MDS-Rely Spring 2026 Brochure - Flipbook - Page 7
Enhancing Battery
Degradation Analysis and
Thermal Runaway Prediction
This project focuses on thermal runaway (TR), a
major safety risk in lithium-ion ba琀琀eries where
self-hea琀椀ng can lead to fire, explosion, or gas
release. While rela琀椀vely rare in single cells, the
risk increases significantly in systems with many
cells, making TR a cri琀椀cal issue in modern
electronics and vehicles.
Dr. Iyengar and his team are collec琀椀ng data from
experiments, incident reports, and related
phenomena to iden琀椀fy early indicators of TR.
They are applying autoregressive integrated
moving average (ARIMA) models for sta琀椀onary
voltage and temperature data, and nonlinear
approaches such as splines and neural networks
for highly nonsta琀椀onary periods. A known PDE
model for temperature is incorporated with noise
to explore TR behavior in a stochas琀椀c se琀�ng.
This work aims to improve predic琀椀on accuracy,
generate simula琀椀on data, and provide insights
into the mechanisms underlying thermal
runaway.
Dr. Sa琀椀sh Iyengar is a Professor
and Associate Chair within the
Department of Sta琀椀s琀椀cs at the
University of Pi琀琀sburgh. He
completed his PhD in Sta琀椀s琀椀cs at
Stanford University a昀琀er receiving
his BA in Mathema琀椀cs from
Harvard. His research explores
stochas琀椀c models, mul琀椀variate
analysis,
meta-analysis
and
applica琀椀ons in neuroscience. His
publica琀椀ons have explored the
theory behind analysis methods
such as the snowflake plot and
modeling neural ac琀椀vity using
inverse Gaussian distribu琀椀on.
6
