Generalized regression neural network and fitness dependent optimization: Application to energy harvesting of centralized TEG systems

The thermoelectric generator (TEG) system has attracted extensive attention because of its applications in centralized solar heat utilization and recoverable heat energy. The operating efficiency of the TEG system is highly affected by operating conditions. In a series-parallel structure, due to diverse temperature differences, the TEG modules show non-linear performance. Due to the non-uniform temperature distribution (NUTD) condition, several maximum power points (MPPs) appear on the P/V curve. In multiple MPPs, the true global maximum power points (GMPP) are very important for optimum action. The existing conventional technologies have slow tracking speed, low productivity, and unwanted fluctuations in voltage curves. To overcome the TEG system behavior and shortcomings, A novel control technology for the TEG system is proposed, which utilizes the improved generalized regression neural network and fitness dependent optimization (GRNNFDO) to track the GMPP under dynamic operating conditions. Conventional TEG system control techniques are not likely to trace true GMPP. Our novel GRNNFDO can trace the true GMPP for NUTD and under varying temperature conditions In this article, some major contributions in the area of the TEG systems are investigated by solving the issues such as NUTD global maxima tracking, low efficiency of TEG module due to mismatch, and oscillations around optimum point. The results of GRNNFDO are compared with the Cuckoo-search algorithm (CSA), and grasshopper optimization (GHO) algorithm and particle swarm optimization (PSO) algorithm. Results of GRNNFDO are verified with experiments and authenticated with MATLAB/SIMULINK. The proposed GRNNFDO control technique generates up to 7% more energy than PSO and 60% fast-tracking than meta-heuristic algorithms.