Conformalized Transfer Learning for Li-ion Battery State of Health Forecasting under Manufacturing and Usage Variability

TL;DR

This paper introduces a transfer learning framework that combines LSTM, domain adaptation, and conformal prediction to improve lithium-ion battery health forecasting across different manufacturing and usage conditions.

Contribution

It presents a novel uncertainty-aware transfer learning approach integrating MMD and conformal prediction for better generalization and reliable SOH forecasts.

Findings

Enhanced prediction accuracy across diverse battery conditions

Calibrated prediction intervals with reliable uncertainty quantification

Improved generalization over traditional models

Abstract

Accurate forecasting of state-of-health (SOH) is essential for ensuring safe and reliable operation of lithium-ion cells. However, existing models calibrated on laboratory tests at specific conditions often fail to generalize to new cells that differ due to small manufacturing variations or operate under different conditions. To address this challenge, an uncertainty-aware transfer learning framework is proposed, combining a Long Short-Term Memory (LSTM) model with domain adaptation via Maximum Mean Discrepancy (MMD) and uncertainty quantification through Conformal Prediction (CP). The LSTM model is trained on a virtual battery dataset designed to capture real-world variability in electrode manufacturing and operating conditions. MMD aligns latent feature distributions between simulated and target domains to mitigate domain shift, while CP provides calibrated, distribution-free prediction intervals. This framework improves both the generalization and trustworthiness of SOH forecasts across heterogeneous cells.