Show simple item record

dc.contributor.advisorPlett, Gregory L.
dc.contributor.authorSmiley, Adam J.
dc.date.accessioned2019-05-15T21:46:20Z
dc.date.available2019-05-15T21:46:20Z
dc.date.submitted2019-05
dc.identifierSmiley_uccs_0892D_10471.pdf
dc.identifier.urihttps://hdl.handle.net/10976/167104
dc.descriptionIncludes bibliographical references.
dc.description.abstractLithium-ion batteries provide unmatched volumetric and gravimetric energy density as well as extremely low self-discharge rates, but require the use of sophisticated battery management systems to prevent the rapid (and in some cases, catastrophic) progression of aging processes that reduce the energy storage capacity and power delivery capabilities of the battery. Physics-based models have been developed that enable battery management systems to slow the advancement of aging processes while also enabling improved battery performance over what is available via equivalent-circuit cell models. The superior control enabled by physics-based models is conditional on the accuracy of the parameterization of the physics-based model, which must change as the cell ages to account for the evolution of battery behavior due to chemical processes that will occur regardless of the control method. In previous work, we have developed the interacting multiple-model Kalman filter method for ensuring accurate parameters over time via a “parameter-selection” scheme, where the most appropriate set of parameter values is selected from a collection of possible “pre-aged” sets of parameter values. In this work several modifications to the “pre-aging” and parameter selection process are explored to improve the ability to make correct age estimates while simultaneously reducing computational cost. Several approaches to improvement are explored. In the first approach, we consider how the Viterbi and BCJR algorithms commonly employed for communication systems can be used to post-process the output of the baseline multiple-model algorithm. In the second approach, the “pre-aging” process is altered such that resistance modeling in the positive electrode is in better agreement with experimental parameter values available in the literature. In the third approach, the Kalman filters utilized by the interacting multiple-model algorithm are altered such that we realize a reduction in computation without sacrificing estimation accuracy. In the final approach, a variable structure interacting multiple-model Kalman filter is implemented, which dynamically selects an optimal subset from the collection of possible “pre-aged” parameters for consideration. As part of this approach, we consider how to choose the models included in the model space based on a discernibility criterion.
dc.languageEnglish
dc.language.isoeng
dc.publisherUniversity of Colorado Colorado Springs. Kraemer Family Library
dc.rightsCopyright of the original work is retained by the author.
dc.subjectLithium-ion batteries
dc.subjectKalman filtering
dc.subjectMultiple-model state estimation
dc.titleImproved Approach to State-of-Age Estimation for Lithium-Ion Battery Cells Using Interacting Multiple Model Kalman Filters, An
dc.typeThesis
dc.contributor.committeememberTrimboli, M. Scott
dc.contributor.committeememberWickert, Mark
dc.contributor.committeememberCascaval, Radu
dc.contributor.committeememberShi, Ying
thesis.degree.nameDoctor of Philosophy (Ph.D.)
thesis.degree.levelDoctoral
thesis.degree.disciplineCollege of Engineering and Applied Science–Electrical Engineering
thesis.degree.grantorUniversity of Colorado Colorado Springs


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record