Tapered Piezoelectric Energy Harvesters and Their Uncertainty Quantification

Abstract

Piezoelectric energy harvesting (PEH) has emerged as a strong candidate to provide a viable self-powering solution for the edge devices of the Internet of Things. This thesis is focused on a speci c category of such PEH devices, namely, those that use a tapered beam as their mechanical element. Such PEH devices exhibit greater uniformity in stress distribution, which increases their power harvesting potential while reducing their risk of failure due to mechanical fatigue. The primary contribution of this thesis is a novel semi-analytical model and its solution to the vibrational dynamics of a multi-layered tapered prismatic micro cantilever leading to a full electromechanical model for the tapered-beam PEH. The derivation of the model is based on a perturbation expansion of the tapered cantilever's partial di erential equations using Euler-Bernoulli beam theory as a starting point. To evaluate the impact of manufacturing tolerances on device performance, Monte-Carlo (MC) analysis has been used. Furthermore, probabilistic models for uncertainty propagation from manufacturing process parameters to device performance parameters have been implemented. In particular, a polynomial chaos expansion has been used to further quantify the impact of process uncertainties on the tapered beam PEH model. The proposed semi-analytical solution enables the modal analysis to run 200x faster compared with a computationally e cient commercial FEM solution. The model prediction error on resonance frequency is below 0.1% for 20 perturbation components and below 10% for 4 components. The run times of steady state and transient analysis show similar gains in speedup. In addition, the semianalytical model has been used to derive an equivalent electrical circuit of the device that can be readily simulated, along with the power conversion and signal conditioning circuitry, using commercial CAD environments for circuit design and analysis. Lastly, the semi-analytical model has been used to implement a MEMS design work ow, incorporating uncertainty quanti cation, for the design, analysis, and yield prediction of tapered beam piezoelectric energy harvesters. This work has been supported under the MEMS TwinLab Program in collaboration with GLOBALFOUNDRIES and the Institute of Microelectronics, Singapore.

Date of Award	May 2017
Original language	American English
Supervisor	Ibrahim Elfadel (Supervisor)