Maximizing onboard power generation of large-scale railway vibration energy harvesters with intricate vehicle-harvester-circuit coupling relationships

Abstract

Large-scale vibration energy harvesters (VEHs) have the potential to produce power of tens of watts and offer a distributed and flexible power supply for onboard devices in unpowered freight wagons. However, research on railway vibration energy harvesting systems (VEHSs) is often limited to individual points and lacks systematic exploration and optimization. This paper proposes a systematic modeling approach for VEHS that considers the intricate interaction and coupling in the vehicle-harvester-circuit system. Firstly, a model is established for a rotary electromagnetic VEH using the equivalent circuit method, with mechanical friction considered and identified via the Equilibrium Optimizer (EO) to improve prediction accuracy. The energy harvesting circuit (EHC) incorporating a bridge rectifier, a DC/DC converter, and a power management module with a speed-driven maximum power point tracking (MPPT) algorithm is designed for efficient energy extraction and storage under stochastic vehicle suspension vibrations. In addition, the freight wagon is modeled spatially based on railway vehicle-track dynamics, accounting for the nonlinearities of primary and secondary suspensions to obtain more accurate vibration response and mechanical interaction with the harvester-circuit module for the coupling of the whole system. Finally, a performance-enhanced control strategy is proposed with the dynamic tuning of voltage coefficient to maximize onboard harvestable energy based on the developed system model. The results indicate that the harvester power can be increased by up to 60%, and the force decreased by up to 11% at various vehicle speeds and loads. The prototype of the whole railway VEHS with MPPT closed-loop control is implemented in the embedded environment, and its engineering-oriented design will significantly improve the system robustness and practicality in onboard environments.