Real Time GA and ANN Based Selective Harmonic Elimination in 9 Level Ups Inverter

High quality power is very much critical and essential for medical, research and industrial applications to bring good quality results with accurate evaluation. Hence all the sensitive equipments and critical loads need to be provided with high quality and reliable power, where Uninterruptible Power Supplies (UPS) are mainly used to supply reliable power to these loads. Inverter is the main component of a UPS. In the recent times, by the advanced usage of semiconductor devices and non linear loads, harmonics are unavoidable. So the actual challenge for UPS is, under a non linear condition of load, it has to maintain a high quality sinusoidal output voltage. In this paper, the inverter of UPS is replaced by a nine level cascaded H bridge multilevel inverter with equal DC sources and harmonics can be eliminated by the optimal selection of switching angle by using Selective Harmonic Elimination Pulse Width Modulation (SHE-PWM) technique along with a hybrid technique to optimize and minimize the Total Harmonic Distortion(THD). The proposed hybrid technique utilizes the Genetic Algorithm (GA) and Neural Network (NN). The switching angles are calculated offline using Genetic Algorithm. Then the NN is trained by these switching angles and the real time switching angles are found out by Neural Network. The proposed technique is tested over a nine level cascaded H-bridge inverter and the resultant fundamental and harmonic voltages are analysed. Then, simulation is carried out in Matlab/Simulink environment and the results indicate that the switching angles obtained using this method results in efficient harmonic minimization.