Improved Maximum Power Control of Photovoltaic Systems Using Quadratic Boost Converter Based on Deep Recurrent Neural Network

Both temperature and solar radiation cause variations in photovoltaic output power. Nev-ertheless, each variation has a maximum power point, which represents the photovoltaic output's maximum efficiency. Photovoltaic power must be managed at the highest point in order to achieve optimal efficiency. This can be accomplished by employing a converter to regulate the photovoltaic output voltage at the maximum power. This study proposes a quadratic boost converter (QBC) to control photovoltaic output power by using the Deep Recurrent Neural Network (DRNN) algorithm. The goal of DRNN is to decrease ripple at the maximum point and speed up time to reach the maximum power point. QBC is de-signed to obtain a higher DC output voltage than a regular boost converter, so it can elim-inate the use of a step-up transformer if the photovoltaic is connected to an inverter. The proposed method is applied to a 50 Wp solar panel with an Arduino microcontroller as the controller device. The experimental results demonstrate that the DRNN algo-rithm-based QBC has effectively controlled the solar panel output power at the maximum point with a smoother ripple and a faster response. QBC has also been able to produce higher voltage output according to its characteristics.