Application of Artificial Neural Networks for Maximum Power Point Tracking of Photovoltaic Panels

Abstract

Maximum power point tracking technique for PV panels with support of online learning artificial neural network is offered. Mathematical model of the system is implemented in Matlab/Simulink environment. Maximum power point tracking is performed using IncCond algorithm and radial basis function artificial neural network. Several criteria for estimation of system performance were derived. It is shown that ANN can increase overall system efficiency by 10%.

Maximum Power Point Tracker for Photovoltaic Systems using On-line Learning Neural Networks

“…An on-line ANN controller, in conjuncion with the wellknown P&O technique [7], and using the back-propagation algorithm in order to minimize the controlled error , has been utilized for online estimation of the reference voltage in the feed-forward loop. The simulated results using the on-line training of the ANN show that the efficiency of the MPPT and the effectiveness of this control method is higher than the classical one.…”

“…Here, however, similar training samples are used by the artificial neural network (ANN). To deal with these training samples, we have used an MLP [7]- [8] with the help of Matlab neural networks toolbox in order to ensure fast and correct learning. The main idea is that the neural network learns each sample online because it is difficult to store all learning samples in small devices.…”

“…During normal operation, the solar irradfiance is considered to be 900W/m 2 , and the ambient temperature is 47°C [7]. The total simulation time is taken 0.1s [8].…”

“…The figures above mentionned show that the MPP control tracks the charging and dischanging of the back-up battery quickly when the shading of the PV panel changes considerably [6] and quickly, and has high control precision and stability in charging and discharging from the state of charge of the battery (Figure 14 (j1)). According to the noticed difference in the simulation results obtained by [ [1]- [6]- [7]- [8]- [10] and [11], that the output PV panel (Fig. 14(c1)) and the bus converter output power (Fig.…”

“…The algorithms commonly used to track the MPPT are the P&O method has been widely used because it is easy to implement. There are many techniques to adapt the load impedance value in order to obtain the optimal solar panel output voltage [6]- [7] and [8]. A general diagram is presented in figure 3 which consists of an MPPT computing system and a DC/DC converter.…”