Neural network sliding mode control for a photovoltaic pumping system

Abstract EN

This paper presents a method for neural network sliding mode control design to track the maximum power point (MPPT) for a photovoltaic pumping system.

For the best use, the photovoltaic (PV) panel must operate at its maximum power point (MPP).

Sliding mode control (SMC) can be used for non linear systems with small uncertainties.

However, for complex nonlinear systems, the uncertainties are large and produce higher amplitude of chattering due to the higher switching gain.

In this work, sliding mode control approach is combined with the neural network (NN) to adjust the duty cycle control law.

NN is used for the prediction of model unknown parts.

The proposed control law uses the full state of the system.

However only the rotation speed variable is available for measurement.

For this particular task, a robust differentiator via SMC is employed.

Performance of the proposed controller is compared with the traditional SMC and investigated by simulation.