Lindsay M. Miller

Modeling and experimental verification of low-frequency MEMS energy harvesting from ambient vibrations

Micro-fabricated piezoelectric vibration energy harvesters with resonance frequencies of 31–232 Hz are characterized and deployed for testing on ambient vibration sources in the machine room of a large building. A survey of 23 ambient vibration sources in the machine room is presented. A model is developed which uses a discretization method to accept measured arbitrary acceleration data as an input and gives harvester response as output. The modeled and measured output from the energy harvesters is compared for both vibrometer and ambient vibration sources. The energy harvesters produced up to 43 nWrms g−2 on a laboratory vibrometer and 10 nW g−2 on ambient vibration sources typically in large buildings.

Integration of a low frequency, tunable MEMS piezoelectric energy harvester and a thick film micro capacitor as a power supply system for wireless sensor nodes

This work presents an integration approach towards manufacturing a MEMS piezoelectric vibration energy harvester and an electrochemical capacitor on the same substrate. Vibration energy harvesters have been fabricated to resonate at low frequencies, matching ambient vibrations found abundantly in buildings. For cost-effective resonance tuning, a direct write dispenser printer can be used to print additional mass at the tips of the beams, and is also used to deposit a capacitor in the open space surrounding the beam. The implementation of a power supply on a single platform is of great value especially for autonomous wireless sensors with long lifetime and small device volume requirements.

Coulomb-damped resonant generators using piezoelectric transduction

Switching interface circuits employed with piezoelectric energy harvesters can increase the electrical damping considerably over that achievable with passive rectifiers. We show that a piezoelectric harvester coupled to certain types of switching circuits becomes a Coulomb-damped resonant generator. This allows analysis of such harvester systems within a well-known framework and, subject to practical constraints, allows the optimal electrical damping to be achieved. In the piezoelectric pre-biasing technique, the Coulomb damping is set by a pre-bias voltage whose optimal value is derived as a function of piezoelectric harvester parameters.

Technologies for an Autonomous Wireless Home Healthcare System

We present a design study highlighting our recent technological developments that will enable the implementation of autonomous wireless sensor networks for home healthcare monitoring systems. We outline the power requirements for a commercially available implantable glucose sensor which transmits measurements to an external wireless sensor node embedded in the home. A network of these sensor nodes will relay the data to a base station, such as a computer with internet connection, which will record and report this data to the user. We explore the feasibility of powering these sensors using energy scavenging from both body temperature gradients and vibrations in the home, and discuss our developments in energy storage and low power consuming hardware.

Self-tuning behavior of a clamped-clamped beam with sliding proof mass for broadband energy harvesting

Real world systems rarely vibrate at a single resonance frequency and the frequencies drift over time. Tunable devices exist, but generally need additional energy to achieve frequency adaptation. This means that the benefits in power output from this tuning need to be large enough to power the mechanism itself. Passively self-tuning systems go into resonance without requiring active control. This paper focuses on a passively self-tuning system with a proof mass that can slide freely along a clamped-clamped beam. Under external vibration, the slider moves along the beam until the system goes into resonance. A proof-of-concept design is introduced using either a copper or a steel beam and a 3D-printed ABS thermoplastic proof mass. Successful self-tuning is demonstrated in both cases. The frequencies range from 80 – 140 Hz at accelerations as low as 0.007 g rms. Results show the resonance of the beam and the position of the slider along the beam with time. Furthermore, the dynamic magnification and the proof mass position at resonance are discussed, together with the inherent non-linearities of double-clamped beam resonators. The findings support the hypothesis that the effect of the ratio between proof mass and beam mass outweighs the Duffing spring stiffening effects.

A Design Methodology for Energy Harvesting: With a Case Study on the Structured Development of a System to Power a Condition Monitoring Unit

Abstract A design methodology is proposed for electronic systems powered by energy harvesting. The methodology first considers the operating environment. It then evaluates the supply-side (the attributes of the harvester), the demand-side (the engineering application or load which receives and uses the converted power), and the power conditioning needed between supply and demand. A test case is presented in which the vibrations of an electromagnetic device are harvested, converted, and used to power a wireless sensor node. Such a node is being used for the condition based monitoring of manufacturing equipment.

