|
Activity Number:
|
148
- Design and Analysis of Experiments
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 8, 2022 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Physical and Engineering Sciences
|
|
Abstract #323431
|
|
|
Title:
|
Degradation Analysis with Longitudinal Functional Data
|
|
Author(s):
|
Quyen N Do* and Pang Du and Yili Hong
|
|
Companies:
|
Virginia Tech and Virginia Tech and Virginia Tech
|
|
Keywords:
|
reliability analysis;
degradation analysis;
functional data analysis;
functional data;
longitudinal data;
renewable energy
|
|
Abstract:
|
Advancement in measurement technology has allowed data to be collected continuously in the form of functional data. The number of these functional curves can also amass temporally. The change of these curves through time can reveal patterns of degradation. This presents a new challenge and opportunity to expand the methodology of degradation analysis to functional data. In this presentation, we introduce an approach to analyzing the degradation of longitudinal functional data via a two-step procedure: prediction of shape and domain end point of functional data collected longitudinally using functional data regression models, and linking the predictions to degradation analysis. This method is advantageous to addressing simultaneously two modes of degradation in domain end point and temporal movement of the curve data. We present an application of this method to a motivating dataset studying the degradation of rechargeable lithium-ion batteries going through repeated cycles of discharge where voltage levels are collected throughout.
|
Authors who are presenting talks have a * after their name.
