International Journal of Performability Engineering, 2018, 14(12): 3184-3194 doi: 10.23940/ijpe.18.12.p28.31843194

Lithium-Ion Battery Management System for Electric Vehicles

Linjie Lia, Zhaojun Li,b, Jingzhou Zhaob, and Wei Guoa

a Department of Industrial Engineering, University of Electronic Science and Technology of China, Chengdu, 611731, China

b College of Enginerring, Western New England University, Springfiled, Massachusettes, 01119, USA

*Corresponding Author(s): * E-mail address: zhaojun.li@wne.edu

First author contact:

Linjie Li is a Master’s student from the School of Mechanical and Electrical Engineering at University of Electronic Science and Technology of China. His research interests include prognostics and health management (PHM).
Zhaojun 'Steven' Li is with the Department of Industrial Engineering and Engineering Management, Western New England University. He received his PhD in Industrial Engineering from the University of Washington, USA. His current research interests include data analytics and its applications in engineering, such as prognostics and system health management, applied statistics and operations research, reliability management for new product development.
Jingzhou 'Frank' Zhao received his BS degree in Mechanical Engineering from Shanghai Jiao Tong University in 2011, MS degree from University of Wisconsin-Madison in 2013, and Ph.D. degree from University of California, Los Angeles, in 2017. He is currently an Assistant Professor in the Department of Mechanical Engineering, Western New England University, MA. His research focus is on the establishment of novel manufacturing processes, theoretical models, computational methods, and control strategies for the high-volume production of structured nanomaterials and advanced sensors.
Wei Guo received her BS degree in Control Theory and Engineering from Dalian University of Technology in 2005, PhD degree in Systems Engineering and Engineering Management from the City University of Hong Kong in 2011. Her main research interests are signal processing of rotating machinery, data analytics, and system health diagnostics and prognostics.

Accepted:  Published:   

Abstract

Lithium-ion batteries have been widely used as energy storage for electric vehicles (EV) due to their high power density and long lifetime. The high capacity and large quantity of battery cells in EV as well as the high standards of vehicle safety and reliability call for the agile and adaptive battery management system (BMS). BMS is one of the key technologies for electric vehicle development, which contributes to the overall performance of lithium-ion batteries in operations. Through a comprehensive literature review, this paper presents a review of lithium-ion battery management systems, including the main measurement parameters within a BMS, state estimation methods, cell equalization issues, thermal management strategies and research trends and progresses. The paper discusses and highlights the key elements and challenges with recommendations in terms of the development of next generation BMS technologies.

Keywords: lithium-ion battery in electric in electric vehicles; battery management system; parameters measurement; state estimation; equalization; thermal management

PDF (454KB) Metadata Related articles Export EndNote| Ris| Bibtex

Cite this article

Linjie Li, Zhaojun Li, Jingzhou Zhao, Wei Guo. Lithium-Ion Battery Management System for Electric Vehicles. International Journal of Performability Engineering, 2018, 14(12): 3184-3194 doi:10.23940/ijpe.18.12.p28.31843194

1. Introduction

As the shortage of fossil fuels and environmental problems become more and more serious [1], new energy vehicles will eventually substitute and replace conventional ones in the future. The electric vehicles (EVs) have attained huge attention due to their performance and efficiency in recent decades. Furthermore, most conventional automobile manufacturers are also developing electric vehicles. Thereby, electric vehicles are the developing trend of the automobile industry for the future.

One of the key components in EVs is the energy storage device. Various energy storages including lithium-ion batteries as well as lead acid and NiMH have been adopted in EVs [2]. Compared with the other types of batteries, lithium-ion batteries have advantages of long cycle lifetime, large power density and low self-discharge, and so on. Therefore, lithium-ion battery is widely accepted for the application in EVs [3]. Although lithium-ion batteries are ubiquitous in many portable electronic devices, it is still a challenging task for EV application. Battery pack needs to supply enormous power and energy for secure and reliable EVs [4]. As battery technology grows, it is equally important to develop effective battery management systems (BMS) to make the utilization of the batteries in EVs reliable, resilient, safe, efficient, economically and environmentally friendly.

BMS is one of the key technologies for EV development, which contributes to the overall performance and economic efficiency of lithium-ion battery applications when operated in severe environments. According to the measured voltage, current, temperature and other information from lithium-ion battery, BMS can estimate the battery states including state of health (SOH), remaining useful life (RUL), and state of charge (SOC). Based on these information, the functions like charging/discharging control, cell balancing and thermal management can be performed [5]. BMS is a bridge connecting the EV’s power system and battery, and its performance determines the efficient and safe utilization of the battery that accounts for the majority of cost in the EVs.

This paper presents and facilitates a comprehensive understanding of lithium-ion battery management systems through an extensive literature review.

2. Overview of Battery Management System

There has been various research involved in development of battery technology. Guo et al. [6] investigated guidelines for battery capacity fade modelling through designed experiments. As the battery technology develops, it is also equally important to develop an effective management system to meet the safety requirements of battery operations. In order to take advantage of the maximum performance of lithium-ion batteries when EVs operate, special attention must be paid to prevent the abnormality and avoid catastrophic failures. This is why an efficient and effective BMS is a vital unit in EVs. A high-end BMS can accurately estimate the health state of the battery and provide important information for the driver. In addition, BMS will respond to the abnormal conditions to ensure the reliability of battery and maximize the battery’s life. We will present an overview of BMS from two perspectives: the main functions and the structure of BMS.

2.1. The Main Function of Battery Management System (BMS) in Electric Vehicles (EVs)

With the wide application of batteries in EVs, BMS has become an essential component. In the meantime, the function of BMS is getting better gradually. A fundamental BMS may only contain a switch mechanism to prevent the battery from being overcharged or discharged in some small portable electronic products. However, for EVs, it is apparent that more functions are indispensable to sustain the safe operation. At present, a BMS of EVs generally covers the following functions and has been updated continually.

2.1.1. Cell Monitoring

In order to manage the battery, the first step is the acquisition of the battery’s information, including battery voltage, current, and temperature. Therefore, various sensing systems are needed to measure these parameters. The data are the prerequisite for estimating the state of the battery pack.

2.1.2. States Estimation

Battery states estimation is a very important part in a BMS. In this paper, we mainly refer to the two critical states of lithium-ion battery: state of charge (SOC) and state of health (SOH).

SOC is regarded as the remaining useful capacity of the battery, which acts as a gauge in gasoline vehicles. An accurate SOC estimation can suggest to a driver how long the EV can run and whether a charge is needed. [7]. Therefore, SOC can prevent an inappropriate operation prolonging the battery’s life. Despite the significance of SOC, it is a universal challenge to accurately estimate and predict SOC since it cannot be measured directly by utilizing current technologies. Consequently, SOC estimation can only be performed indirectly by designing effective algorithms based on the knowledge of battery mechanisms and measured data through cell monitoring [8].

SOH is another important indicator to evaluate the health condition of batteries and the degree of degradation. Currently, the definition of SOH is not as clear as SOC. SOH may be influenced by various factors and should be assessed by multiple parameters like capacity and impedance. Fuzzy statements like “fresh” and “aged” were previously used to represent the SOH roughly [9]. It is of course impractical to provide useful information for drivers. Pattipati et al. [5] proposed a quantitative measure that replaces fuzzy statements by the ratio between battery’s initial total capacity value and current total capacity value. This definition has been widely accepted and used for SOH estimation. A comprehensive assessment of battery health is very complex, which is the precondition when predicting the battery’s remaining useful life (RUL). RUL refers to the remaining number of times/cycles that a battery can be charged and discharged before its life is determined to be at the end-of-life (EoL). In addition, the EoL threshold of the battery is usually considered to be the condition when the capacity is reduced to around 75%-80% of its rated capacity value. Again, SOH cannot be measured directly. Special methods or algorithms must be established to assess SOH.

2.1.3. Cell Balancing

In order to provide adequate operation voltage and power to support the application to EVs, a battery pack consists of dozens or even hundreds/thousands of single battery cells connected in parallel and series configurations [10]. However, due to the constraints of the raw materials and manufacturing process of the single power battery, and the influence of the operation environment of the battery pack, the inconsistency between the battery cells is inevitable. This inconsistency can result in over-charge or over-discharge of individual cells in the charging and discharging cycling process, thereby reducing the battery capacity and shortening the battery life. In severe cases, the battery may also burn or explode, which may directly lead to safety issues. Cell balancing can manage the battery pack and balance the SOC or voltage of each individual cell by external hardware. The cell balancing technique is also a challenge and attracts a great deal of attentions.

2.1.4. Thermal Management

Temperature has great impacts on the lithium-ion battery performance. High temperature can result in significant reduction in terms of battery capacity, life, and energy usage efficiency. On the other hand, extremely low temperature can lead to the failure or malfunction of batteries. Thermal runaway can emerge under serious environmental conditions, causing the battery to be in the danger of explosion. Due to the variability of environment where EVs are running, batteries will operate under different temperatures, which results in bad performance of batteries. Effective thermal management can keep battery operating under optimum temperature so that the battery can reach its best status [11]. Therefore, the thermal management for battery is essential in BMS.

