Switching algorithms for extending battery life in Electric Vehicles
Highlights
► A battery's life is greatly affected by the method of use. ► Discharging in non-optimal currents has negative effects on the battery's life. ► We propose advanced switching algorithms to minimize these effects. ► The battery's life can be significantly extended by the proposed algorithms.
Introduction
Electric Vehicles (EVs) are the next generation of cars in the world of automobiles. The propulsion solutions for EVs are based on hybrid or fully battery powered electric vehicles [1]. The critical part of EVs, that determine their performance, is the battery [2]. In this work we focus on one of its problems, namely the battery's life.
The life of operating batteries is usually measured as the number of cycles, that can be obtained, until the capacity per cycle declines beyond a pre-determined threshold [3]. The energy extracted from the battery during full discharge is the integration of voltage as a function of capacity throughout the discharge process, until its cut off point (measured in Watt-hours). However, an alternative definition, which we use throughout this paper, can be the total accumulated energy extracted from a battery during its life.
Batteries are very expensive [4] and hence it is critically important to prolong their life as much as possible. The method of operating the EV battery may have a dominant effect on its life. According to results presented in Ref. [5] operating it incorrectly might reduce the battery's life to even a third of its expected duration. Thus, it is logical that the Battery Management System (BMS) in the EV includes a component that will control the battery's operation in order to extend its life.
An EV battery is actually a pack of strings of cells connected in a series, in order to reach the required high voltage. These strings of cells, i.e. cell-series, are connected in parallel, in order to provide the required current. For safety reasons, e.g. issues related to heat dissipation during operation, it is important to construct such large batteries from small individual cells. Thus, the focus in this paper is on the level of the individual cell-series in the battery, which determine the total voltage of the battery.
Developments in the Lithium-ion (Li-ion) battery technology in the last decade have made Li-ion batteries the standard choice of power sources for EVs [6], [7]. Thus, in this paper we focus on Li-ion batteries. Based on the chemistry and engineering of Li-ion batteries, it is clear that the values of the current upon discharge of a single cell and cell-series, i.e. the rate of operation, may have a significant impact on the battery's life. The use of discharge currents that are too low or too high, may have a detrimental effect on the battery's life, as shall be discussed later in Section 3.2. These effects justify the optimization efforts described in this paper. Based on these insights we propose a penalty function, which for each discharge current, defines a penalty in terms of the detrimental effect on the battery's life.
The motivation for our research emerges from the possibility of extending the life of EV batteries via their smart operation. The discharge method that is commonly used for EV batteries, presented in Section 4.1, is very simple as the current demand is supplied using all cells in the battery simultaneously, hence the load is equally divided among them. The rationale behind this method is simplicity of implementation and the assumption that the lower the current drawn from each cell in the battery the better. However, as described in Section 3.2, the behavior and performance of a real battery is more complex, and this assumption is not always true.
In this paper we focus on optimizing the process of smart distribution of the load of each demand over the cells in the EV battery. Our main theory is that not all the cell-series in the battery should be discharged together but rather, each time only part of them should be discharged. We propose advanced switching algorithms that select the cell-series to be discharged for each current demand and control the discharge current drawn from each.
The rest of this paper is organized as follows. In Section 2 we provide an overview of related work on switching methods. We formally describe the problem in Section 3. Our proposed switching algorithms are presented in Section 4. A description of the simulation data we used to evaluate our algorithms is provided in Section 5. The simulation results are presented in Section 6 and discussed in Section 7. Finally we present our conclusions and direction for future work in Section 8.
Section snippets
Related work
The subject of battery management for multiple battery systems has been widely studied. Most of the studies were intended to maximize the battery lifetime, i.e. the time until most of the cells in the battery have lost an essential part of their capacity in the course of prolonged cycling. Basically there are two kinds of discharge algorithms: sequential and parallel. In the sequential algorithms only one battery supplies the workload each time, while in the parallel algorithms a subset of
Problem description
The problem is defined as follows. The battery pack consists of m identical cell-series, s1,s2,…,sm, all having the same initial capacity C. There are n current demands, d1,d2,…,dn, where di is the required current in amperes (A) for the i-th second. The sequence of the current demands, and in particular its length n, is not known in advance. However, the total capacity of the cell-series is guaranteed to satisfy all demands, i.e. . A summary of the notations is presented in Table 1.
