration source is sufficient to support the node provided the rechargeable batteries from the printing process are present to store power if the vibration source is intermittent or is temporarily absent.

4. Societal Impacts of Low-Cost Micro-Integrated Sensor Nodes
Smart infrastructure monitoring and control projects are typical applications of such self-powered nodes relying on energy scavenging. The workshop will cover a variety of applications in energy efficiency - both for industrial and residential settings - and in "demand response" which uses wireless-based technologies to monitor energy use and lower it when prices or demand are high. Wireless sensor networks and displays help emergency responders prioritize search and rescue missions. The same visual information is combined with other occupancy sensors and egress displays so that the 'command and control officers' overseeing a rescue can monitor the progress of rescue operations and communicate with selected firefighters.


Paul Wright is a Professor of Mechanical Engineering at the University of California, Berkeley and Director of the Center for information Technology Research in the Interest of Society (CITRIS). Among his many publications on this topic, he is a co-author (with Shad Roundy and Jan Rabaey) of the book Energy Scavenging for Wireless Sensor Networks with Special Focus on Vibrations, (Kluwer Academic Publishers).

Copyright 2009, All Rights Reserved, Sensors 2010 and interteq.com