Particle Filtering framework for Health Monitoring of Lithium-Ion Batteries using Ampere-hour Throughput based Semi-Empirical Model

Battery health monitoring typically refers to the monitoring of its State of Health (SoH). This paper explores a method for the health monitoring of a Li-ion battery based on a previously developed capacity fade model which is a function of ampere-hour throughput. The model parameters is identified by using Levenberg-Marquardt algorithm. Next, the model is used in a particle filtering (PF) framework for health monitoring. Further, the results corresponding to the C-rate dependency on initial model parameters variations have been also presented. The test has been carried out on 18650 cylindrical cells (ICR) with a nominal capacity of 2.6Ah and a nominal voltage of 4.2V and experiments were performed using a Biologic (BCS-815) battery testing equipment.The accuracy of the results is within about ±2% which is quite encouraging.