Combined State of Charge and State of Health Estimation for Lithium-ion Batteries

Open Access
- Author:
- Sharma, Aabhas
- Graduate Program:
- Mechanical Engineering
- Degree:
- Master of Science
- Document Type:
- Master Thesis
- Date of Defense:
- July 16, 2012
- Committee Members:
- Hosam Kadry Fathy, Thesis Advisor/Co-Advisor
- Keywords:
- State of charge
state of health
lithium ion batteries
estimation
engineering
online prediction - Abstract:
- This thesis examines the challenges faced in combined State of Charge(SoC) - State of Health(SoH) estimation for lithium ion (Li-ion) batteries. To monitor battery state (both long term and short term), it is imperative to analyze the factors affecting the errors in estimation of SoC and health parameters such as capacity and internal resistance. A brief survey of the literature is done for the state-of-the-art in battery state estimation and control with respect to the above stated parameters. Estimation of SoC, capacity and resistance over an extended period of use is studied using a first order equivalent circuit model to demonstrate the fundamental challenges faced in simultaneous estimation. Identifiability of system parameters pertaining to persistence of excitation (PE) in input signal and flatness of open circuit voltage(OCV) vs. SoC curve is studied using an Extended Kalman Filter,least squares estimation algorithm and Cram\'{e}r-Rao bounds. Mathematical derivations supported by simulation studies are done to observe the effect of individual factors including noise levels, PE dither, SoC ranges, time and initial conditions on estimation error. A trade-off discussion supported by Monte-Carlo simulations is finally presented to explain the influence and importance of each factor pertinent to SoC-SoH diagnostics, prognostics and control.The results suggest that battery SoC-SoH parameter identifiability is dependent on: (1) Higher regions of the OCV-SoC curve which give better capacity estimates (2) PE dither which helps gain accuracy in resistance estimation.