Paper Abstract and Keywords |
Presentation |
2020-11-20 10:30
[Invited Talk]
Power Device Degradation Estimation by Machine Learning of Gate Waveforms Hiromu Yamasaki, Koutaro Miyazaki, Yang Lo, A. K. M. Mahfuzul Islam, Katsuhiro Hata, Takayasu Sakurai, Makoto Takamiya (Univ. of Tokyo) SDM2020-29 Link to ES Tech. Rep. Archives: SDM2020-29 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
A method to detect bonding wire lift-off of SiC MOSFETs using machine learning from the gate voltage waveform is proposed. In this paper, we proposed a new method that can be applied to 3-terminal SiC MOSFETs without the need for insulation compared to the conventional bonding wire lift-off detection method and demonstrated its effectiveness by SPICE simulation. By applying a linear regression algorithm to the two parameters extracted from the gate voltage waveform, we succeeded in constructing a bonding wire lift-off detection method that is robust to load current and temperature fluctuations. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
Linear regression / Machine learning / Power device / Reliability / Gate / / / |
Reference Info. |
IEICE Tech. Rep., vol. 120, no. 239, SDM2020-29, pp. 32-35, Nov. 2020. |
Paper # |
SDM2020-29 |
Date of Issue |
2020-11-12 (SDM) |
ISSN |
Online edition: ISSN 2432-6380 |
Copyright and reproduction |
All rights are reserved and no part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Notwithstanding, instructors are permitted to photocopy isolated articles for noncommercial classroom use without fee. (License No.: 10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
Download PDF |
SDM2020-29 Link to ES Tech. Rep. Archives: SDM2020-29 |
|