ANFIS-Based MPPT Controller of the Thermoelectric Energy Harvesting System for DC Micro-grid Applications

Abstract

The power-current characteristics of a thermoelectric generator are nonlinear, which relies on the hot and cold side temperature difference. It is, therefore, vital that the thermoelectric modules (TEMs) are made to work at maximum power point (MPP). For continuous tracking of the MPP, a maximum power point tracking (MPPT) techniques may be used. In this research work, for thermoelectric energy harvesting systems (TEEHSs), an adaptive neuro-fuzzy inference system (ANFIS)-based MPPT controller is proposed to optimize the output power from the TEM array. The TEMs are connected in series–parallel configuration and that are placed near to the stator windings (SWs) of the wind generator (WG) for generating electrical energy from waste heat energy. The TEM array with ANFIS-based MPPT controller is compared to the fuzzy logic (FL)-based MPPT controller under a uniform wind velocity (UWV) and non-uniform wind velocity (NUWV) condition. From the obtained results, it is observed that the MPPT controller based on ANFIS tracks the maximum power extremely faster than the FL controller. The time taken by the ANFIS controller to reach the MPP is decreased by 0.15 s compared to the FL MPPT controller. The average maximum power obtained from ANFIS-based MPPT controller is 109 W and 119 W higher than the FL-based MPPT controller under UWV and NUWV conditions, respectively. In comparison, the tracking accuracy of the ANFIS-based MPPT controller is higher than the FL-based MPPT controller with an average relative error of 1.696% and 1.171% under both conditions.