Design of an Artificial Neural Network Control Based on Levenberg-Marquart Algorithm for Grid-Connected Packed U-Cell Inverter

In this paper an Artificial Neural Network (ANN) controller based on Levenberg-Marquart algorithm is designed and applied to a five-level sensor-less Packed U-cell (PUC5) inverter operating as grid connected mode. The advantage of the proposed design is that there is no longer required to the conven...

Full description

Saved in:
Bibliographic Details
Main Authors: Babaie, Mohammad, Sebaaly, Fadia, Sharifzadeh, Mohammad, Kanaan, Hadi Y., Al-Haddad, Kamal
Format: Conference Proceeding
Language:English
Subjects:
Artificial Neural Network
Artificial neural networks
Capacitors
Grid Tied system
Inverters
Multilevel Converter
Neurons
Power Quality
Sensor-less PUC5 Inverter
Topology
Training
Voltage control
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In this paper an Artificial Neural Network (ANN) controller based on Levenberg-Marquart algorithm is designed and applied to a five-level sensor-less Packed U-cell (PUC5) inverter operating as grid connected mode. The advantage of the proposed design is that there is no longer required to the conventional parameters tuning snag confronted with the traditional Proportional Integral (PI) method. An offline algorithm takes in action of the controller parameters tuning calculation. The proposed ANN regulates the grid line current while the DC bus voltage is maintained constant with no need of regulator as the capacitor is self-voltage balanced in PUC5 topology. The overall design is simple as less tuning, sensors and regulators are requested compared to the conventional PI compensator case. Simulation results on MATLAB/Simulink are conducted to validate the performance of the controller in both steady and dynamic state where low Total Harmonic Distortion (THD) of the line current and high dynamic system response can be identified.
ISSN:2643-2978
DOI:10.1109/ICIT.2019.8755098