Efficiency improvement and uncertainty quantification of stochastic resonance based tristable energy harvesting models

The multistability, including bistability and tristability, has been widely introduced into energy harvesting systems so that the effect of stochastic resonance can be applied to utilize the ambient vibration. Nevertheless, there are still issues about utilization of an additional magnet within the stochastic resonance effect. To resolve these issues, a typical tristable energy harvesting model driven by weak stochastic periodic excitation is explored by the moment method of derivative matching closure and the analysis of uncertainty quantification for the first time. With the stochastic resonance effect to optimize the stiffness and damping parameters, it is revealed that the tristable model can acquire a 12%–26% improvement in energy conversion efficiency compared to the deterministic model and demonstrates 8%–12% higher efficiency than the bistable energy harvester. It is also revealed that the noise intensity dominates the sensitivity of the output power, but the reciprocal time constant is the most influential factor for the energy conversion efficiency under stochastic resonance conditions. In particular, it is found that the importance ranking of parameters changes with noise levels and there is a significant difference between noisy and noise-free scenarios. These findings should provide a useful reference for relevant engineering designs.