Indico is back online after maintenance on Tuesday, April 30, 2024.
Please visit Jefferson Lab Event Policies and Guidance before planning your next event:

Oct 16 – 21, 2022
Marriott Newport News at City Center
US/Eastern timezone

Multi-Module Based VAE to Predict HVCM Faults in the SNS Accelerator

Oct 20, 2022, 4:35 PM
Oral New & Innovative Accelerator Technology New & Innovative Accelerator Technology


ALANAZI, Yasir (Old Dominion University)


The anomalies in the High Voltage Converter Modulator (HVCM) remain a major down time for the Spallation Neutron Source (SNS) facility. To improve the reliability of the HVCMs, several studies using machine learning techniques were to predict faults ahead of time in the SNS accelerator using a single modulator. In this study, we present a multi-module framework based on Conditional Variational Autoencoder (CVAE) to detect anomalies in the power signals coming from multiple HVCMs that vary in design specifications and operating conditions. By conditioning the VAE according to the given modulator system, the model can capture different representations of the normal waveforms for multiple systems. Our experiments with the SNS experimental data show that the trained model generalizes well to detecting several fault types for several systems, which can be valuable to improve the HVCM reliability and SNS as a result.

Primary authors

ALANAZI, Yasir (Old Dominion University) Dr SCHRAM, Malachi (Thomas Jefferson National Accelerator Facility) RAJPUT, Kishansingh (Thomas Jefferson National Accelerator Facility) Dr PAPPAS, Chris (Oak Ridge National Laboratory ) Dr RADAIDEH, Majdi (Oak Ridge National Laboratory) Dr LU, Dan (Oak Ridge National Laboratory)

Presentation materials