USP Electronic Research Repository

Fault Diagnosis using Shallow Neural Networks for Voltage Source Inverters in SynRM Drives

Riccio, Jacopo and Kumar, Rahul R. and Cirrincione, Giansalvo and Zanchetta, Pericle and Cirrincione, Maurizio (2022) Fault Diagnosis using Shallow Neural Networks for Voltage Source Inverters in SynRM Drives. [Conference Proceedings]

[thumbnail of Fault Diagnosis using Shallow Neural Networks for Voltage Source Inverters in SynRM Drives.pdf] PDF - Published Version
Restricted to Repository staff only

Download (793kB) | Request a copy

Abstract

This paper presents a neural-based fault diagnosis system for a two-level Voltage Source Inverter that is used to drive the Synchronous Reluctance Motors. In particular, three classes are considered: Healthy, Open Circuit Fault (OSF) and Short Circuit Fault (SCF). The proposed strategy relies on the data generated by the mathematical models of OSF and SCF together with the healthy configuration. For each category of fault, multi-switch faults have been emulated. Following the acquisition of healthy and faulty three-phase currents, exploratory analysis of the data is conducted using Principal Component Analysis. Thereafter, various Machine Learning techniques have been utilized to develop different types of classifiers. The best classification model after considering factors like time complexity, storage, confidence level of outputs was the shallow Long-Short-Term-Memory neural network.

Item Type: Conference Proceedings
Subjects: T Technology > TK Electrical engineering. Electronics Nuclear engineering
T Technology > TL Motor vehicles. Aeronautics. Astronautics
Divisions: School of Information Technology, Engineering, Mathematics and Physics (STEMP)
Depositing User: Rahul Kumar
Date Deposited: 04 Feb 2024 21:42
Last Modified: 04 Feb 2024 21:42
URI: https://repository.usp.ac.fj/id/eprint/14016

Actions (login required)

View Item View Item