2.2. The Architecture of Battery Management System

Because cells are connected with each other, a rational approach is needed to manage all cells in battery packs. Jung et al. [12] proposed three methods to create a BMS. It is proven that a modular approach is suitable to build up a BMS for electric vehicle, as shown in Figure 1. Each lithium-ion battery module consisting of several battery cells is controlled and managed by a modular management system (MMS). Furthermore, all MMS are controlled by a central management system (CMS). Data exchange throughout the BMS is needed since BMS modules operate in a stand-alone mode. In order to exchange data within the BMS, a controllable transceiver is required. This architecture is flexible, scalable and has been commonly accepted.

Figure 1

Figure 1.   Modular concept of a battery management system [12]


3. Challenges and the State-of-the-Art of Related Techniques

Extensive research has been conducted to enhance the performance and augment the functionalities of the BMS such that the required standards can be met in EV applications. Designing BMS for EV applications can encounter many challenges and opportunities. In the following, related challenges in battery parameters measures, states estimation, and cell balancing are discussed.

3.1. Temperature Measurement

Temperature sensors are essential elements of BMS to ensure safety and performance of lithium-ion batteries. There are strong needs for advanced BMS that have real-time access to both local and distributed temperature information of each battery cell, enabling more precise estimations of SOC and RUL, as well as early detection and prevention of catastrophic events such as thermal runaway. Various sensing principles have been employed and developed in the past decade including but not limited to resistive sensors [13], thermoelectric sensors [14], infrared thermography [15], electrochemical impedance measurement [16], giant magnetoresistance (GMR) based Johnson noise thermometry (JNT) sensor [17], and fiber Bragg grating sensors [18]. These sensors are either utilized for direct temperature measurements at the desired locations or combined with physics-based or data-driven modelling methods for estimation of other key parameters. Non-commercially available sensors are also designed and fabricated by Screen Printing [13], or Micro-electro-mechanical-systems (MEMS) fabrication techniques [15] for higher spatial and temporal resolution, easier assembly, or lower cost. Challenges remain in the design and manufacturing of low cost, compact, non-intrusive, and high dynamic range (i.e. high spatial and temporal resolution with large coverage area) temperature sensors for smart BMS.

3.2. State of Charge Estimation Algorithms

The determination of battery SOC is the key in BMS. It is challenging to estimate the SOC accurately due to the high non-linearity property of the electric vehicle over the varying courses of driving processes.

3.2.1. Conventional Methods

One of the common conventional methods is Coulomb counting, in which the remaining capacity is calculated by simply accumulating the charge transferred in or out of a battery [19]. Although straightforward to perform, this method can only make a rough estimation because the measurement noise of current can make a great difference and the errors can be accumulated as it is an open-loop algorithm [20]. Furthermore, an accurate assessment of initial SOC value is the pre-condition of this method. Otherwise, the following calculation will be meaningless. This method is easy but very limited.

Another conventional method of SOC estimation is to measure the open circuit voltage (OCV) of a battery. It is a simple method based on the static relationship between the SOC and OCV. However, this approach is limited to certain types of batteries as the relationship between SOC and OCV varies from battery materials [21]. For instance, a lead-acid battery remains a linear relationship between SOC and OCV during the whole discharging process, while a lithium-ion battery will remain flat at a certain period [22]. This characteristic of lithium-ion batteries indicates that a small measurement error will generate an unacceptable result of SOC estimation when mapping the relationship. In addition, it is found that the graph about OCV and SOC is different at the period of discharging and charging according to plenty of experiments. This phenomenon is called “hysteresis”. The hysteresis of batteries must be considered as it will result in a low OCV when the battery is discharged and a high OCV when the battery is charged, as it can be seen from Figure 2.

Figure 2

Figure 2.   OCV curves of the LiFePO4-based batteries depending on the current direction, measured after various rest durations at each step [23]


In consideration of the individual shortcoming of the aforementioned methods, the OCV method is often combined with the coulomb counting method to resolve the error accumulation problem in coulomb counting [24], where we can implement periodic calibration using the OCV method. However, due to the long-term dynamics behaviour of batteries, it takes a long time for batteries to reach from an operating state to a stable state, which is commonly called the relaxation effect. This requirement makes the online estimation of SOC impractical.

According to the relationship between SOC and internal impedance, electrochemical impedance spectroscopy (EIS) is another method for SOC estimation. The study of EIS used to assess and estimate the SOC of lithium-ion batteries is relatively rare. The internal impedance of the batteries is assessed by imposing small current signals under varying frequencies to the battery and measuring the corresponding voltage using EIS analysers. This procedure costs a longer time and needs special equipment, leading to a high cost. Therefore, this method can be implemented in labs but is not suitable for EV application.

Due to the limitation of coulomb counting and OCV, multiple improvement methods are proposed on the basis of these methods. Kong, et al. [19] takes into account the charging and discharging efficiencies to improve the estimation accuracy. Zhang et al. [25] proposed an approach to estimate the initial SOC by taking into consideration of the relaxation effect and temperature of OCV. Furthermore, the quantitative relationship, i.e. the function describing the relationship among the terminal voltage u, resting time t, and temperature T to SOC0, is obtained by experimentation. This improvement method can realize the online estimation in EV application despite the relaxation effect of OCV. Although these improvements enhance the performance of SOC estimation to some extent, there also exists a series of problems in practical application.

3.2.2. Model-based Methods

The model-based approaches can be mainly divided into equivalent circuit models (ECMs) and electrochemical models (EMs). The model-based methods can realize the online estimation.

3.2.2.1. Equivalent Circuit Models

In ECMs, the battery is abstracted into an electric circuit. The electric circuit with diverse elements can approximately describe the physical phenomena inside a battery. Some classic ECMs, including Rint model, Thevenin model and PNGV model are put forward in succession. Recently, a lot of research is devoted to improving the model accuracy based on aforementioned models. On the basis of an established model, an algorithm is required to update the model’s parameters online, representing the dynamic characteristics. In the model, SOC is taken as one of internal states and an observer needs to be designed to obtain the state with the identified model and available measurements.

The accuracy of ECM is very important. Simultaneously, we should also take into account the complexity to reduce the computational time. The relaxation effort of a lithium-ion battery is a fundamental characteristic that emerges in the process of charging and discharging cycling. This characteristic can be modelled by series-connected parallel RC circuits. Two RC groups are suggested by [26] as the optimal trade-off between estimation accuracy and model complexity. A typical ECM is shown in Figure 3. Capacitors can store and release energy similarly in the charging and discharging cycles. Two RC groups are adopted to model the relaxation effort of a lithium-ion battery, while R0 represents the internal resistor. The value of R0 is not a constant but may change with the SOC, ambient temperature, and the aging effects of the batteries. VOC is the open-circuit voltage and has a static relationship with the SOC. VT and iL are the terminal voltage and current respectively and can be obtained by measurement.

Figure 3

Figure 3.   Battery model with relaxation effect, internal resistance, and VOC-SOC function [27]


There are also numerous research that focus on the online state estimation algorithms about the equivalent circuit. The actual output can be measured and is compared with the predicted output of the established model. Then, a gain or weight can be calculated to compensate for the error of the predicted state. Simultaneously, the parameters of the model should be updated using an online parameter identification and estimation method. The design of the observer is also equally important. Various techniques have been employed to construct observers to monitor the SOC, ranging from simple observers designed by trial and error to sophisticated, robust, optimal, and recursive techniques (e.g., Kalman filters and sliding mode observers). In [28], a comparative study is performed on three different model-based state observer architectures including Luenberger observer, Extended Kalman Filter (EKF), and Sigma Point Kalman Filter (SPKF) for the same battery model. The results show that a superior algorithm will improve the estimation accuracy as well as estimation robustness greatly.

3.2.2.2. Electrochemical Models

ECMs simply model the battery according to the external features so that plenty of information about internal underlying reactions are lost. Compared with ECMs, electrochemical models (EMs) apply mathematic models such as the partial differential equations (PDEs) to depict the real electrochemical reaction processes inside the cells. It is validated that EMs can capture cell dynamic behaviors with high accuracy compared with ECMs, and EMs have a good interpretability. Nalin [29] believes that an advanced BMS should use a physics-based electrochemical model rather than an equivalent model. However, the EM is very complicated, which involves a large amount of mathematical calculation, limiting the embedment in the BMS of an EV. Therefore, how to simplify the model remains a challenge.

The common EMs involve a porous electrode model and a single particle model. The porous electrode model has an overall description of various physical and chemical processes occurring in a battery, leading to a time-consuming solution. The single particle model treats each electrode as a single spherical particle, which was first presented by Haran et al. [30]. It has a faster computation speed; however, the validity is limited. What needs to be illustrated is that the single particle model is as valid as the porous electrode models for low discharge rates [31].

