Implementation and comparison of metaheuristically modified ANN MPPT controllers under varying solar irradiance conditions.

The maximum power point tracking (MPPT) mechanism is one of the important components of a solar photovoltaic system (SPS). For efficient working of the SPS, the MPPTs must be flexible to adapt to any change in irradiance level, as the output from the Solar PV system majorly depends on the level of irradiance present at a geographical location. This study deals with the scope of improvement in artificial neural network (ANN) by incorporating the metaheuristic tuning. The modified version of ANN MPPT controller has been developed with the help of a novel Horse herd optimization (HHO) algorithm that strategically optimizes the weights of the neural network. The systematic implementation and comparison of different versions of ANN MPPT controller have been performed, taking into account different scenarios like dynamically changing irradiance levels and several cases of partial shading condition. For the training of the neural network, the irradiance and temperature dataset has been collected from the NASA website. The analysis of the results has been done using parameters like fill factor, maximum power tracked, percentage mismatch power loss and efficiency, which shows that the ANN MPPT controller tuned with HHO algorithm tracks maximum power and attains maximum efficiency under various considered scenarios. Further, the EN50530 MPPT efficiency test has been performed under the fast- and slow-changing irradiance level to compare the efficacy of different algorithms. [ABSTRACT FROM AUTHOR]