A method for the estimation of the battery pack state of charge based
on in-pack cells uniformity analysis
Liang Zhong, Chenbin Zhang, Yao He, Zonghai Chen
⇑
Department of Automation, University of Science and Technology of China, Hefei 230027, PR China
highlights
Build a method for battery pack SOC estimation.
Analyze the effect of the uneven cells problems to the pack SOC.
The SOC is estimated with consideration of different balance control strategies.
The UPF method is used to estimate the SOC to improve the accuracy.
article info
Article history:
Received 21 March 2013
Received in revised form 28 May 2013
Accepted 3 August 2013
Keywords:
State-of-charge
Inconsistency analysis
Unscented particle filter
Li-ion battery pack
Battery model
abstract
The state-of-charge (SOC) is a critical parameter of a Li-ion battery pack. Differences among in-pack cells
are inevitable and can change the total capacity of a pack and the remaining available capacity. Because
the traditiona l methods for the estimation of the SOC of a pack did not consider the difference among the
cells and the impact of balance control, we developed a new method that accounts for these problems. To
accurately estimate the pack SOC, we establish the relationship between the parameters of the pack and
those of in-pack cells under different balance control strategies. This paper also studies the two different
types of connections of a battery pack: in series and in parallel. Based on the model of the first over-
charged cell and that of the first over-discharged cell, the estimation of the SOC of a battery pack is real-
ized by the Unscented Particle Filter (UPF) algorithm. A simulation experiment verified the method for
the estimation of the SOC for a battery pack based on actual data and proved that an accurate estimation
value can be obtained by the method.
Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction
With the development of electric vehicles and smart grids, lith-
ium-ion batteries are becoming widely used as large-scale energy
storage systems. As we know, to meet the requirements of high-
energy and high-power applications, a battery system is usually
composed of hundreds or thousands of cells through series and
parallel electrical connections. An accurate estimation of the SOC
of a battery system will enable the protection of the battery pack
from being over-discharged or over-charged and thereby extend
the service life [1].
Many methods currently exist to estimate the SOC of cells or
battery packs in real-time, with the primary methods being the
current integral method [1], the neural network model method
[2], the fuzzy logic method [3] and the battery model-based meth-
od. The current integral method is simple to implement and is of-
ten used with correction by open circuit voltage. In the battery
model-based method, the battery model is first established, and
then an algorithm such as the Kalman Filter (KF) [4], Extended
Kalman Filter (EKF) [5], Unscented Kalman Filter (UKF) [6,7], Parti-
cle Filter (PF) [8] or Unscented Particle Filter (UPF) [9], is used to
estimate the SOC. These approaches have been widely studied in
many reports in the literature, and most of them have achieved
acceptable results.
However, the methods discussed above, usually do not take into
account the difference between each individual cell when calculat-
ing the SOC of the entire pack. The in-pack cell with the lowest
available capacity will determine the available capacity of the
entire pack, because that cell will be the first to be completely dis-
charged during the discharging of the pack. Similarly, the charging
of the pack will stop when the in-pack cell with the highest avail-
able capacity is full, even though the other cells are still not fully
charged. Hence, uneven cells problem will limit the total capacity
and remaining available capacity of the pack and thus affect the
SOC of the entire battery pack.
Differences among the in-pack cells are inevitable, mainly be-
cause of the production process used and the external environment
0306-2619/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.apenergy.2013.08.008
⇑
Corresponding author. Tel.: +86 551 63606104; fax: +86 551 63603244.
E-mail address: chenzh@ustc.edu.cn (Z. Chen).
Applied Energy 113 (2014) 558–564
Contents lists available at ScienceDirect
Applied Energy
journal homepage: www.elsevier.com/locate/apenergy