Scheme for improved integration and lifetime for piezoelectric energy harvesters

The power output of piezo harvesters can be significantly increased by using charge modification techniques such as piezoelectric pre-biasing or synchronous switched harvesting, but in order to achieve a significant power gain high Q electrical resonant circuits must be used. For integrated systems with a significant size constraint, or in systems using on-chip inductors, achieving high Q can be difficult. Here we present an improved interface circuit which performs better than all previously presented techniques, especially in cases where the Q-factor of the resonant circuit is low. This has the added advantage of extending the useful life of the harvester because repeated cycling reduces the quality of the piezoelectric film, increasing the series resistance and lowering the electrical Q and the performance. Maximum power extraction by this new circuit is also less sensitive to the supply rail voltage than previously presented implementations, which reduces control power overhead.

Maximum Performance of Piezoelectric Energy Harvesters When Coupled to Interface Circuits

This paper presents a complete optimization of a piezoelectric vibration energy harvesting system, including a piezoelectric transducer, a power conditioning circuit with full semiconductor device models, a battery and passive components. To the authors awareness, this is the first time and all of these elements have been integrated into one optimization. The optimization is done within a framework, which models the combined mechanical and electrical elements of a complete piezoelectric vibration energy harvesting system. To realize the optimization, an optimal electrical damping is achieved using a single-supply pre-biasing circuit with a buck converter to charge the battery. The model is implemented in MATLAB and verified in SPICE. The results of the full system model are used to find the mechanical and electrical system parameters required to maximize the power output. The model, therefore, yields the upper bound of the output power and the system effectiveness of complete piezoelectric energy harvesting systems and, hence, provides both a benchmark for assessing the effectiveness of existing harvesters and a framework to design the optimized harvesters. It is also shown that the increased acceleration does not always result in increased power generation as a larger damping force is required, forcing a geometry change of the harvester to avoid exceeding the piezoelectric breakdown voltage. Similarly, increasing available volume may not result in the increased power generation because of the difficulty of resonating the beam at certain frequencies whilst utilizing the entire volume. A maximum system effectiveness of 48% is shown to be achievable at 100 Hz for a 3.38-cm3 generator.

Which is better, electrostatic or piezoelectric energy harvesting systems?

This paper answers the often asked, and until now inadequately answered, question of which MEMS compatible transducer type achieves the best power density in an energy harvesting system. This question is usually poorly answered because of the number of variables which must be taken into account and the multi-domain nature of the modelling and optimisation. The work here includes models of the mechanics, transducer and the power processing circuits (e.g. rectification and battery management) which in turn include detailed semiconductor models. It is shown that electrostatic harvesters perform better than piezoelectric harvesters at low accelerations, due to lower energy losses, and the reverse is generally true at high accelerations. At very high accelerations using MEMS-scale devices the dielectric breakdown limit in piezoelectric energy harvesters severely decreases their performance thus electrostatics are again preferred. Using the insights gained in this comparison, the optimal transduction mechanism can be chosen as a function of harvesting operating frequency, acceleration and device size.

Comparison between MEMS and Meso Scale Piezoelectric Energy Harvesters

Manufacture of piezoelectric energy harvesters typically assumes bulk piezoelectric material for the transducer until the reduction in size of the device prevents this. However when designing piezoelectric harvesters, the complete system must be taken into account including the transducer, power circuit, and battery, as these will impose restrictions on what can be achieved. Therefore a comparison between MEMS and meso-scale piezoelectric energy harvesting systems using a fully parametrised model is required. The comparison was restricted to a piezoelectric beam with a mass at the end connected to a single supply pre-biasing circuit to provide the optimal damping force and rectification. A buck converter was used to transfer extracted energy to a 1.5V battery. The results indicate that for devices with a volume side length less than 16.25 mm, no device using meso-scale properties can be made to resonant at 100 Hz or less due to the length and stiffness of the beam. Whereas above this limit, the voltage required to damp devices with MEMS scale properties causes a breakdown in the dielectric. We present a comparison of the theoretical limits of MEMS and meso-scale piezoelectric harvesters to provide design insight for future devices to maximise power generation.