In [24], the battery health is taken into account when estimating SOC, including up-to-date capacity and resistance. A single particle model is presented by simplifying the 1-D-spatial model. This simplified model can significantly reduce the complexity while retaining a similar accuracy with the original one. Based on the simplified model, trinal observers are applied for the co-estimation task. The effectiveness of this method is verified through experiments with high estimation precision and robustness. However, this method is not compared with other methods on accuracy and speed to validate practical application.

3.2.2.3. Learning Algorithms

With the fast development of computer technology, learning algorithm has been very popular and is applied to many domains. Especially, many studies are being focused on the SOC estimation using the learning algorithm. The intelligent techniques estimate SOC by treating the battery as a “black-box” [32], which means that the complicated electrochemical reaction inside the battery is unnecessary to be known. With the well trained model offline using special algorithms, the learning method can accurately depict the relationship between SOC and related influential factors. The learning method is convenient but the training process is time-consuming. In addition, the prerequisite is that the historical data is abundant and reliable.

In [33], a three-layer BP network is used for SOC estimation. Through the training using experimental data, the results show that the estimation curve is in coincidence with the curve of practical value. However, the accuracy is not very high. In addition, this model just simply takes discharge voltage and discharge current as the input and SOC as the output, respectively while ignoring the effects of temperature, discharging rate, aging, etc.

3.3. Remaining Useful Life Prediction Approaches

Battery RUL information provides operators with a means in decision making by quantitatively knowing how much more time a battery can be used until its functionality is lost. Therefore, it is a key issue to accurately predict RUL for the battery health management system even though there are many challenges such as modelling insufficiencies, system noise, and degraded sensing fidelity. In recent years, many methods have been proposed and developed to predict battery remaining useful life. They are mainly divided into three categories: electrochemical mechanism analysis, data-driven method, and hybrid methods of combining previous two methods.

3.3.1. Remaining Useful Life Prediction based on Aging Mechanisms

This method firstly analyzes the electrochemical reaction inside a lithium-ion battery. Then a degradation model can be established according to the collected data. RUL is thus predicted using the model. The degradation model is focused on the covariate factors that can affect the battery capacity over time. This method takes advantage of deep learning about battery aging mechanisms. An accurate result can be obtained if the aging model is precise. However, in most instances, it is difficult to establish a precise degradation model. Moreover, the result of this method is the lack of the expression of uncertainties [34].

Pinson [35] proposed a theory of solid-electrolyte interphase (SEI) formation from the viewpoint of degradation mechanism. The theory indicates that the capacity loss is mainly attributed to the SEI formation at the negative electrode in a lithium ion battery by underlying physics analysis. A basic single particle reaction model is developed to quantitatively understand the capacity fade, focusing on the mechanism of SEI formation. Specifically, this method’s models are based on the electrochemical mechanism of a battery, using diffusion-reaction equation and Butler-Volmer equation. In addition, the diffusivity of electrolyte molecules through the SEI is an important parameter, which is temperature dependent. An Arrhenius dependence is utilized to model the temperature dependent diffusivity. Based on the established model, we can make clear how the battery gradually degrades and guides accelerated aging experiments. RUL can be inferred from the evolution of this model. However, such models are mostly limited to certain types of batteries [36].

3.3.2. Data-Driven Remaining Useful Life Prediction

With the development of artificial intelligence and machine learning, research works on data-driven prognostics have been shifted to the use of various statistical and deep learning methods, such as neural networks (NNS), Support Vector Machine (SVM), and Relevance Vector Machine (RVM) to predict battery RUL [37]. For example, a continuous hidden Markov model (HMM) is presented in [38] for RUL prediction. The data-driven method mines the hidden information in data to perform the prediction, so it is data-based. It does not need a precise model based on physics, which makes it popular with RUL prediction challenges. In addition, this method has high adaptability regardless of different types of batteries. However, this method requires the ability of data analysis and processing. A common drawback of this method is that the prognostic process can be opaque and has a poor interpretability. As a result, the parameters selection becomes a challenge when predicting RUL.

One other challenge of data-driven approaches is the health indicator (HI). Battery capacity is frequently used as the health indicator (HI) for degradation modeling and RUL estimation when the data of capacity is available because capacity is directly related to battery degradation. In [39], capacity and internal resistance are both predicted to obtain an accurate health status assessment of the lithium-ion batteries. However, considering the practical applications, battery is not always fully discharged at each cycle, so capacity cannot be calculated by integration. Therefore, we must extract an indirect HI parameter from the charging and discharging voltage, current, and time that can be directly measured for on-line degradation analysis. To tackle the problems and challenges, Liu et al. [40] proposed the time interval of equal discharging voltage difference (TIEDVD) as an HI. Then the Grey Correlation Analysis is conducted to analyze the consistency of TIEDVD and battery’s capacity as battery ages. The results show a strong similarity between the extracted TIEDVD series and the capacity series. Therefore, TIEDVD is one effective candidate as an HI to predict the RUL.

3.3.3. Hybrid Methods of Remaining Useful Life Prediction

Recently, the fusion methods of RUL prediction have become a hot topic, which combine multiple methods to overcome the shortage of individual ones. The hybrid methods can develop the advantages of different models, leading to a preferable performance.

In [41], a fusion of both data-driven and model-based approaches are proposed to realize the prognostics of lithium-ion RUL. A summation of exponential functions is used to describe the battery aging dynamics, while particle filtering (PF) is used to predict the RUL. The results show that the proposed fusion method has a more reliable RUL prediction. Saha et.al [42] proposed a combination of PF and neural networks using EIS to build an electrochemical model. Relevance vector machine (RVM) is used for the estimation of battery model parameters. PF achieves the adaptive selection of RVM parameters. An aging mechanism model is built using RVM and RUL prediction is obtained with extrapolation. In addition, based on the fact that the measurement model is hard to establish and the state cannot be updated at the period of prediction in model-based PF, two data-driven methods are introduced [43]. This fusion framework can update particles and weight in long term prediction and shows a superior prediction performance. In [34], a mean-covariance decomposition modelling method is proposed for battery capacity prognostics. The basic functional form of the mean function is obtained from chemical degradation mechanisms while SVM is applied to model the logarithm variance and angles. Although the hybrid method can perform a better performance than the respective one, the complexity and computational burden increases. In addition, the fusion approach is various and we should choose suitable one.

3.4. Cell Balancing

In general, lithium-ion cell balancing methods can be divided into two categories: passive balancing method and active balancing method, both of which are based on an equalization circuit. Passive balancing is relatively easy to achieve but it has a low efficiency. The extra energy stored in the overly charged cells is completely wasted as heat using a resistor. This method should be adopted only during the charging processes [44]. The balancing time is long. In contrast, active balancing transfers more charged cells to less charged ones. The circuit design is complicated but it can reach a very high efficiency. Due to the advantage of active balancing, it has become the main research direction. A tradeoff analysis between the circuit complexity of the active balancing method as well as cost and achieved efficiency must be considered. The relevant circuits have been proposed and the efficiency is up to 90% [45]. In addition, the balancing time is also an important indicator. A quick change requires fast balancing.

In addition, there are two types of balancing criteria: voltage based balance and SOC based balance. The voltage based balancing method is to control the operation state of balancing module through the voltage difference between single cells. This method is feasible because voltage variation can reflect SOC variation to some extent. It is easy to achieve using only the voltage sensing module. However, a lithium-ion battery has small voltage variation even in large difference of SOC, which cannot guarantee the consistency of residual capacity of each cell. It is not available for cells in parallel. The existing balancing methods are mostly based on voltage [45].

A fuzzy logic control for battery equalization based on SOC is proposed in [46]. Difference in SOC less than 0.3 will trigger the balancing process. The energy efficiency can reach around 95%. Besides SOC, the capacity degradation caused by the aging process is also taken into account in [44], using both the SOC and the capacity of each cell as the balancing criterion. It is demonstrated that the drawback of voltage balancing can be eliminated through analysis. Moreover, the charge time can also be decreased. However, these balancing methods must depend on a very accurate SOC and capacity estimation, which is also a challenge.

In the future, developing balancing methods of high efficiency, low cost, and fast equilibrium speed will be a trend.

3.5. Thermal Management

Some research is devoted to improving the thermal stability of batteries [47]; however, the precondition is to sacrifice battery capacity. High capacity is essential to ensure the required mileage and power of an EV. Therefore, the use of external battery thermal management systems will be a more suitable method. According to the different heat transfer medium, the battery thermal management system (TMS) can be classified into air cooling, liquid cooling, and Phase change material cooling (PCM). The details of these three types of TMSs will be discussed in the following sections.

3.5.1. Air Cooling

Air cooling method can be divided into natural convection and forced convection. Forced convection has high efficiency and is the most studied heat dissipation method. Air cooling system has the advantages of simple structure, low quality and low cost, which is generally regarded as the first choice for electric car cooling [48]. Some studies have focused on improving battery cell arrangement and ventilation pattern.

