OMILOLI, KOTO ANDREW and Covenant University, Theses (2022) ROBUST ESTIMATION OF THE STATE OF CHARGE IN A LITHIUM-ION BATTERY FOR A BATTERY MANAGEMENT SYSTEM. Masters thesis, COVENANT UNIVERSITY.
![[img]](http://eprints.covenantuniversity.edu.ng/style/images/fileicons/application_pdf.png) |
PDF
Download (385kB) |
Abstract
The hazardous effect fossil-based systems have on the planet requires transition to nonpolluting energy sources which lithium-ion batteries (LiBs) fall. An implication is, the state of charge (SOC), must be determined to provide an indication for the available energy left in LiBs. However, there exists the difficulty in measuring the SOC, more so existing methods for SOC estimation are not robust to accommodate battery parameter sensitivity to disturbance. Hence, in this research battery models coupled with SOC estimation techniques, namely Extended Kalman Filter (EKF), Linear Matrix Inequality (LMI) and Sliding Mode Observer (SMO), are developed and implemented to solve the observability problem. Model validation was carried out via primary data while the secondary data was used for validating the state estimators using charge/discharge voltage and current inputs. Performance results showed the LMI and SMO, yielded RMSE and MAE values equal to zero (0), offering a superior accuracy than the EKFs having RMSE values in range of [0.00000861, 0.00680] and MAE in range of [0.00000214, 0.00410]. In addition, by means of a modified priori estimate and a compensating proportional gain, an improved extended Kalman filter (EKFmod) for the estimation task was carried out. Amongst the improved estimators, the fourth order EKFmod had an accuracy of six (6) decimal places with the smallest error bound of ±2.05%. In terms of robustness, the SMO and LMI algorithms demonstrated capability in disturbance rejection from measurement input data at battery temperature of 0oC having both RMSE and MAE values of zero (0) in contrast to the EKF having lesser metrics values. Furthermore, a hybrid (EKF-SMO) estimator developed showed a 93% decrease from the EKF performance metrics and faster convergence. This research recommends the selection of battery models, and the estimators should be a trade-off between model complexity, accuracy, and present computation power.
| Item Type: | Thesis (Masters) |
|---|---|
| Uncontrolled Keywords: | Lithium-ion Battery, Extended Kalman Filter, Sliding Mode Observer, State of Charge, Linear Matrix Inequality, Robust. |
| Subjects: | T Technology > T Technology (General) T Technology > TK Electrical engineering. Electronics Nuclear engineering |
| Divisions: | Faculty of Engineering, Science and Mathematics > School of Electronics and Computer Science |
| Depositing User: | AKINWUMI |
| Date Deposited: | 11 Oct 2022 11:00 |
| Last Modified: | 11 Oct 2022 11:00 |
| URI: | http://eprints.covenantuniversity.edu.ng/id/eprint/16303 |
Actions (login required)
 |
View Item |
