Power quality improvement in solar photovoltaic system to reduce harmonic distortions using intelligent techniques

This paper presents the harmonic elimination in a solar fed cascaded fifteen level inverter supplied with varying input sources from solar Photo Voltaic's (PV) using the modern intelligent techniques such as Artificial Neural Networks (ANNs), Genetic Algorithm (GA) and Particle Swarm Optimization (PSO). In ANN based approach the harmonic equations are solved and switching angles are obtained such that the fundamental is kept constant and lower order harmonics are minimized or eliminated. The data set with varying input voltages and switching angles are trained with ANN. The trained network is integrated with solar PV system to reduce the harmonic distortions. Similarly the objective function for the proposed methodology is formulated and corresponding operations are performed with respect to optimization techniques (GA and PSO) to obtain the optimal switching angles required for the inverter switches. The modeling of solar panel is developed which serves as the input in simulation. The results show that ANN based method provide a significant improvement of power quality with respect to the reduction of Total Harmonic Distortion when compared to the other techniques. A 3 kWp solar PV plant with a fifteen level inverter incorporated with ANN based technique is implemented in hardware to show the effectiveness of the proposed method.