Optimal arrangement of battery cells can realize a uniform temperature distribution and improve cooling efficiency. In [49], parallel channels are adopted in TMS. Computational fluid dynamics (CFD) is used to simulate the temperature distribution in a cylinder battery pack. Three different battery arrangements are compared and optimally chosen. In [50], a battery pack with eight prismatic cells arranged in different ways is simulated using CFD. The experimental results show that when the gap between individual cells widens under a constant flow rate, the uniformity of temperature will improve. Also, a design of asymmetrical cells’ space will benefit the center cells by generate a relatively unblocked cooling environment.

Serial ventilation and parallel ventilation are two common ventilation modes. The effect of parallel ventilation is better than that of serial ventilation [51]. A parallel airflow is used in [52]. The test results show that the maximum temperature of battery pack can be controlled below 45℃ and the difference between monomer batteries is below 5℃ under the ambient temperature of 25℃. However, with the increase of the battery module and the size of the battery pack, both the parallel ventilation and the serial ventilation can easily cause the high temperature difference between the cells in different positions. A reciprocating air flow is proposed in [53]. Using the model, a 4℃ drop in temperature difference was achieved with a switching time interval of 120s and the maximum cell temperature reduced by 1.5℃ in contract with uni-directional cases.

3.5.2. Liquid Cooling

The air cooling can meet the requirement of thermal management under normal conditions. However, when the ambient temperature is high, the air cooling will not attach the expected heat dissipation effect, leading to the development of liquid cooling system. The advantages of liquid cooling include fast cooling velocity, higher cooling efficiency and small volume of cooling system.

Pesaran et al. [51] compared the air cooling with oil cooling with the 30W battery heat production and the initial temperature of 25℃. The experimental results show that the maximum temperature of oil cooling is 45℃ and air cooling is 54℃. When battery operates in extreme environment, liquid cooling can behave efficiently in controlling temperature. This remarkable merit has attracted popular attention.

3.5.3. Phase Change Material Cooling

Phase change material (PCM) can absorb enormous heat but the temperature change is very small. Al-Hallaj et al. [54] are the first ones to apply PCM to TMS. Battery or battery pack can be directly soaked in the PCM to achieve thermal management. Sabbah et al. [55] compared PCM cooling with active air cooling of high power lithium-ion batteries. Under high environmental temperature (40 ~ 45℃), the air cooling was out of work but PCM could keep temperatures lower than 55℃. Recently, PCM cooling has shown some good application potentials.

Compared to air and liquid, battery thermal management system using PCM as heat transfer medium is simple and doesn’t need battery to offer energy. It’s suitable under high environmental temperature.

4. Future Research Areas and Opportunities

A smart BMS is crucial for EVs industry. Although plenty of studies have been focused on different aspects of BMS, there still exist many challenges in developing next generation BMS for advanced EV applications. There are also many opportunities in different aspects for researchers, including battery states estimation, thermal management, cell balancing and so on. The future research areas and opportunities include, but are limited to the following.

$\cdot$ Various methods have been proposed and used for SOC estimation. However, each method has its own limitations. Many factors must be taken into account for accurate SOC estimation in EV applications. Future research on SOC estimation methods need to consider and incorporate both conventional learning algorithms and advanced analytical methods, e.g., artificial intelligence based methods to account for the effects of aging, discharge rate, and environmental operating conditions such as temperature and humidity.

$\cdot$ A proper definition of SOH needs to be settled for RUL prediction.

$\cdot$ An advanced and sophisticated BMS should be built on valid physics-based electrochemical models. Therefore, developing a simple but accurate electrochemical model will be needed.

$\cdot$ Developing effective methods to actively balance different cells and modules within the battery pack with high efficiency and low cost is also demanded.

$\cdot$ Improving battery thermal management systems to reduce the temperature difference between different individual cells and keep the optimal operation condition of battery needs to be studied.

5. Conclusion

BMS initially only had the function of battery voltage, current, and temperature measurement. The main purpose of BMS is to achieve battery performance monitoring. As technology evolves, BMS has many more features that not only monitors battery packs, but also controls and manages battery packs based on battery pack information through advanced sensing mechanisms. Advanced BMS can well improve battery operation efficiency and extend service life. Improving the practicality of BMS has become one of the core technologies of electric vehicles.

Currently, the research on BMS still has potential to improve. There are some challenges in the establishment of battery models, such as the accurate SOC estimation, the dynamic balance of the battery pack and so on. It is projected that with further research of BMS, the technology advancement of BMS will continue to improve. The endurance and safety of electric vehicle will be better guaranteed, and the popularization of electric vehicle become more reliable.

Acknowledgements

This research is partially supported by the National Science of Foundation of China, No. 71771038.

Reference

D. A. Notter, M. Gauch, R. Widmer, P. Wager, A. Stamp, R. Zah , et al., “

Contribution of Li-Ion Batteries to the Environmental Impact of Electric Vehicles

,” Environmental Science & Technology, Vol. 44, No. 17, pp. 6550-6, 2010

DOI:10.1021/es903729a      URL     PMID:20695466      [Cited within: 1]

Battery-powered electric cars (BEVs) play a key role in future mobility scenarios. However, little is known about the environmental impacts of the production, use and disposal of the lithium ion (Li-ion) battery. This makes it difficult to compare the environmental impacts of BEVs with those of internal combustion engine cars (ICEVs). Consequently, a detailed lifecycle inventory of a Li-ion battery and a rough LCA of BEV based mobility were compiled. The study shows that the environmental burdens of mobility are dominated by the operation phase regardless of whether a gasoline-fueled ICEV or a European electricity fueled BEV is used. The share of the total environmental impact of E-mobility caused by the battery (measured in Ecoindicator 99 points) is 15%. The impact caused by the extraction of lithium for the components of the Li-ion battery is less than 2.3% (Ecoindicator 99 points). The major contributor to the environmental burden caused by the battery is the supply of copper and aluminum for the production of the anode and the cathode, plus the required cables or the battery management system. This study provides a sound basis for more detailed environmental assessments of battery based E-mobility.

S. Manzetti and F. Mariasiu, “

Electric Vehicle Battery Technologies: From Present State to Future Systems

,” Renewable & Sustainable Energy Reviews, Vol. 51, pp. 1004-1012, 2015

DOI:10.1016/j.rser.2015.07.010      URL     [Cited within: 1]

Electric and hybrid vehicles are associated with green technologies and a reduction in greenhouse emissions due to their low emissions of greenhouse gases and fuel-economic benefits over gasoline and diesel vehicles. Recent analyses show nevertheless that electric vehicles contribute to the increase in greenhouse emissions through their excessive need for power sources, particularly in countries with limited availability of renewable energy sources, and result in a net contribution and increase in greenhouse emissions across the European continent. The chemical and electronic components of car batteries and their waste management require also a major investment and development of recycling technologies, to limit the dispersion of electric waste materials in the environment. With an increase in fabrication and consumption of battery technologies and multiplied production of electric vehicles worldwide in recent years, a full review of the cradle-to-grave characteristics of the battery units in electric vehicles and hybrid cars is important. The inherent materials and chemicals for production and the resulting effect on waste-management policies across the European Union are therefore reported here for the scope of updating legislations in context with the rapidly growing sales of electric and hybrid vehicles across the continent. This study provides a cradle-to-grave analysis of the emerging technologies in the transport sector, with an assessment of green chemistries as novel green energy sources for the electric vehicle and microelectronics portable energy landscape. Additionally, this work envisions and surveys the future development of biological systems for energy production, in the view of biobatteries. This work is of critical importance to legislative groups in the European Union for evaluating the life-cycle impact of electric and hybrid vehicle batteries on the environment and for establishing new legislations in context with waste handling of electric and hybrid vehicles and sustain new innovations in the field of sustainable portable energy.

J. T. Baer, B. C. Davis, R. J. Blanyer , “

Battery Management System

,” WO, US5698967, 1997

[Cited within: 1]

N. A. Chaturvedi, R. Klein, J. Christensen, J. Ahmed, A. Kojic , “

Algorithms for Advanced Battery-Management Systems

,” IEEE Control Systems, Vol. 30, No. 3, pp. 49-68, 2010

DOI:10.1109/MCS.2010.936293      URL     [Cited within: 1]

Lithium-ion (Li-ion) batteries are ubiquitous sources of energy for portable electronic devices. Compared to alternative battery technologies, Li-ion batteries provide one of the best energy-to-weight ratios, exhibit no memory effect, and have low self-discharge when not in use. These beneficial properties, as well as decreasing costs, have established Li-ion batteries as a leading candidate for the next generation of automotive and aerospace applications. In the automotive sector, increasing demand for hybrid electric vehicles (HEVs), plug-in HEVs (PHEVs), and EVs has pushed manufacturers to the limits of contemporary automotive battery technology. This limitation is gradually forcing consideration of alternative battery technologies, such as Li-ion batteries, as a replacement for existing leadacid and nickel-metal-hydride batteries. Unfortunately, this replacement is a challenging task since automotive applications demand large amounts of energy and power and must operate safely, reliably, and durably at these scales. The article presents a detailed description and model of a Li-ion battery. It begins the section "Intercalation-Based Batteries" by providing an intuitive explanation of the fundamentals behind storing energy in a Li-ion battery. In the sections "Modeling Approach" and "Li-Ion Battery Model," it present equations that describe a Li-ion cell's dynamic behavior. This modeling is based on using electrochemical principles to develop a physics-based model in contrast to equivalent circuit models. A goal of this article is to present the electrochemical model from a controls perspective.

