System design and modeling of a piezoelectric energy harvesting module
Energy harvesting systems (EHS) use externally-derived energy to power small autonomous devices. They provide unique capabilities for numerous applications, such as tire pressure monitoring systems (TPMS). Conventional TPMS are powered by batteries and are mounted on a wheel rim. Assembling the monitoring system on the inner liner of the tire enables additional parameters to be detected,1 such as tire temperature, friction, wear, and side slip. These parameters are useful for optimized tracking and engine control. However, this innovative use of EHS sets severe requirements for implementation. For instance, the EHS must be robust within the tire environment, weigh less than seven grams to avoid tire imbalance, and last at least eight years. Implementing these last two requirements is especially challenging with a battery-based approach, and favors implementation of energy harvesting microelectromechanical systems (MEMS).
MEMS implementation faces numerous challenges. Large dynamic forces can occur within the tire environment, ranging from tens to thousands of gravitational acceleration units. Movable parts of the generator device must be designed to tolerate these forces. Therefore, a conventional mass-loaded cantilever design2 with a seismic mass in the gram range is critical. Furthermore, there is no stable frequency spectrum available within the tire environment. Therefore, the conventional concept of tapping mechanical energy by stimulating the generators' seismic mass with an excitation at the resonant frequency is not suitable. Alternative generator concepts must be developed that exhibit a minimum of mass and operate with a non-resonant excitation scheme. The harvesting module should also be less than 100mm2 in area to replace the battery. Furthermore, a minimum average power of 3μW is required to send data every 60 seconds, considering the power consumption of the pressure sensor, microcontroller, and radio-frequency front end.
Addressing these challenges, we considered a piezoelectric MEMS generator to power an automotive wireless sensor node for tire pressure monitoring. The system design and modeling included the material properties impact, generator design, interface circuit topology, and operation conditions on relevant system parameters (such as output voltage and power level). Our system modeling was based on analytical calculations and allowed identification of a suitable design space.
Our design (see Figure 1) used a piezoelectric MEMS generator without a seismic mass. The intrinsic mass of the cantilever was in the microgram region, and the resulting acceleration forces within the tire environment were very small. Instead of the mass-acceleration force, we used the deformation energy during tire shuffle entry to mechanically stress the cantilever up to a defined maximum value and released it at the shuffle exit. The cantilever then started oscillating. During each cycle, electrical energy was extracted by the interface circuit to provide it to the load (see Figures 2 and 3). The cantilever amplitude decayed exponentially until it was reset to the initial excitation value at the shuffle entry.
: Piezo layer permittivity. Sp11: Piezo layer Young's modulus. Sc11: Carrier layer Young's modulus.
The generator possessed a triangular shape to induce a uniform stress distribution in the horizontal surface direction. The stress in the vertical direction is schematically shown in Figure 3. The cantilever consisted of a silicon carrier layer and a piezo layer.3, 4 The carrier layer served three purposes: providing mechanical stability to the structure, containing the neutral axis, and acting as a storage element for harvested mechanical energy. Designing the MEMS generator required finding suitable values for these parameters for a given carrier and piezoelectric material. We used an analytical system model for this purpose.
The interface circuitry must rectify the generator voltage and transfer the primary energy from the piezoelectric capacitance into the buffer capacitance. The capacitor elements must be electrically connected for some period of time. A crucial system design issue is whether or not to use an inductive element in that process. If the connection is established by a resistive path,5 an external inductor as well as related costs and additional weight are avoided. However, in this case, at most a quarter of the energy harvested at the piezoelectric element can be extracted. We used an inductive interface scheme,6, 7 and the transferred energy was increased. The costs related to the external inductor are negligible because of the reduced MEMS generator area. Therefore, we used a system model to investigate the tradeoffs on the interface scheme.
Powerful simulation tools are available for component design. Typically, the finite element method is used for MEMS generator design, and Spice is suitable to simulate the interface circuitry. However, there are no well-established general interfaces between the two modeling domains to perform a system simulation. So, we used an analytical approach8 to gain system-level insight. Then we analytically calculated all relevant system parameters by following the scheme outlined in Figure 4. Next, we determined the mechanical energies based on the mechanical stress distribution in the cantilever layers and obtained the electrical energy for a particular cantilever state with the piezo material constants. Finally, we directly calculated the open-circuit voltage from this energy, using the piezo capacitance. Transformation of the primary electric piezo energy into a load power depends on the operational conditions (e.g. tire rotation speed) and on the interface circuit. As a result, we considered resistive5 and inductive6, 7 schemes. Based on the analytical modeling, we calculated the system parameters for a given generator area over the remaining geometrical design space of the carrier and piezoelectric layer thicknesses. The calculation results were used to identify a design space consistent with the given requirements (see Figure 4).
Our analytical modeling approach has predicted optimal piezo and carrier thicknesses for resistive and inductive behavior in an energy harvesting module for TPMS applications. Future work will involve redesigning the MEMS generator, designing a dedicated application-specific integrated circuit interface, and system integration.
Alexander Frey received his MA from the University of Texas in 1994, his Diploma in Physics from the University of Wuerzburg, Germany, in 1997, and his PhD from Saarland University, Germany, in 2010. His current research interests are in the field of complimentary metal-oxide-semiconductor-based biosensors and energy-autonomous microsystems.
Ingo Kuehne received his diploma in engineering physics, and his MEng in micro- and nanotechnology, from Munich University of Applied Sciences in 2002 and 2004, respectively. He received his PhD from Saarland University in 2009. His current research interests are in MEMS pressure sensors, biosensors, and energy-autonomous microsystems.
Julian Seidel received his Diplom-Ingenieur in Mechatronics from the University of Applied Sciences in Munich, Germany, in 2010. He is currently working towards an MEng. His current interests relate to designing an MEMS energy harvester for energy-autonomous microsystems.
