Synchronous switched circuits for enhanced energy extraction from triboelectric kinetic energy harvesters

Abstract

Internet of Things (IoT) sensors are finding increasing usage in today's ‘smart’ City/Building/Factory/Agriculture/Healthcare settings. A challenge with remote-monitoring IoT sensors lies in their battery-powered operation that requires human intervention to periodically replace batteries, increasing maintenance cost. Harvesting ubiquitous background kinetic energy is an attractive green solution to self-power the IoTs. In this context, Triboelectric Nanogenerators (TENGs), a class of ambient kinetic energy harvesters with their material and design versatility, have emerged as a promising cost-effective solution to self-power the IoTs in the past decade.

This research fundamentally investigates the TENG’s cyclic transduction operation and proposes novel synchronous-switched Energy Extraction Circuits (EECs) that \emph{actively} condition the TENG charge (equivalently voltage) in sync with TENG operation to enhance the TENG's per-cycle transduced energy. The overall objective is developing an EEC interfacing the AC TENG source with the IoT's on-board DC storage (battery/ capacitor) that (a) provides rectification, (b) maximizes TENG's energy output, and (c) maintains the energy optimality even as the external mechanical input condition (accordingly the TENG parameters) changes. In that context, this dissertation reports the design, simulation, mathematical modeling, optimization, and comparison of the following seven circuit architectures as TENG EEC candidates, • Standard Full Wave Rectifier circuit (FWR) (reference circuit for comparison) • Synchronous Switched Harvesting on Inductor (SSHI) class of circuits: – Parallel-SSHI – Series-SSHI • Synchronous Charge Extraction (SCE) class of circuits: – Standard SCE – Pre-biased SCE (pSCE) – Self-propelled Pre-biased SCE (spSCE) – Multi-Shot SCE (MS-SCE). Barring the Standard FWR and Standard SCE, the rest of the architectures have been proposed as TENG EEC for the first time in this work.

In the first set of works, detailed analytical circuit models for FWR, P-SSHI, and the SSHI circuits are developed to derive their per-cycle energy output, optimal load (as in MPP: Maximum Power Point), and upper bound on load. The analytical results are validated against the simulated and those measured by testing using a custom-made experimental TENG. SSHI circuits showed up to 8.5x experimental gain over the Standard FWR; however, their optimal load operation over varying ambient input conditions requires a second-stage DC/DC converter with closed-loop MPP Tracker, increasing the overhead power consumption and, thus, reducing the net output.

Next, we study the SCE circuit with the desirable trait of providing a constant level of transduced energy, regardless of the load voltage, i.e., MPPT-free operation and also one that exceeds the reference FWR circuit's optimal energy output. Building on the SCE circuit architecture, we propose a pre-biased SCE (pSCE) circuit that is analytically shown to provide a further boost beyond the SCE output through its synchronous pre-biasing (pre-charging) technique. Measured results with the discrete component implementation of pSCE on PCB provide validation with its up to 2.1x gain over the SCE circuit. A drawback with pSCE lies in its complex control that requires TENG-specific manual tuning. The next proposed architecture, Self-propelled pre-biased SCE (spSCE), addressed this challenge by offering plug-and-play self-tuned pSCE action, i.e., applicable for any TENG and under any operating ambient input condition, thus, meeting the set EEC design goals.

To achieve compact and low-loss EEC, we move from discrete component to on-chip design with our novel on-chip compatible Multi-shot SCE (MS-SCE) architecture that extracts energy from TENG in multiple short bursts to maintain operating voltage just below the technology’s breakdown limit. The MS-SCE is mathematically shown to be MPPT-free and universally-optimal, as in extracting energy higher than the optimal FWR for any given TENG under any operating conditions. The implementation and simulation in TSMC 0.18μm 70 V BCD process provide validation, showing >1.91× net gain over the optimal FWR.