B. Pattipati, K. Pattipati, J. P. Christopherson, S. M. Namburu, D. V. Prokhorov, “

Automotive Battery Management Systems

” in Proceedings of 2008 IEEE Autotestcon, pp. 581-586, 2008

[Cited within: 2]

J. Guo, Z. J. Li, T. Keyser , “

Modeling Li-Ion Battery Capacity Fade Using Designed Experiments

,” in Proceedings of the 2014 Industrial and Systems Engineering Research Conference, 2014

URL     [Cited within: 1]

The battery life cycle has been continuously improved over time due to advanced battery management systems, newly developed battery materials, and mature cell and battery pack manufacturing processes. However, the current magnitude of Li-Ion battery life cycle still remains a big challenge for many applications such as satellite and energy storage power sources which expect over decades of life time. This paper investigates the capacity fade mechanisms of Li-Ion battery over various discharge rates, battery types, and usage profiles in terms of the number of charge/discharge cycles. The designed experiments provide insights about the impact of discharge rates and battery types as well as their interactions on battery performance metric of capacity fade over prolonged usage cycles. The experimental and analysis results provide generic guidelines for the optimal battery cell design and battery life optimization through appropriate battery usage and operations.

W. He, N. Williard, C. Chen, M. Pecht , “

State of Charge Estimation for Li-Ion Batteries Using Neural Network Modeling and Unscented Kalman Filter-based Error Cancellation

,” International Journal of Electrical Power & Energy Systems, Vol. 62, No. 11, pp. 783-791, 2014

DOI:10.1016/j.ijepes.2014.04.059      URL     [Cited within: 1]

Lithium-ion batteries have been widely used as the energy storage systems in personal portable electronics (e.g. cell phones, laptop computers), telecommunication systems, electric vehicles and in various aerospace applications. To prevent the sudden loss of power of battery-powered systems, there are various approaches to estimate and manage the battery's state of charge (SOC). In this paper, an artificial neural network ased battery model is developed to estimate the SOC, based on the measured current and voltage. An unscented Kalman filter is used to reduce the errors in the neural network-based SOC estimation. The method is validated using LiFePO4 battery data collected from the Federal Driving Schedule and dynamical stress testing.

H. Rahimi-Eichi, U. Ojha, F. Baronti, M. Chow , “

Battery Management System: An Overview of its Application in the Smart Grid and Electric Vehicles

,” Industrial Electronics Magazine IEEE, Vol. 7, No. 2, pp. 4-16, 2013

DOI:10.1109/MIE.2013.2250351      URL     [Cited within: 1]

With the rapidly evolving technology of the smart grid and electric vehicles (EVs), the battery has emerged as the most prominent energy storage device, attracting a significant amount of attention. The very recent discussions about the performance of lithium-ion (Li-ion) batteries in the Boeing 787 have confirmed so far that, while battery technology is growing very quickly, developing cells with higher power and energy densities, it is equally important to improve the performance of the battery management system (BMS) to make the battery a safe, reliable, and cost-efficient solution. The specific characteristics and needs of the smart grid and EVs, such as deep charge/discharge protection and accurate state-of-charge (SOC) and state-of-health (SOH) estimation, intensify the need for a more efficient BMS. The BMS should contain accurate algorithms to measure and estimate the functional status of the battery and, at the same time, be equipped with state-of-the-art mechanisms to protect the battery from hazardous and inefficient operating conditions.

E. Meissner and G. Richter, “

Battery Monitoring and Electrical Energy Management: Precondition for Future Vehicle Electric Power Systems

,” Journal of Power Sources, Vol. 116, No. 1-2, pp. 79-98, 2003

DOI:10.1016/S0378-7753(02)00713-9      URL     [Cited within: 1]

New vehicle electric systems are promoted by the needs of fuel economy and ecology as well as by new functions for the improvement of safety and comfort, reliability, and the availability of the vehicle. Electrically controlled and powered systems for braking, steering and stabilisation need a reliable supply of electrical energy. The planned generation of electrical energy (only when it is economically beneficial meaningful), an adequate storage, and thrifty energy housekeeping with an intelligent integration of the battery as the storage medium into the overall concept of the vehicle Energy Management, and early detection of possible restrictions of reliability by Battery Monitoring allows for actions by the Energy Management well in advance, while the driver need not be involved at all. To meet today’s requirements for Battery Monitoring and Energy Management, solutions have been developed for series vehicles launched in years 2001–2003, operating at the 14 V level.

M. Einhorn, W. Roessler, J. Fleig , “

Improved Performance of Serially Connected Li-Ion Batteries with Active Cell Balancing in Electric Vehicles

,” IEEE Transactions on Vehicular Technology, Vol. 60, No. 6, pp. 2448-2457, 2011

DOI:10.1109/TVT.2011.2153886      URL     [Cited within: 1]

This paper presents an active cell balancing method for lithium-ion battery stacks using a flyback dc/dc converter topology. The method is described in detail, and a simulation is performed to estimate the energy gain for ten serially connected cells during one discharging cycle. The simulation is validated with measurements on a balancing prototype with ten cells. It is then shown how the active balancing method with respect to the cell voltages can be improved using the capacity and the state of charge rather than the voltage as the balancing criterion. For both charging and discharging, an improvement in performance is gained when having the state of charge and the capacity of the cells as information. A battery stack with three single cells is modeled, and a realistic driving cycle is applied to compare the difference between both methods in terms of usable energy. Simulations are also validated with measurements.

C. Jiang, Z. G. Tang, J. X. Hao, H. Q. Li , “

A Novel Design of Air-Cooled Battery Thermal Management System for Electric Vehicle

,” Applied Mechanics & Materials, Vol. 563, pp. 362-365, 2014

DOI:10.4028/www.scientific.net/AMM.563.362      URL     [Cited within: 1]

Effective thermal management of battery pack is essential for electric vehicle to adapt to different kinds of external environment, achieve desired working efficiency and life cycle of the power batteries. In this paper, a novel thermal management system is designed for battery pack in electric vehicle. Visualized simulation analysis of the thermal management system is carried on under different working conditions by CFD, then the structure parameters will be optimized. According to the conclusion, the thermal management system has been indicated to be effective to ensure an appropriate temperature range and the normal work of the power batteries in electric vehicle.

M. Jung and S. Schwunk, “

High End Battery Management Systems for Renewable Energy and EV Applications

,” Green, Vol. 3, No. 1, pp. 19-26, 2013

[Cited within: 2]

J. Grosch, E. Teuber, M. Jank, V. Lorentz, M. März, L. Frey, “

Device Optimization and Application Study of Low Cost Printed Temperature Sensor for Mobile and Stationary Battery based Energy Storage Systems

” in Proceedings of 2015 IEEE International Conference on Smart Energy Grid Engineering (SEGE), pp. 1-7, 2015

DOI:10.1109/SEGE.2015.7324599      URL     [Cited within: 2]

One of the most important physical parameters for state estimation in battery based Energy Storage Systems (ESS) is the temperature. This physical quantity does not only strongly influence state estimation for battery management systems, but also significantly affects lifetime and return on investment finally. Thus, monitoring the cell temperature is essential when high performance and efficiency is demanded. Contrary to this fact, less temperature sensors than battery cells are implemented in state of the art battery systems, to limit system costs. In this paper a low cost temperature sensor is presented. Based on printed electronics technology, a broad spectrum of designs and substrates is processable which leads to a variety of possible applications. After the selection of design and concept for battery applications, the processing of the sensor device is described. The main part of the paper is about the experimental validation of the printed temperature sensor performance. In a high power charge and discharge cycle of a single battery cell, the printed sensor is directly compared to state of the art temperature sensors implemented in mobile or stationary battery systems. Finally, the results are discussed and future perspectives are given. Both, the advantages and disadvantages of the printed temperature sensor are shown, whereas for the latter possible solutions are pointed out with respect to further developments.