Switching algorithms
We present three discharge algorithms: naive in Section 4.1 and two heuristic algorithms in Sections 4.2 PreferOPT algorithm, 4.3 EqualLoad algorithm. The inputs of the algorithms are the cell-series statuses, values, i.e. their remaining capacities, and the current demand, di. The output of the algorithms is the current values allocated to the demand by the cell-series, values. For simplicity, as our algorithms refer to a single demand at a given time, di, we drop the index i whenever
Simulation data
We evaluated our proposed switching algorithm using computer simulations. The current demands were simulated using the main world-wide driving cycles used in the United States, Europe and Japan [26], [27], [28]. Below is a list of the driving cycles we used with a short description of their purpose (sources: [26], [29]).
United States driving cycles:
- •
The Urban Dynamometer Driving Schedule (UDDS), also called FTP-72 (Federal Test Procedure) or LA-4 cycle or FUDS, was developed by the United States
Simulation results
The presentation of the results is divided into two. First we present the penalty of the proposed heuristic algorithms compared to the naive algorithm which is used as a benchmark. Later we describe the differences between the performances of the proposed algorithms.
In general, apparently when using the proposed algorithms the penalty can be significantly reduced and almost totally avoided as the number of cell-series increases. In Fig. 4 the penalties of the EqualLoad algorithm compared to the
Discussion
In general, the results presented in Fig. 4 imply that by using the proposed heuristics a non-linear penalty reduction is possible for almost all driving cycles. In addition, that as the number of cell-series increases the improvement is better, i.e. the total penalty is lower. This behavior is reasonable and can be explained easily. As the number of cell-series increases there are more allocation options and more penalties can be avoided.
A statistical analysis of the results and the current
Conclusions and future work
A battery's life is greatly affected by the method of use and in this paper we focus on the discharge current. We defined a penalty function to measure the negative effects of discharge in non-optimal discharge currents and presented a switching algorithm that minimizes this function. We proposed switching algorithms that select a subset of the battery's cells for each current demand and control the discharge current from each.
The algorithms were evaluated by simulations on world-wide driving
References (33)
- et al.
Energy Conversion and Management
(2006) - et al.
Transportation Research Part D: Transport and Environment
(2001) - et al.
Journal of Power Sources
(1988) - et al.
Renewable and Sustainable Energy Reviews
(2012) - et al.
Electrochimica Acta
(2010) - et al.
Journal of Power Sources
(2007) - et al.
Journal of Power Sources
(2003) - et al.
IEEE Transactions on Industrial Electronics
(2005) A Guide to Understanding Battery Specifications
(December 2008)- J. Gartner, B. Gohn, Electric Vehicle Batteries, Tech. Rep., Pike Research,...
Extending Lifetime of Portable Systems by Battery Scheduling
Analysis of Discharge Techniques for Multiple Battery Systems
IEEE Transactions on Computers
IEEE Transactions on Vehicular Technology
Electrochemical and Solid-State Letters
IEEE Design & Test of Computers
Cited by (29)
Grid integration of battery swapping station: A review
2021, Journal of Energy StorageConfiguration method of hybrid energy storage system for high power density in More Electric Aircraft
2020, Journal of Power SourcesCitation Excerpt :In MEA, ESS not only decreases the weight but saves the cost [9]. There are many reports about cost [9], cycle life [10], and state of health [11] for ESS. However, the essential relation between the weight and load power are rarely published.
A review of Battery Electric Vehicle technology and readiness levels
2017, Renewable and Sustainable Energy ReviewsCitation Excerpt :In order to meet these goals, mathematical models of the behaviour of each battery technology have been elaborated [81]. It has been proposed models that take into account the change in capacity and impedance of batteries during their lifetime (decreasing the available power), the model allowing for the monitoring and prediction of degradation, and the development of advanced charging algorithms to maximize the battery life [82–84]. These, however, necessitate from cell manufacturers a high level of consistency in their products [85].
Simulation of lithium ion battery replacement in a battery pack for application in electric vehicles
2017, Journal of Power SourcesPhysics based modeling of a series parallel battery pack for asymmetry analysis, predictive control and life extension
2016, Journal of Power SourcesCitation Excerpt :A novel way to mitigate this issue is by using a reconfigurable pack, wherein when a certain number of cells fail or go faulty during operation they can be replaced with new cells and the pack can continue to be used. Adany et al. [32] and Kim and Shin [33] used switching algorithms where select cells of the pack are used at different demand conditions by controlling the current flowing through each of these groups of cells. Hamidi [34] propose a hardware level solution to develop a reconfigurable battery pack.