Surrogacy-Based Maximization of Output Power of a Low-Voltage Vibration Energy Harvesting Device

Abstract

The coreless microgenerators implemented in electromagnetic vibration energy harvesting devices usually suffer from power deficiency. This can be noticeably improved by optimizing the distribution of separate turns within the armature winding. The purposeful optimization routine developed in this work is based on numerical identification of the turns that contribute most to the electromotive force and the elimination of those with the least contribution in order to reduce the internal impedance of the winding. The associated mixed integer nonlinear programming problem is solved comparatively using three approaches employing surrogate models based on kriging. The results show very good performance of the strategy based on a sequentially refined kriging in terms of the ability to accurately localize extremum and reduction of the algorithm execution time. As a result of optimization, the output power of the system increased by some 300 percent with respect to the initial configuration.

Keywords

• topological optimization
• mixed integer nonlinear programming
• surrogate models
• energy harvesting