Paper Title
Neural Network-Based DC to AC Converter: A Cascaded Inverter Replacement for PV System

Abstract
The use of Photovoltaic (PV) systems in generating electrical energy has increased in recent years. However, the most famous choice of inverter, the cascaded h-bridge (CHB) inverter, used in converting the DC power output of PV panels to AC power, consists of several semiconductor switches. These switches contribute to the switching losses of the system. Therefore, the focus of this research work is on designing a neural network (NN) based inverter with a minimal number of semiconductor switches as an alternative to CHB. The proposed system was realized by initially generating sinusoidal pulse width modulation (PWM) signals using mathematical equations. These signals serve as inputs to the proposed NN, which is based on feedforward/backpropagation. The output of the NN is combined with the previously generated sinusoidal PWM signals to produce pulses used to control the switches in the inverter circuit. Both the proposed system and the CHB inverterbased PV system were simulated using MATLAB/Simulink software for comparison. The results indicate that the total harmonic distortion (THD) content of the output waveform for the proposed and CHB inverter-based PV systems is 5.64% and 3.83%, respectively. Additionally, the switching losses for the proposed system and the CHB inverter-based PV system are 48.8 kW and 171.64 kW, respectively, resulting in corresponding output powers of 99.95 kW and 99.99 kW. Based on the results obtained from both systems, this study concludes that the proposed ANN based inverter can serve as a replacement for the CHB inverter, leading to a reduction in the power losses of the system. Keywords - Harmonic Distortion, Switching Losses, PV System, Inverter, MATLAB/Simulink