Power optimization of a piezoelectric-based energy harvesting cantilever beam using surrogate model

Energy harvesting is a conventional method to collect the dissipated energy of a system. In this paper, we investigate the optimal location of a piezoelectric element to harvest maximum power concerning different excitation frequencies of a vibrating cantilever beam. The cantilever beam oscillates by a concentrated sinusoidal tip force, and a piezoelectric patch is integrated on the beam to generate electrical energy. To this end, the system is modeled with analytical governing equations, then a Deep Neural Network (DNN)-based surrogate model is developed to appropriately model the system within the range of its first three natural frequencies. The surrogate model has significantly abated the computation cost. Thus, the optimization time is reduced drastically. Our investigations led to an optimal piezoelectric location for different excitation frequencies, which can result in maximum electrical output power. This location is highly dependent on the excitation frequency. When excitation frequency equals to natural frequencies, the maximum harvested power increases considerably.