Thesis - Open Access
Master of Science (MS)
Department / School
Battery state monitoring is one of the key techniques in Battery Management System (BMS). Accurate estimation can help to improve the system performance and to prolong the battery lifetime. The main challenges for the state online estimation of Li-ion batteries are the flat characteristic of open circuit voltage (OCV) with the function of the state of charge. Hence, the focus of this thesis study is to estimation of the state of charge (SOC) of Li-ion with high accuracy, more robustness. A 2nd order RC equivalent circuit model is adapted to battery model for simulation, mathematical model analysis, and dynamics characteristic of battery study. Model parameters are identified with MATLAB battery model simulation. Although with more lumped RC loaders, the model is more accurate, high computation with a higher nonlinear function of output will be. So, a discrete state space model for the battery is developed. For a complex battery model with strong nonlinearity, Sequential Monte Carlo (SMC) method can be utilized to perform the on-line SOC estimation. An SMC integrates the Bayesian learning methods with sequential importance sampling. SMC approximate the posterior density function by a set of particles with associated weights, which is developed in MATLAB environment to estimate on-line SOC. A comparison is presented with Kalman Filtering and Extended Kalman Filtering to validated estimation results with SMC. Finally, the comparison results provide that SMC method is more accurate and robust then KF and EKF. Accurately prediction of Remaining Useful Life of Li-ion batteries is necessary to reliable system operation and monitoring the BMS. An empirical model for capacity degradation has been developed based on experimentally obtained capacity fade data. A nonlinear, non-Gaussian state space model is developed for empirical model. The obtained empirical model used in Sequential Monte Carlo (SMC) framework is to update the on-line state and model parameters to make a prediction of remaining useful life of a Li-ion battery at various lifecycle.
Library of Congress Subject Headings
Lithium ion batteries -- Testing.
Includes bibliographical references (pages 65-69)
Number of Pages
South Dakota State University
Daravath, Suresh, "Nonlinear Stochastic Filtering for Online State of Charge and Remaining Useful Life Estimation of Lithium-ion Battery" (2018). Electronic Theses and Dissertations. 2421.