BatteryLife: A Comprehensive Dataset and Benchmark for Battery Life Prediction

TL;DR

BatteryLife tackles the data scarcity, limited diversity, and benchmarking inconsistency in Battery Life Prediction by integrating 16 datasets into 90,000 samples across Li-ion, Zn-ion, Na-ion, and CALB, and by introducing a unified benchmark and CyclePatch, a plug-in token-based framework that models within-cycle and cross-cycle dynamics using an intra-cycle and an inter-cycle encoder. The study evaluates 18 methods and demonstrates that many time-series techniques underperform on BLP, while CyclePatch consistently improves performance and enables scalable modeling across diverse aging domains. The results highlight substantial domain gaps in unseen aging conditions and partial cross-domain transferability, underscoring the need for domain-aware architectures and transfer strategies. Overall, BatteryLife provides a public, diverse resource and a principled benchmark that can accelerate development of accurate, generalizable battery life prediction models with practical impact on battery management and production optimization.

Abstract

Battery Life Prediction (BLP), which relies on time series data produced by battery degradation tests, is crucial for battery utilization, optimization, and production. Despite impressive advancements, this research area faces three key challenges. Firstly, the limited size of existing datasets impedes insights into modern battery life data. Secondly, most datasets are restricted to small-capacity lithium-ion batteries tested under a narrow range of diversity in labs, raising concerns about the generalizability of findings. Thirdly, inconsistent and limited benchmarks across studies obscure the effectiveness of baselines and leave it unclear if models popular in other time series fields are effective for BLP. To address these challenges, we propose BatteryLife, a comprehensive dataset and benchmark for BLP. BatteryLife integrates 16 datasets, offering a 2.5 times sample size compared to the previous largest dataset, and provides the most diverse battery life resource with batteries from 8 formats, 59 chemical systems, 9 operating temperatures, and 421 charge/discharge protocols, including both laboratory and industrial tests. Notably, BatteryLife is the first to release battery life datasets of zinc-ion batteries, sodium-ion batteries, and industry-tested large-capacity lithium-ion batteries. With the comprehensive dataset, we revisit the effectiveness of baselines popular in this and other time series fields. Furthermore, we propose CyclePatch, a plug-in technique that can be employed in various neural networks. Extensive benchmarking of 18 methods reveals that models popular in other time series fields can be unsuitable for BLP, and CyclePatch consistently improves model performance establishing state-of-the-art benchmarks. Moreover, BatteryLife evaluates model performance across aging conditions and domains. BatteryLife is available at https://github.com/Ruifeng-Tan/BatteryLife.

Figures (5)

• Figure 1: Partial cycling data (voltage and current time series) from three batteries with the same specifications tested under a protocol.
• Figure 2: Partial cycling data in degradation tests from (a) a lab-tested Li-ion battery, (b) a lab-tested Zn-ion battery, (c) a lab-tested Na-ion battery, and (d) an industry-tested Li-ion battery. The positive and negative values in current represent charging and discharging, respectively. The gray dotted lines are separators of different cycles. The charging and discharging periods are labeled in the first cycle as illustrations.
• Figure 3: SOH degradation trajectories in BatteryLife.
• Figure 4: Overview of the CyclePatch framework.
• Figure 5: The performance of the top-three models with an increasing number of usable cycles.