Performance analysis of frequency up-converting energy harvesters for human locomotion

Energy harvesting from human locomotion is a challenging problem because the low frequencies involved are incompatible with small, light-weight transducers. Furthermore, frequency variations during changing levels of activity greatly reduce the effectiveness of tuned resonant devices. This paper presents the performance analysis and parameter study of energy harvesters utilizing magnetic interactions for frequency up-conversion. Ferrous structures are used to periodically attract a magnetic tip mass during low-frequency oscillations, producing a series of impulses. This technique allows resonant structures to be designed for much higher natural frequencies and reduces the effects of excitation frequency variation. Measured vibrational data from several human activities are used to provide a time-varying, broadband input to the energy harvesting system and are recreated in a laboratory setting for experimental validation. Optimal load resistances are calculated under several assumptions including sinusoidal, white noise, and band-limited noise base excitations. These values are tested using simulations with real-world accelerations and compared to steady-state power optimization results. The optimal load is presented for each input signal, and an estimation of the maximum average power harvested under idealized conditions is given. The frequency up-conversion technique is compared to linear, resonant structures to determine the impact of the nonlinearities. Furthermore, an analysis is performed to study the discrepancies between the simulated results and the predicted performance derived from frequency response functions to determine the importance of transients.