1.3 Motivation of the research described in this thesis
As described in section 1.1, more efficient use of the energy inside a battery is becoming increasingly important in the rapidly growing market for portable products. Manufacturers of portable devices are consequently paying even more attention to battery management. This is reflected in many commercial electronics magazines, such as Electronic Design and EDN, containing examples of implementations of battery management functions in a system. Many examples are to be found of ICs that implement certain charge algorithms [4]-[7]. Adding intelligence to batteries in portable products to enable e.g. SoC monitoring is also receiving a great deal of attention. The term ‘smart battery’ is a general buzzword that pops up in many articles [5],[8]-[10]. However, no explanation of battery behaviour is given in this kind of magazines. Therefore, the reason why one battery management IC performs better than another is often not understood. Moreover, it is hard to determine how the functionality of a BMS can be improved.
Besides in the magazines mentioned above, a lot of information on battery management can also be found in the literature. For example, much attention is paid to finding ways of accurately determining a battery’s SoC [11]-[13]. The battery management functions described in these articles are derived from extensive battery measurements, for example measurements of battery discharge curves under various conditions. Most of these measurements are very time-consuming. The conclusions are moreover often empirical.
In practice, battery management functions are implemented in portable devices by electrical engineers. These engineers usually treat the battery as a black box. It is usually assumed that the battery is a voltage source with some series resistance. However, in order to improve the functionality of a BMS, at least some understanding of battery behaviour in the system is needed. A prototype of (part of) the portable device is needed for measurements of actual battery charge and discharge behaviour. On the other hand, simulation is a helpful tool in obtaining a better understanding of the behaviour of complex systems under a wide variety of conditions. Simulations take less time than measurements and no prototype is needed. Therefore, the availability of simulation models for batteries would be very helpful for the development of BMS.
Simulation models of batteries are given a lot of attention in the literature. However, some models have been constructed by researchers with an electrochemical background and are very complex and based on many mathematical equations [14],[15]. Consequently, such models are usually not very suitable for electrical engineers who have to simulate a battery in a system. Other models have
Introduction 5
been constructed by electrical engineers [16]. These models can be readily used in system simulations, but do not lead to a better understanding of battery behaviour. Therefore, simulations with these models do usually not lead to better views on battery management.
This thesis presents the results of research into battery modelling obtained by combining the expertise of electrical engineering with that of electrochemistry. The result is a method for modelling batteries that can be readily applied to all kinds of batteries [17]-[24]. The models result from translating (electro)chemical theory into equivalent-network models using the principle of physical system dynamics [25]. This enables the use of conventional electronic-circuit simulators that electrical engineers are accustomed to. In addition to the battery voltage and current, the internal gas pressure and battery temperature can also be simulated integrally and coherently under a wide variety of conditions. Apart from the electrical interaction with the battery’s environment, it is also possible to simulate the thermal interaction. The modelling method even allows the simulation of effects such as venting under high-pressure build-up inside the battery.
Obviously, close quantitative agreement between the results of simulations using a battery model and measurements of battery behaviour is important. Part of the research described in this thesis is devoted to optimizing this quantitative agreement. The results of this research will be described in chapter 4.
As is revealed by the title of this thesis, the derived battery models will be applied in the design of Battery Management Systems. The battery models can be viewed as ‘transparent’, in which the course of the various reactions can be investigated. For example, the charging efficiency can be easily plotted. Charging efficiencies of different charging algorithms can be compared for optimization. An example of an optimized charging algorithm that was found in simulations with a battery model will be described in chapter 5 [24].
Internal battery behaviour that is normally hard to measure can be visualized with the models. For example, overpotentials of all reactions inside a battery and electrode equilibrium potentials can be easily plotted during the battery’s operation. This makes the model a very useful tool in the quest for more accurate SoC indication algorithms. The model can be used to gain insight into the combined action of the various processes taking place inside a battery. Based on this insight, compact descriptions of battery behaviour can be derived. The results of research efforts in this field will be described in chapter 6.
Sometimes a designer is not interested in what goes on inside a battery, for example when simulating the run time of a portable device. In such cases the designer is merely interested in the battery’s discharge behaviour under various load conditions. Part of the research described in this thesis is aimed at finding an optimum method for powering a PA inside a cellular phone. As will be shown in chapter 7, the battery models described in this thesis offer a simple way of comparing the run times of various PA supply strategies in a cellular phone. In addition to battery models, use will be made of simulation models of a DC/DC converter and a PA to design the BMS.
1.4 Scope of this thesis
Chapters 2 and 3 provide general information required as a background in the remaining chapters of this thesis. Chapter 2 describes the various parts of a BMS and their functionality in more detail. Some examples of BMS in different portable products are given to clarify the influence of several factors on the complexity.

6 Chapter 1
Chapter 3 deals with the central part of a BMS, which is the battery itself. Some general information is given on the construction, types, operational mechanism and behaviour of batteries.
The research results are described in chapters 4 to 7. Battery modelling forms the core of this thesis. The construction of simulation models for rechargeable batteries is described in chapter 4. Those who want to have a thorough understanding of the background and construction of the models should read this chapter. The adopted modelling approach is explained in detail and the model equations for both a NiCd and a Li-ion battery model are derived. Further, the efforts to improve the quantitative agreement between the results of simulations and measurements are thoroughly discussed for the NiCd model. Chapters 5, 6 and 7 deal with the design of BMS. These chapters describe the use of the battery models of chapter 4 and other models to find improved BMS schemes. One does not have to read chapter 4 first in order to understand the content of these chapters.
Battery charging algorithms are discussed in chapter 5. It is shown that battery models can readily be used in the development of new, more efficient charging algorithms. Chapter 6 deals with the determination of a battery’s SoC. Several possible methods are compared. In addition, a new SoC indication system is proposed and tested, based on simulations using a battery model and knowledge obtained in battery measurements. An optimum method for efficiently powering PAs in cellular phones is described in chapter 7. This strategy is named efficiency control [26]. As a DC/DC converter is necessary to implement efficiency control, some basic information on voltage conversion techniques is given. Measurement results are discussed to define the benefits of efficiency control in practice and to make a comparison with simulation results. Conclusions are drawn in chapter 8 and recommendations are made for further research in the exciting field of BMS in general and battery modelling in particular.