J. Smith, M. Hinterberger, C. Schneider, J. Koehler , “

Energy Savings and Increased Electric Vehicle Range Through Improved Battery Thermal Management

,” Applied Thermal Engineering, Vol. 101, pp. 647-656, 2016

DOI:10.1016/j.applthermaleng.2015.12.034      URL     [Cited within: 1]

Lithium-ion cells are temperature sensitive: operation outside the optimal operating range causes premature aging and correspondingly reduces vehicle range and battery system lifetime. In order to meet consumer demands for electric and hybrid-electric vehicle performance, especially in adverse climates, a battery thermal management system (BTMS) is often required. This work presents a novel experimental method for analyzing BTMS using three sample cooling plate concepts. For each concept, the input parameters (ambient temperature, coolant temperature and coolant flow rate) are varied and the resulting effect on the average temperature and temperature distribution across and between cells is compared. Additionally, the pressure loss along the coolant path is utilized as an indicator of energy efficiency. Using the presented methodology, various cooling plate layouts optimized for production alternative techniques are compared to the state of the art. It is shown that these production-optimized cooling plates provide sufficient thermal performance with the additional benefit of mechanical integration within the battery and/or vehicle system. It is also shown that the coolant flow influences battery cell thermal behavior more than the solid material and that pressure drop is more sensitive to geometrical changes in the cooling plate than temperature changes at the module.

C. Heubner, C. Lämmel, N. Junker, M. Schneider, A. Michaelis , “

Microscopic in-Operando Thermography at the Cross Section of a Single Lithium Ion Battery Stack

,” Electrochemistry Communications, Vol. 48, pp. 130-133, 2014

DOI:10.1016/j.elecom.2014.09.007      URL     [Cited within: 2]

61A cell design for microscopic in-operando thermography at the cross section of a single battery stack has been developed.61The thermographic measurements point out differences in the local heat generation rates of the individual components.61The experimental setup allows the investigation of the heat transport and the observation of lateral inhomogeneity.

L. H. Raijmakers, D. L. Danilov, J. P. V. Lammeren, T. J. Lammers, H. J. Bergveld, and P. H. Notten, “

Non-Zero Intercept Frequency: An Accurate Method to Determine the Integral Temperature of Li-Ion Batteries

,” IEEE Transactions on Industrial Electronics, Vol. 63, No. 5, pp. 3168-3178, 2016

DOI:10.1109/TIE.2016.2516961      URL     [Cited within: 1]

A new impedance-based approach is introduced in which the integral battery temperature is related to other frequencies than the recently developed zero-intercept frequency (ZIF). The advantage of the proposed non-ZIF (NZIF) method is that measurement interferences, resulting from the current flowing through the battery (pack), can be avoided at these frequencies. This gives higher signal-to-noise ratios (SNRs) and, consequently, more accurate temperature measurements. A theoretical analysis, using an equivalent circuit model of a Li-ion battery, shows that NZIFs are temperature dependent in a way similar to the ZIF and can therefore also be used as a battery temperature indicator. To validate the proposed method, impedance measurements have been performed with individual LiFePO4 batteries and with large LiFePO4 battery packs tested in a full electric vehicle under driving conditions. The measurement results show that the NZIF is clearly dependent on the integral battery temperature and reveals a similar behavior to that of the ZIF method. This makes it possible to optimally adjust the NZIF method to frequencies with the highest SNR.

X. Liu, C. Zheng, C. Liu, P. W. Pong , “

Experimental Investigation of a Johnson Noise Thermometry Using GMR Sensor for Electric Vehicle Applications

,” IEEE Sensors Journal, Vol. 18, No. 8, pp. 3098-3107, 2018

DOI:10.1109/JSEN.2018.2805309      URL     [Cited within: 1]

A new temperature measurement method using a giant magnetoresistance (GMR) sensor based on Johnson noise thermometry (JNT) is proposed in this work. This paper presents the principle of this GMR-based JNT and demonstrates its feasibility by measuring power spectral density (PSD) of noise voltage in the bandwidth from 10 kHz to 22.8 kHz as well as resistance of GMR sensor and analyzing their relationship with the absolute temperature. The measuring errors throughout the measurement of temperature from 6140 °C and 150 °C in the thermal chamber were less than ±1.8 °C with integration time of 58.6s. Its dynamic sensing performance under both varying temperature and changing external magnetic field was characterized and demonstrated. A practical demonstration of the GMR-based JNT for measurement of surface temperature of a battery pack on an electric-vehicle testbed was also provided. Therefore, it is feasible to implement this new thermometry with a GMR sensor, making spintronic sensor multifunctional in EV applications by being a temperature sensor as well.

M. Nascimento, M. S. Ferreira, J. L. Pinto , “

Real Time Thermal Monitoring of Lithium Batteries with Fiber Sensors and Thermocouples: A Comparative Study

,” Measurement, Vol. 111, pp. 260-263, 2017

DOI:10.1016/j.measurement.2017.07.049      URL     [Cited within: 1]

For this specific application, the results show that the fiber Bragg grating sensors have better resolution and a rise time 28.2% lower than the K‐type thermocouples, making them a better choice for the real time monitoring of the battery surface temperature as well as a useful tool for failure detection and an optimized management in batteries.

S. N. Kong, C. S. Moo, Y. P. Chen, Y. C. Hsieh , “

Enhanced Coulomb Counting Method for Estimating State-of-Charge and State-of-Health of Lithium-Ion Batteries

,” Applied Energy, Vol. 86, No. 9, pp. 1506-1511, 2009

DOI:10.1016/j.apenergy.2008.11.021      URL     [Cited within: 2]

The coulomb counting method is expedient for state-of-charge (SOC) estimation of lithium-ion batteries with high charging and discharging efficiencies. The charging and discharging characteristics are investigated and reveal that the coulomb counting method is convenient and accurate for estimating the SOC of lithium-ion batteries. A smart estimation method based on coulomb counting is proposed to improve the estimation accuracy. The corrections are made by considering the charging and operating efficiencies. Furthermore, the state-of-health (SOH) is evaluated by the maximum releasable capacity. Through the experiments that emulate practical operations, the SOC estimation method is verified to demonstrate the effectiveness and accuracy.

E. Leksono, I. N. Haq, M. Iqbal, F. X. N. Soelami, and I. G. N. Merthayasa, “

State of Charge (SoC) Estimation on LiFePO4 Battery Module Using Coulomb Counting Methods with Modified Peukert. Joint

,” in Proceedings of IEEE International Conference on Rural Information & Communication Technology and Electric-Vehicle Technology, pp. 1-4, 2014

DOI:10.1109/rICT-ICeVT.2013.6741545      URL     [Cited within: 1]

Batteries have an important role in the development of electrical energy utilization, such as in renewable energy and electric vehicles. Batteries with good performance would support the devices which utilized them. Because the amount of energy stored in a battery is limited, so battery getting charging and discharging cycles. Improper charge and discharge process could decrease the battery's performance. Therefore battery management system (BMS) is very necessary so that could maintain performance of the battery at optimum condition. One of the important aspect in BMS is state of charge (SoC), which indicate the amount of energy left in the battery. In this work, SOC estimation calculated using Coulomb Counting method which is calculate and compare the electric charge that came in and came out the battery. Upon Coulomb counting on discharging and charging process, battery's SoC estimated by different ways, namely calculate the Peukert's effect and calculate charging efficiency, respectively. Modified Peukert relationship includes current rate parameters are affecting discharge capacity of a battery. Higher Peukert constant would lead to decreasing of operational battery module voltages. Moreover, it could be studied that Peukert only has an effect in the changes of discharge capacity at zero point of SoC. Furthermore, from experimental result it has been known that energy stored in the battery might be has different deep of discharge depends on Peukert constant.

T. Dong, J. Li, F. Zhao, Y. You, Q. Jin, “

Analysis on the Influence of Measurement Error on State of Charge Estimation of LiFePO4 Power Battery

” in Proceedings of 2011 International Conference on Materials for Renewable Energy & Environment, pp. 644-649, 2011

DOI:10.1109/ICMREE.2011.5930893      URL     [Cited within: 1]

The error sources of state of charge(SOC) estimation of li-ion battery based on Kalman filter is analyzed in this paper. To LiFePO4 power battery, the extremely flat part during the normal SOC range of SOC-OCV curve will require strong restriction to the voltage detection especially. This paper aims at the influence of the measuring precision of the voltage and current signal of battery management system (BMS) to SOC estimation of the li-ion battery including LiFePO4 battery. A simulation analysis is performed independently in Simulink on the assumption that other factors are under ideal conditions, in which the effects of Gaussian white noise and offset error of measurement of BMS are discussed respectively to simulate the actual vehicle condition. The principle to the precision index design of BMS is proposed according to the analysis result, in addition, a high-precision data acquisition system is developed as a precise calibration benchmark device for BMS.

X. Tang, Y. Wang, Z. Chen , “

A Method for State-of-Charge Estimation of LiFePO4, Batteries based on a Dual-Circuit State Observer

,” Journal of Power Sources (ISSN), Vol. 296, pp. 23-29, 2015

DOI:10.1016/j.jpowsour.2015.07.028      URL     [Cited within: 1]

61A dual-circuit based observer is proposed to estimate the SOC of LiFePO4 batteries.61The computation of the algorithm is light-weighted with no matrix operations.61The phenomenon of quasi-unobservability is discussed in this study.61An easy drifting current corrector is proposed in this study.

