Online Capacity Estimation of Lithium-Ion Batteries Based on Deep Convolutional Time Memory Network and Partial Charging Profiles

Data-driven methods have been widely employed for capacity estimation of lithium-ion batteries through exploiting machine learning models to build a mapping relationship between extracted health features and capacity. However, existing machine learning based approaches require plentiful and intricate data processing for feature extraction. To remedy this limitation, this paper presents a deep learning method for online capacity estimation of lithium-ion batteries. A predictive model, namely deep convolutional time memory network with the properties of automatic feature abstraction and target estimation, is established via fusing the convolutional neural network and long short-term memory unit. Partial charging voltage and current data are selected from the raw charging data and can be directly fed into the proposed model without complicated pre-processing, resulting in easier training preparation and lower computational intensity. Experimental results demonstrate that the proposed algorithm can precisely estimate the battery capacity with the absolute error of less than 0.021 Ampere-hour and 0.11 Ampere-hour for two types of batteries. The proposed method requires only short charging voltage and current profiles and pledges high estimation accuracy, contributing to fast online capacity estimation.

Automated summary

This paper proposes a deep learning method for online capacity estimation of lithium-ion batteries. The proposed algorithm utilizes a predictive model that combines a convolutional neural network and a long short-term memory unit for automatic feature extraction and target estimation. Experimental results demonstrate that this method can accurately estimate battery capacity using only short voltage and current data.