Investigation of load dependent behavior of non-isolated DC-DC converters for thermoelectric generators

Thermoelectric generators (TEG) are semiconductor materials that convert the temperature difference in places where waste heat is found into electrical energy. Considering the fact that TEGs have low efficiency, it is of great importance to obtain maximum power from them. Maximum power point tracking (MPPT) algorithms are used to extract maximum power from TEGs. In addition, since the temperature difference between the surfaces of the TEGs is continuously variable, the voltage value produced varies. Converters are used with TEGs for voltage regulation and MPPT. MPPT algorithms do not provide adequate performance in every converter due to variable loads. In this study, simulation and experimental studies of the MPPT performances of DC-DC converters that are a boost, a buck and a buck-boost converters with perturb & observation (P&O) MPPT algorithm at variable loads were carried out. In applications and simulation studies, when variable loads are connected to the boost converter with P&O algorithm in the TEG system, the P&O algorithm carried out MPPT at the TEG internal resistance and the load resistance values above it. When a buck converter with P&O algorithm was connected to the TEG system and the load values were changed gradually, the P&O algorithm fulfilled MPPT at values equal to or below the TEG internal resistance. When a buck-boost converter with P&O algorithm and variable loads were connected to the system, the P&O algorithm executed good MPPT performance at every variable load. In the simulation studies, the MPPT error rate was at least 5.16% in the boost converter, 0.24% in the buck converter, and 0.45% in the buck–boost converter. The characteristics of the P&O algorithm of the three converters in the simulation studies in the load changes have been experimentally verified.