Modeling of partial shading in photovoltaic systems based on MLP artificial neural networks

摘要

In this work, a technique for modeling photovoltaic characteristics is proposed and this for operation in various environmental circumstances, including partially shaded conditions, which has been developed based on practical measures using an artificial neural network type of Multi Layer Perceptron to study the influence of partial shading on photovoltaic systems and therefore, make predictions and prevention in order to keep the system functional in an acceptable way. The results show that, when the area under shading is highly shaded, the short-circuit current and the open-circuit voltage of the photovoltaic panel are more influenced and the tracked maximum power decreases greatly. Thus, the lost power is dissipated in the shaded cell as heat. It is envisaged that the proposed technique is very useful for photovoltaic system who requires simple, fast and accurate PV model to make their systems.