Abstract
The solar panel has Maximum Power Point (MPP) by which the output becomes the maximum. Since the MPP depends on insolation and panel temperature, it is never constant over time. So, Maximum Power Point Tracker (MPPT) is often used for solar systems.
Perturbation and Observation (P&O) method is widely used for MPPT because of its simple structure and relatively good performance. But in rapidly changing insolation conditions, the efficiency of P&O decreases.
To solve this, MPPT controller using Neural Network (NN) is proposed. Although MPPT controller using NN shows quick responses to rapidly changing insolation conditions, almost all of such models need to do the pre-learning using the solar panel specific data in advance.
We propose a novel MPPT controller that achieves online-learning and control. The approach of proposed method is a combination of NN and P&O method. The simulation results show that the proposed method is highly efficient in rapidly changing insolation conditions.
