Remaining Useful Life Estimation of Lithium-Ion Batteries Based on Optimal Time Series Health Indicator

Zhonghua Yun,Wenhu Qin

doi:10.1109/access.2020.2981947

Zhonghua Yun, Wenhu Qin

Open Access

https://doi.org/10.1109/access.2020.2981947

Copy DOI

Abstract

The evaluation of lithium battery performance is a complex and very important issue. Generally, manufacturers perform battery burn-in tests and evaluate the performance of lithium batteries based on capacity, internal resistance, voltage, and other parameters in the cycle. However, due to the complexity of practical applications and the difficulty of parameter measurement, it is necessary to evaluate the status of health (SOH) of lithium-ion batteries from the side. Analysis of battery charge and discharge data found that using the charge and discharge time to evaluate the health of the battery is effective and feasible, especially the time during the discharge/charge platform period, and the parameter measurement is more convenient. In this paper, three time-health indicators are constructed and analyzed in detail, and then the health of the battery is evaluated using a simple Bayesian Monte Carlo theory. The experimental results of four batteries show that the scheme is simple and convenient, and can effectively evaluate the SOH of lithium-ion batteries.

Highlights

In recent years, non-renewable energy sources such as oil have gradually been exhausted, the research in electric vehicles and hybrid vehicles has been intensified around the world and it will be regarded as the important vehicle models in the future
The main advantages of lithium-ion batteries are high energy density, low self-discharge rate, long life, etc.it cannot discharge at a large current and has the disadvantage of poor safety, the deterioration monitoring of lithium-ion batteries and the prediction of remaining useful life are extremely important in practical applications
In the experiments of these two batteries, four voltage drops of 4V-3.8V, 3.9V-3.6V, 3.7V-3.3V and 3.63V-3.58V are selected respectively, the dotted line represents the end of life (EOL)

Summary

INTRODUCTION

Non-renewable energy sources such as oil have gradually been exhausted, the research in electric vehicles and hybrid vehicles has been intensified around the world and it will be regarded as the important vehicle models in the future. Studied the health status and remaining service life of lithium-ion batteries using the Dempster-Shafer theory and Bayesian Monte Carlo theory, respectively. Qin: Remaining Useful Life Estimation of Lithium-Ion Batteries Based on Optimal Time Series Health Indicator. The internal mechanism of the battery involves many complicated factors such as materials and chemical reactions, but it is not convenient to monitor these internal parameters in the application environment It requires a simple and unambiguous estimation of the remaining useful life and the state of health of lithium-ion batteries from the side. Experimental results show that using optimal time difference indicator (discharge/ charge/hybrid ) to estimate the state of health of the lithium-ion battery and the life prediction track of lithium-ion battery more accurately.

ANALYSIS OF THE EXPERIMENTAL DATASET

THE RESAMPLING

THE SOH PROGNOSTICS

CONCLUSION