Christou’s Group presents short course, research at CS ManTech

news story image

Yizhou Lu, a graduate student in Professor Aristos Christou’s research group, presented the short course ‘Application of Unsupervised Machine Learning Techniques in Prognostics of Power Electroncs’ at the 2018 International Conference on Compound Semiconductor Manufacturing Technology (CS ManTech), in Austin, TX.

Anomaly and fault detection is an integral part of prognostics and health management, and in many engineering problems that are not well understood, it is often difficult and costly to achieve fully labeled training datasets. Thus, unsupervised or semi-supervised machine learning a preferable choice under such circumstances.

The short course focused on unsupervised machine learning techniques commonly used in anomaly detection, such as principal component analysis (PCA) and k-means clustering, and the implementation of these techniques, in combination with statistical assessment of extracted principal components on time-series of recorded degradation data of IGBT modules and GaN HEMTs, to determine the probability of anomaly of test data. Additionally, implementing particle filter techniques coupled with anomaly detection techniques were presented.

Also at the event, Prof. Christou and his group presented the latest results on diamond power electronics in a presentation titled ‘Fabrication and Characterization of Diamond FETs with 2D Conducting Channels.”

This presentation discussed the fabrication and characterization of diamond field-effect transistor switches with 2-D carrier channels created by hydrogen termination of smooth diamond surfaces and by growth of epitaxial boron-doped delta layers.

Published May 11, 2018