Performance optimization of symmetrical multi-level boost converter using hybrid MPPT-ANN for solar energy applications

This paper introduces an artificial neural network (ANN) methodology for maximum power point tracking (MPPT) to regulate a symmetrical multilevel boost converter, aiming to enhance the output power of an autonomous photovoltaic system under both constant and variable solar radiation conditions. The paper's originality is in employing a design distinct from the traditional boost converter in autonomous photovoltaic systems and analyzing the system's efficiency through the non-conventional architecture. The autonomous photovoltaic system is developed using Matlab/Simulink software, emphasizing the performance of combining hybrid MPPT-ANN control and the characteristics of the symmetrical multilevel boost converter. The findings indicate that integrating the symmetrical multilevel converter with the suggested method reduces response times to 4.82 ms for settling time and 3.623 ms for rising time while minimizing voltage overshoot by 0.816%. Furthermore, the technique significantly diminishes voltage and current ripple to 0.006325 V and 0.00741 A, improving the generated energy quality. The proposed methodology is also evaluated in real time using a cost-effective implementation technique called the processor-in-the-loop technique. The Arduino Mega is employed to implement this strategy. The implementation results corroborate the simulation results. Ultimately, a comparison with other studies demonstrates that the proposed methodology improves the efficiency of the autonomous photovoltaic system by 99.99 % and decreases both the rise and settling times, as well as the voltage overshoot, in contrast to systems utilizing traditional boost converters or flyback converters.