Maximum Power Point Tracking MPPT Control Based on Neural Networks for Photovoltaic Panels Implemented in Typhoon HIL

Due to the increase in energy consumption and the need to care for the environment, the use of clean energy has increased, being photovoltaic systems the most used. However, due to its low efficiency, the solar resource is not fully exploited, being necessary control strategies for monitoring the maximum power point (MPPT), in this context, this research presents the design, comparison and implementation of to two algorithms, one traditional and intelligent control of a Boost DC/DC converter of the photovoltaic panel implemented in a Hadware-in-the-loop (HIL) system. The MPPT control algorithms evaluated are one based on Perturb and Observe (P&O) and an intelligent algorithm based on neural networks (NN) multilayer perceptron with Back Propagation training, which takes as inputs the voltage V<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">pv</inf> and current I<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">pv</inf> at the panel and duty cycle output and a hidden layer of ten neurons with offline training.In the Typhoon HIL 402 system, the PV panel with its power converter is emulated in real-time while the controller is implemented on the digital signal processing (DSP) board. The two controllers are compared under irradiance and load changes, demonstrating that the NN algorithm performs better and faster control actions under cut-off and irradiance and load changes, this leads to 98% efficiency in the DC/DC Boost converter and solar resource utilization.