Roscher, A. Michael, D. U. Sauer, “

Dynamic Electric Behavior and Open-Circuit-Voltage Modeling of LiFePO4 -based Lithium ion Secondary Batteries

,” Journal of Power Sources, Vol. 196, No. 1, pp. 331-336, 2011

[Cited within: 1]

L. Zheng, L. Zhang, J. Zhu, G. Wang, J. Jiang , “

Co-Estimation of State-of-Charge, Capacity and Resistance for Lithium-Ion Batteries based on a High-Fidelity Electrochemical Model

,” Applied Energy, Vol. 180, pp. 424-434, 2016

DOI:10.1016/j.apenergy.2016.08.016      URL     [Cited within: 2]

Lithium-ion batteries have been widely used as enabling energy storage in many industrial fields. Accurate modeling and state estimation play fundamental roles in ensuring safe, reliable and efficient operation of lithium-ion battery systems. A physics-based electrochemical model (EM) is highly desirable for its inherent ability to push batteries to operate at their physical limits. For state-of-charge (SOC) estimation, the continuous capacity fade and resistance deterioration are more prone to erroneous estimation results. In this paper, trinal proportional-integral (PI) observers with a reduced physics-based EM are proposed to simultaneously estimate SOC, capacity and resistance for lithium-ion batteries. Firstly, a numerical solution for the employed model is derived. PI observers are then developed to realize the co-estimation of battery SOC, capacity and resistance. The moving-window ampere-hour counting technique and the iteration-approaching method are also incorporated for the estimation accuracy improvement. The robustness of the proposed approach against erroneous initial values, different battery cell aging levels and ambient temperatures is systematically evaluated, and the experimental results verify the effectiveness of the proposed method.

Y. Zhang, W. Song, S. Lin, Z. Feng , “

A Novel Model of the Initial State of Charge Estimation for LiFePO4 Batteries

,” Journal of Power Sources, Vol. 248, No. 4, pp. 1028-1033, 2014

DOI:10.1016/j.jpowsour.2013.09.135      URL     [Cited within: 1]

61Build a dynamic multi-parameter SOC0 model to enhanced the estimation accuracy.61The effect of relaxation effect and temperature of open circuit voltage also is described.61A new method to calculate energy efficiency is used successfully and described.61The estimation approaches are verified with typical driving profiles.

M. Chen and G. A. Rincon-Mora, “

Accurate Electrical Battery Model Capable of Predicting Runtime and i-v Performance

,” IEEE Transactions on Energy Conversion, Vol. 21, No. 2, pp. 504-511, 2006

DOI:10.1109/TEC.2006.874229      URL     [Cited within: 1]

Low power dissipation and maximum battery runtime are crucial in portable electronics. With accurate and efficient circuit and battery models in hand, circuit designers can predict and optimize battery runtime and circuit performance. In this paper, an accurate, intuitive, and comprehensive electrical battery model is proposed and implemented in a Cadence environment. This model accounts for all dynamic characteristics of the battery, from nonlinear open-circuit voltage, current-, temperature-, cycle number-, and storage time-dependent capacity to transient response. A simplified model neglecting the effects of self-discharge, cycle number, and temperature, which are nonconsequential in low-power Li-ion-supplied applications, is validated with experimental data on NiMH and polymer Li-ion batteries. Less than 0.4% runtime error and 30-mV maximum error voltage show that the proposed model predicts both the battery runtime and I-V performance accurately. The model can also be easily extended to other battery and power sourcing technologies.

H. Rahimi-Eichi, F. Baronti, M. Y. Chow , “

Online Adaptive Parameter Identification and State-of-Charge Coestimation for Lithium-Polymer Battery Cells

,” IEEE Transactions on Industrial Electronics, Vol. 61, No. 4, pp. 2053-2061, 2013

DOI:10.1109/TIE.2013.2263774      URL     [Cited within: 1]

Real-time estimation of the state of charge (SOC) of the battery is a crucial need in the growing fields of plug-in hybrid electric vehicles and smart grid applications. The accuracy of the estimation algorithm directly depends on the accuracy of the model used to describe the characteristics of the battery. Considering a resistance-capacitance (RC)-equivalent circuit to model the battery dynamics, we use a piecewise linear approximation with varying coefficients to describe the inherently nonlinear relationship between the open-circuit voltage (VOC) and the SOC of the battery. Several experimental test results on lithium (Li)-polymer batteries show that not only do the VOC-SOC relationship coefficients vary with the SOC and charging/discharging rates but also the RC parameters vary with them as well. The moving window least squares parameter-identification technique was validated by both data obtained from a simulated battery model and experimental data. The necessity of updating the parameters is evaluated using observers with updating and nonupdating parameters. Finally, the SOC coestimation method is compared with the existing well-known SOC estimation approaches in terms of performance and accuracy of estimation.

J. Li, J. K. Barillas, C. Guenther, M. A. Danzer , “

A Comparative Study of State of Charge Estimation Algorithms for LiFePO4, Batteries used in Electric Vehicles

,” Journal of Power Sources, Vol. 230, No. 10, pp. 244-250, 2013

DOI:10.1016/j.jpowsour.2012.12.057      URL     [Cited within: 1]

One of the most important aspects in battery management systems (BMS) in electric vehicles is the state of charge (SOC) estimation. SOC needs to be accurately determined for safety and performance reasons but cannot be measured directly due to the flatness and hysteresis of the open circuit voltage (OCV) curve of Lithium-ion chemistries as LiFePO4. The classical approach of current integration (Coulomb counting) cannot solve the problems of accumulative error and inaccurate initial values, thus advanced estimation algorithms are applied to determine the sate of charge. In this work, three model-based state observer designs including Luenberger observer, Extended Kalman Filter (EKF) and Sigma Point Kalman Filter (SPKF) are carried out and studied. These estimation approaches are verified using measurement data acquired from commercial LiFePO4 cells. In addition, computational tests analyze the systems performances in terms of tracking accuracy, estimation robustness against temperature uncertainty, sensor drift, and convergence behavior with an initial SOC offset.

N. A. Chaturvedi, R. Klein, J. Christensen, J. Ahmed, A. Kojic , “

Algorithms for Advanced Battery-Management Systems

,” IEEE Control Systems, Vol. 30, No. 3, pp. 49-68, 2010

DOI:10.1109/MCS.2010.936293      URL     [Cited within: 1]

Lithium-ion (Li-ion) batteries are ubiquitous sources of energy for portable electronic devices. Compared to alternative battery technologies, Li-ion batteries provide one of the best energy-to-weight ratios, exhibit no memory effect, and have low self-discharge when not in use. These beneficial properties, as well as decreasing costs, have established Li-ion batteries as a leading candidate for the next generation of automotive and aerospace applications. In the automotive sector, increasing demand for hybrid electric vehicles (HEVs), plug-in HEVs (PHEVs), and EVs has pushed manufacturers to the limits of contemporary automotive battery technology. This limitation is gradually forcing consideration of alternative battery technologies, such as Li-ion batteries, as a replacement for existing leadacid and nickel-metal-hydride batteries. Unfortunately, this replacement is a challenging task since automotive applications demand large amounts of energy and power and must operate safely, reliably, and durably at these scales. The article presents a detailed description and model of a Li-ion battery. It begins the section "Intercalation-Based Batteries" by providing an intuitive explanation of the fundamentals behind storing energy in a Li-ion battery. In the sections "Modeling Approach" and "Li-Ion Battery Model," it present equations that describe a Li-ion cell's dynamic behavior. This modeling is based on using electrochemical principles to develop a physics-based model in contrast to equivalent circuit models. A goal of this article is to present the electrochemical model from a controls perspective.

B. S. Haran, B. N. Popov, R. E. White , “

Determination of the Hydrogen Diffusion Coefficient in Metal Hydrides by Impedance Spectroscopy

,” Journal of Power Sources, Vol. 75, No. 1, pp. 56-63, 1998

DOI:10.1016/S0378-7753(98)00092-5      URL     [Cited within: 1]

Cobalt coatings on metal hydrides give rise to an additional capacity due to the cobalt on the surface of the alloy. In such a case, the galvanostatic discharge technique cannot be used to measure the diffusion coefficient of hydrogen in the alloy. We present here an analytical impedance model of the metal hydride electrode. This simple model is used to calculate the diffusion coefficient of hydrogen in the metal alloy. The impedance response of cobalt microencapsulated LaNi Sn electrode was measured at different hydrogen contents. From the slope of the Nyquist plot in the transition region the diffusion coefficient of hydrogen was calculated at various states of charge (SOC). It is seen that the diffusion coefficient increases with the hydrogen content in the alloy. The technique applied for cobalt encapsulated LaNi Sn could however be used for any hydrogen storage alloy.

