State–of–Charge Estimation Enhancing of Lithium batteries through a Temperature–Dependent Cell Model F. Baronti, G. Fantechi, L. Fanucci, E. Leonardi, R. Roncella, R. Saletti, S. Saponara Dipartimento di Ingegneria dell’Informazione: Elettronica, Informatica, Telecomunicazioni University of Pisa, Via G. Caruso 16, 56126 Pisa (Italy) f.baronti@iet.unipi.it Abstract—State–of–Charge estimation is one of the most important task of Battery Management Systems in hybrid and electric vehicles. Knowing the amount of charge stored in each cell of the vehicle battery pack is indeed crucial for effective battery utilisation that prevents cells from damaging and extends the battery lifetime. This is even more relevant for lithium batteries that are less tolerant to overcharging and deep discharging. However, State–of–Charge estimation is a difficult task to be performed online in a vehicle. This is because of the noisy and low accurate measurements and the wide operating conditions in which the vehicle battery can operate. This paper shows that the use of temperature–dependent cell model can improve State–of– Charge estimation, as temperature changes dramatically affect lithium battery behaviour. Experimental evidence is provided using the developed temperature–dependent cell model of Lithium-polimer battery inside a promis- ing State–of–Charge estimator that mixes the standard Coulomb counting, i.e. battery current integration, with the model–based approach. I. I NTRODUCTION Given the high performances in terms of cycle life and especially energy and power specific density, Lithium Ion (Li+) and Lithium-Polymer (LiPo) bat- teries are becoming very promising for use in hybrid (HEV) and electric (EV) vehicles. However, lithium- based batteries require electronic Battery Management Systems (BMSs) to improve the battery lifetime and to guarantee that each cell of the battery pack works in its safe area. Overcharging and deep discharging can indeed damage the cell shortening its life time or even leading to catastrophic failures. BMS should also prevent lithium batteries from working outside their operating temperature range [1]. Thus BMS is a crucial component of HEVs and EVs for safe and reliable operation of the vehicle bat- tery [2]. This objective requires the precise estimation of the amount of charge stored in each cell of the battery pack, namely the State–of–Charge. State–of– Charge is linked to the power that the battery can deliver, as well as, to the residual energy stored in the battery. Consequently, its knowledge is very important to manage the different power sources in HEVs and to predict the runtime of EVs. In addition, State–of– Charge estimation is the basis for implementing charge balancing strategies, which lead to a better utilisation of the battery pack extending its lifetime. State–of–Charge estimation inside a BMS is a very challenging task [3]. This is because the State–of– Charge estimator has to rely on onboard current and voltage measurements, which are typically noisy and inaccurate. In addition, the computational resources required by the estimator should be suitable for its implementation in an embedded system. On the other hand, the estimator should provide a precise State– of–Charge estimate in every possible battery operating condition. This requires that the estimator takes into account the actual cell temperature, as it dramatically affects the cell performance [4]–[6]. Nonetheless, this dependence is usually neglected in State–of–Charge estimation, as in [7], [8]. The objective of this paper is to show that the use of a cell model including thermal effects inside a State–of–Charge mix estimator [7] leads to better performances. The enhanced mix estimation algorithm combines State–of–Charge estimate simply obtained by integrating the battery current with the information provided by the temperature–dependent cell model reported in [5], [6]. The paper is organised as fol- lows. Section II reviews the most interesting State– of–Charge estimation methods. Section III describes the enhanced mix algorithm used in this work and briefly recalls the temperature–dependent cell model used. Experimental results are reported in Section IV. Finally, conclusions are drawn in Section V. II. SOCESTIMATION State–of–Charge is commonly defined as the ratio of the charge amount present in a battery cell at a certain time to its nominal capacity, i.e. the amount of charge that can be stored in the cell when it is fully charged. In fact, this is true only at the beginning of the battery life, as the actual cell capacity gradually decreases with time. This aspect has to be considered when the State–of–Charge is used to obtain the range of an EV vehicle. Assuming that the initial value SoC(0) is known, State–of–Charge can be evaluated by integrating the battery current over time, as shown in the following equation: SoC(t)= SoC(0) - 1 C  t 0 I cell (τ ) dτ , (1)