Implementation of the RBF neural network on a SOPC for maximum power point tracking

In this paper, a radial basis function (RBF) neural network is implemented as a system on a programmable chip (SOPC) to carry out maximum power point tracking (MPPT) for photovoltaic (PV) control systems. The implementation of the SOPC can provide a traditional proportional integral derivative (PID) controller and some additional hardware like a pulse width modulation (PWM) generator by a general purpose field programmable gate array (FPGA) chip, as well as integrate a RBF neural network controller by embedded soft processors. The tracking algorithm changes the duty-cycle of the IGBTs to make the PV converter work at maximum power. As a result, the MPPT unit of the PV system becomes an independent and highly integrated product including peripheral design and a control algorithm.