Maximum Power Tracking for Centralized Temperature Difference Power Generation System Based on Elman Neural Network Combined With Improved Sparrow Search

Since the thermoelectric generation (TEG) sheets will be placed in places with different temperature gradients, it leads to multiple peaks in the duty-power (D-P) characteristic curve of a centralized TEG system under non-uniform temperature distribution (NTD). For this reason, this paper proposes an ENN-ISSA control algorithm, which combines the Elman neural network (ENN) with the sparrow search algorithm (SSA) by adding firefly perturbation. The ENN obtains the centralized TEG system’s single-input and single-output fitting curves, after which the firefly perturbation is introduced into the SSA algorithm. Then the improved SSA algorithm is used to realize the maximum power point tracking (MPPT) control based on the fitted curves. Based on building a centralized TEG system Simulink model and analyzing the output characteristics of the TEG module, temperature constancy experiments, temperature change experiments, and accuracy analysis were conducted. The results of these simulation experiments all show that the algorithm can track the global maximum power point (GMPP) quickly and accurately in the duty-power (D-P) curve with multiple peaks compared with the perturbation observation method and particle swarm algorithm.