A Systematic Review on Latest Features of Neural Network Designs for Power Electronic Systems Using Impedance Modeling

Abstract

For the purpose of this research, the advanced Neural Network (NN) development for the power electronic system’s is discussed with special emphasis on the impedance modeling. Power electronic systems are essential for all today’s electrical networks as a means of energy conversion and management. Recently, used to describe the impedance of these systems, NNs are capable of capturing various relationships between converters, inverters, and related components. That kind of modeling not only increases the system efficiency through the mastery of the energy conversion processes but also allows for controlling the processes that change in response to the changes in the other parameters, thus stabilizing the system’s performance. Also, NNs enable early fault diagnosis and prediction through the analysis of the impedance diagrams, which helps to increase the level of reliability in a system and avoid unnecessary faults. Thus, this paper is concluded by mentioning the current limitations encountered regarding the applicability of NN models for large-scale systems and discussing possible research directions that can enhance model performability, readability, and coexistence with the conventional control strategies for hybrid systems.