MPPT design of centralized thermoelectric generation system under non-uniform temperature distribution based on COOT optimization algorithm

Due to the fact that thermoelectric generation (TEG) system commonly operates in environment where temperature distribution is non -uniform, consequently the output characteristic curve P-V for centralized TEG systems gives rise to several local maximum power points (LMPP). The paper suggests a COOT algorithm-based maximum power point tracking (MPPT) method for centralized TEG system, by imitating the collective behavior of Coot water birds, which involves their leadership structures and chain movements, to enhance its capability of jumping out the local optima. Furthermore, it integrates a dynamic reverse learning strategy and a dynamic termination condition, which shortens the iteration times and convergence time of the algorithm. To evaluate its performance, this paper conducted distinct case studies that included engine speed at 1000, 2500, and 4000 rpm, step changes and randomly changes. The effectiveness of the COOT algorithm is evaluated by comparing its performance with that of the traditional Perturbation and Observation (P&O) algorithm as well as the meta-heuristic algorithms Gray Wolf Optimization (GWO) and Particle Swarm Optimization (PSO). It is shown that the COOT algorithm can achieve optimum power output and effectively reduce fluctuations in output power, while converging to the optimal solution more quickly.