A Fast and Universal Neuro-Based SVM Algorithm for Multi-Level Converters

Author

Saeedifard, M. ; Rad, H. Saligheh ; Bakhshai, A. ; Iravani, R.

Author_Institution

Center for Applied Power Electronics (CAPE), Department of Electrical and Computer Engineering, University of Toronto, 10 King´´s College Road, Toronto, Ontario M5S 3G4, Canada. email: mary.saeedifard@utoronto.ca

fYear

2007

fDate

Feb. 25 2007-March 1 2007

Firstpage

1508

Lastpage

1514

Abstract

This paper proposes a novel, simple and fast classification algorithm for implementation of Space Vector Modulation (SVM) method for a Diode Clamped Multi-level Converter (DCMC) with an arbitrary number of levels. The proposed algorithm is based on a classifier neural network which provides a straightforward and computationally efficient approach without the use of trigonometric calculations or look-up tables to identify (i) the location of reference voltage vector, (ii) its adjacent switching voltage vectors, and (iii) their corresponding on-duration time intervals. The feasibility of the proposed SVM algorithm is validated based on theoretical analysis, simulation studies and experimental tests on a DSP-controlled, 5 kVA, three-level converter system.

Keywords

Algorithm design and analysis; Analytical models; Classification algorithms; Computational modeling; Computer networks; Diodes; Neural networks; Support vector machine classification; Support vector machines; Voltage; Classification Technique; Competitive Neural Network; Multi-Level Converters; Space Vector Modulation;

fLanguage

English

Publisher

ieee

Conference_Titel

Applied Power Electronics Conference, APEC 2007 - Twenty Second Annual IEEE

Conference_Location

Anaheim, CA, USA

ISSN

1048-2334

Print_ISBN

1-4244-0713-3

Type

conf

DOI

10.1109/APEX.2007.357717

Filename

4195920