The Design and Effect of Power Electronics on Vibration-Based Energy Harvesting Methods.

Abstract

Recent advancements in communication and low-power sensor nodes have led to innovative data acquisition systems for applications such as heart monitoring and forest-fire detection. Often these systems are in locations characterized by limited access to electrical power, yet they are in the presence of ambient mechanical vibrations. Therefore, energy harvesting from mechanical vibrations is proposed as a solution for powering these wireless sensor nodes. There are two devices that are commonly used for vibration-based energy harvesting: piezoelectric devices and electrostatic devices. This dissertation focuses on the power electronic interface between vibration energy harvesting devices and electrical energy storage elements.

By including power electronic efficiency as a parameter in the analysis of variable-capacitance energy harvesting, new fundamental properties of these devices are derived: a threshold efficiency necessary for energy harvesting, analytical solutions for optimal harvesting conditions, a comparison of energy harvesting methods at practical power electronic efficiencies, and a comparison of energy harvesting capabilities of various device architectures. Case studies are presented to illustrate practical applications of the theory presented in this work. One case study demonstrates the advantage of using the Charge Pump Method for MEMs applications, and illustrates the use of these new fundamental properties to aid power electronic architecture selection. Ultimately, the analysis-aided design produces more than twice as much power as previous implementations on the same device.

Recently, the dynamic active energy harvesting method has been proposed as a way to widen the bandwidth of resonant piezoelectric energy harvesters; however, the bandwidth extension is dependent on power electronic efficiency. In this dissertation a new energy harvesting system is proposed that includes a resonant inverter topology, in conjunction with new low-power analog control circuitry, in order to produce the first wideband autonomous dynamic active energy harvesting system. Experimental results using the Mide Volture V20w piezoelectric device shows that the harvested power is up to twice that of the adaptive rectifier method. These results include previously ignored loss mechanisms such as control losses, gating losses, and phase detection losses; making this system the first autonomous energy harvesting system of its kind.