S. Santhanagopalan, Q. Guo, P. Ramadass, R. E. White , “

Review of Models for Predicting the Cycling Performance of Lithium Ion Batteries

,” Journal of Power Sources, Vol. 156, No. 2, pp. 620-628, 2006

DOI:10.1016/j.jpowsour.2005.05.070      URL     [Cited within: 1]

A rigorous pseudo two-dimensional model to simulate the cycling performance of a lithium ion cell is compared with two simplified models. The advantage of using simplified models is illustrated and their limitations are discussed. It is shown that for 1 C or less discharge rates a simple ordinary differential equation (ODE) model can be used to predict accurately the potential as a function of time. For rates higher than 1 C, simplifications to the rigorous model are suggested that reduce the solution time for the model.

L. W. Kang, X. Zhao, J. Ma , “

A New Neural Network Model for the State-of-Charge Estimation in the Battery Degradation Process

,” Applied Energy, Vol. 121, No. 5, pp. 20-27, 2014

DOI:10.1016/j.apenergy.2014.01.066      URL     [Cited within: 1]

Battery state-of-charge (SOC) is a key parameter of the battery management system in the electric vehicle. To predict the practicable capacity of the battery in the degradation process, the cycle life model is built based on the aging cycle tests of the 6Ah Lithium Ion battery. Combined with the cycle life model, a new Radial Basis Function Neural Network (RBFNN) model is proposed to eliminate the battery degradation effect on the SOC estimation accuracy of the original trained model. This proposed model is verified through the 6Ah Lithium Ion battery. First, Urban Dynamometer Driving Schedule (UDDS) and Economic Commission of Europe (ECE) cycles are experimented on the batteries under different temperatures and aging levels. Then, the robustness of the new RBFNN model against different aging levels, temperatures and loading profiles is tested with the datasets of the experiments and compared against the conventional neural network model. The simulations show that the new model can improve the accuracy of the SOC estimation effectively and has a good robustness against varying aging cycles, temperatures and loading profiles. Finally, the measurement of actual aging cycles of the battery in electric vehicles is discussed for the SOC estimation.

R. H. Liu, Y. K. Sun, X. F. Ji, “

Battery State of Charge Estimation for Electric Vehicle based on Neural Network

” in Proceedings of 2011 International Conference on Information and Computer Networks, pp. 493-496, 2011

DOI:10.1109/ICCSN.2011.6013952      URL     [Cited within: 1]

Prediction of battery's remaining capacity is always a significant issue to which electric vehicle researchers paid close attention. Batteries of different types or the same type batteries of different model varies in prediction model of remaining capacity, the expert's advice obtained from experiment is not so universal that it is significant to build and improve the prediction model of remaining capacity for batteries of different types. This article takes iron phosphate Li-ion battery as the object of study, based on charge-discharge performance test of iron phosphate Li-ion battery, introduces neural network method to build prediction model for remaining capacity of battery and verify the model with test data in the end.

J. Guo and Z. J. Li , “

A Mean-Covariance Decomposition Modeling Method for Battery Capacity Prognostics

,” in Proceedings of the 2017 International Conference on Sensing, Prognostics, and Control, pp. 16-18, Shanghai, August 2017

[Cited within: 2]

M. B. Pinson and M. Z. Bazant, “

Theory of SEI formation in Rechargeable Batteries: Capacity Fade, Accelerated Aging and Lifetime Prediction

,” Journal of the Electrochemical Society, Vol. 160, No. 2, pp. A243-A250, 2012

[Cited within: 1]

J. Guo and Z. J. Li , “

Prognostics of Lithium Ion Battery Using Functional Principal Component Analysis

,” in Proceedings of the 2017 International Conference on Prognostics and Health Management, pp. 19-21, Allen, TX, June 2017

[Cited within: 1]

M. Rezvani, S. Lee, J. Lee , “

A Comparative Analysis of Techniques for Electric Vehicle Battery Prognostics and Health Management (PHM)

,” SAE Technical Paper, 2011

[Cited within: 1]

Y. Lin, M. Y. Hu, X. H. Yin, J. Guo, Z. J. Li , “

Evaluation of Lithium Batteries based on Continuous Hidden Markov Model

,” in Proceedings of the 2017 IEEE International Conference on Software Quality, Reliability and Security Companion, pp. 25-29, Prague, Czech, July 2017

DOI:10.1109/QRS-C.2017.43      URL     [Cited within: 1]

This paper presented a method of evaluating the health of lithium battery based on the continuous hidden Markov model (CHMM). This paper focuses on how to use CHMM to build the evaluation model. The capacity of battery is chosen as the observation variable. The evaluation process is divided into two phases: leaning phase and evaluation phase. First, learning data is used to estimate the elements of CHMM. Then, the battery state and the next state can be identified and predicted by using this model. At last, the simulation results prove the applicability of CHMM in the application of evaluating the health of lithium battery.

J. Guo and Z. J. Li, “

Bivariate Gamma Processes for Modeling Lithium Ion Battery Aging Mechanism

” in Proceedings of the 2017 Industrial and Systems Engineering Annual Conference, Pittsburg, PA, May 20-23 2016

[Cited within: 1]

D. Liu, H. Wang, Y. Peng, W. Xie, H. Liao , “

Satellite Lithium-Ion Battery Remaining Cycle Life Prediction with Novel Indirect Health Indicator Extraction

,” Energies, Vol. 6, No. 8, pp. 3654-3668, 2013

[Cited within: 1]

C. Chen and M. Pecht, “

Prognostics of Lithium-Ion Batteries Using Model-based and Data-Driven Methods

” in Proceedings of the IEEE 2012 Prognostics and System Health Management Conference, pp. 1-6, 2012

[Cited within: 1]

B. Saha, G. Kai, J. Christophersen , “

Comparison of Prognostic Algorithms for estimating Remaining Useful Life of Batteries

,” Transactions of the Institute of Measurement & Control, Vol. 31, No. 3, pp. 293-308, 2009

[Cited within: 1]

L. X. Liao, F. Kottig , “

A hybrid framework combining data-driven and model-based methods for system remaining useful life prediction

,” Applied Soft Computing, 44. C(2016):191-199.

[Cited within: 1]

M. Einhorn, W. Roessler, J. Fleig , “

Improved Performance of Serially Connected Li-Ion Batteries with Active Cell Balancing in Electric Vehicles

,” IEEE Transactions on Vehicular Technology, Vol. 60, No. 6, pp. 2448-2457, 2011

[Cited within: 2]

F. Baronti, G. Fantechi, E. Leonardi, R. Roncella , “

Hierarchical Platform for Monitoring, Managing and Charge Balancing of LiPo Batteries

,” in Proceedings of IEEE Vehicle Power and Propulsion Conference, pp. 1-6, 2011

[Cited within: 2]

J. Yan, Z. Cheng, G. Xu, H. Qian , “

Fuzzy Control for Battery Equalization based on State of Charge

,” in Proceedings of IEEE Vehicular Technology Conference Fall, Vol. 45, pp. 1-7, 2009

[Cited within: 1]

J. Arai, A. Matsuo, T. Fujisaki, K. Ozawa , “

A Novel High Temperature Stable Lithium Salt (Li2B12F12) for Lithium Ion Batteries

,” Journal of Power Sources, Vol. 193, No. 2, pp. 851-854, 2009

D. Y. Jung, B.H. Lee, W.K. Sun , “

Development of battery management system for nickel-metal hydride batteries in electric vehicle applications

,” Journal of Power Sources, 109(1), 1-10, 2002

[Cited within: 1]

C. Jiang, Z. G. Tang, J. X. Hao, H. Q. Li , “

A novel design of air-cooled battery thermal management system for electric vehicle

,” Applied Mechanics & Materials , 563, 362-365, 2014

[Cited within: 1]

L. Fan, J. M. Khodadadi and A. A. Pesaran, “

A parametric study on thermal management of an air-cooled lithium-ion battery module for plug-in hybrid electric vehicles

,” Journal of Power Sources, 238, 301-312, 2013

[Cited within: 1]

A. A. Pesaran , “

Battery thermal management in evs and hevs: issues and solutions

”. Battery Man, 2001

[Cited within: 2]

M. Zolot, A. A. Pesaran, M. Mihalic , “

Thermal Evaluation of Toyota Prius Battery Pack

,” Future Car Congress, 2002

[Cited within: 1]

R. Mahamud, C. Park , “

Reciprocating air flow for li-ion battery thermal management to improve temperature uniformity

,” Journal of Power Sources, 196(13), 5685-5696, 2011

[Cited within: 1]

act: A Novel Thermal Management System for Electric Vehicle Batteries Using Phase-Change Material

,” Cheminform, Vol. 32, No. 1, 2001

[Cited within: 1]

R. Sabbah, R. Kizilel, J. R. Selman, S. Al-Hallaj , “

Active (Air-Cooled) vs. Passive (Phase Change Material) Thermal Management of High Power Lithium-Ion Packs: Limitation of Temperature Rise and Uniformity of Temperature Distribution

,” Journal of Power Sources, Vol. 182, No. 2, pp. 630-638, 2008

[Cited within: 1]

/