Monitoring State-of-charge of Lithium-Ion Battery with Diverse Series- Parallel Configuration using Particle Filter

Abstract

Lithium-ion battery packs are typically assembled into modules. These modules can be equipped with various series-parallel configurations [1]. SoC estimation is of critical importance to the reliability and safety of a battery pack. SoC Estimation for a battery pack assembly is of great interest [2]. Our work presents a method for estimating SOC for multi-cell assemblies [3]. In this work a feasible method for monitoring SoC is demonstrated by designing a circuit model in Simulink of three different configurations 3s2p (three in series and two in parallel), 3s3p (three in series and three in parallel) and 4s2p (four in series and two in parallel) using Particle Filter. In this model, a current source is attached to the battery in order to maintain and control the circuit balance which is further attached to the subsystem having SoC, previous state of SoC, SoC minimum, charge current and discharge current as input, working on a simple logic; which is connected to the bus selector for respective output in scope. Initially, lithium-ion cells are connected in 3s2p configuration and then we compared our results by adding one cells in parallel (3s3p) and then in series (4s2p) with the estimated SoC of 3s2p configuration. All the three configurations have been simulated in 0.5C and 1C discharging rate with 0.5C charging rate respectively. The circuit parameters are taken as function of SOC, for model accuracy, where PFs algorithm are used for proper estimation of SOC [4]. Rated capacity is kept at 40Ah, nominal voltage of 12.6V, temperature at 25⁰C, discharge current as 20A for 0.5C discharging rate and 40A for 1C discharging rate with 20A of charge current. Current, Voltage and SOC plot for different configuration has been evaluated. By comparing 3s2p and 4s2p configuration (Fig.1), it is observed that the SoC of both the configurations has slight difference in their charging and discharging time, that is, they almost overlap each other with the slope of 0.0133. Upon comparing 3s2p and 3s3p configuration (as shown in Fig.2), we observe the discharging and charging time for 3s3p is increasing. The SoC of the individual configuration does not coincide each other, with the slope of 0.0133 at discharging of 80% and charging of 100%. Upon monitoring we observe that connecting more cells in series, has less contribution towards change in SoC, whereas in parallel connection, there is a difference in the state of charge depending on charge and discharge rates. Simulations has been carried out in T=20000, a complex current pattern is used in simulation to verify PF's performance. Based on given SOC, the discharging pattern is commonly changed in short time battery current (Fig.3). Thereby, changing the position of the cell effects the SOC of the battery packs, can be effectively monitored using Particle filter by setting number of particles, that is M as 1000. As lithium ions batteries have a great importance and widely used in Electrical vehicles because of their power efficiency, high charging rate, more lifetime, so proper monitoring of state of charge plays a vital role in Battery Management System [5].