This content is not included in your SAE MOBILUS subscription, or you are not logged in.

A Linear Parameter Varying Combined with Divide-and-Conquer Approach to Thermal System Modeling of Battery Modules

Journal Article
2015-01-9148
ISSN: 2167-4191, e-ISSN: 2167-4205
Published May 01, 2016 by SAE International in United States
A Linear Parameter Varying Combined with Divide-and-Conquer Approach to Thermal System Modeling of Battery Modules
Sector:
Citation: Asgari, S. and Kaushik, S., "A Linear Parameter Varying Combined with Divide-and-Conquer Approach to Thermal System Modeling of Battery Modules," SAE Int. J. Alt. Power. 5(1):41-49, 2016, https://doi.org/10.4271/2015-01-9148.
Language: English

Abstract:

A linear parameter varying (LPV) reduced order model (ROM) is used to approximate the volume-averaged temperature of battery cells in one of the modules of the battery pack with varying mass flow rate of cooling fluid using uniform heat source as inputs. The ROM runs orders of magnitude faster than the original CFD model. To reduce the time it takes to generate training data, used in building LPV ROM, a divide-and-conquer approach is introduced. This is done by dividing the battery module into a series of mid-cell and end-cell units. A mid-cell unit is composed of a cooling channel sandwiched in between two half -cells. A half-cell has half as much heat capacity as a full-cell. An end-cell unit is composed of a cooling channel sandwiched in between full-cell and a half-cell. A mass flow rate distribution look-up-table is generated from a set of steady-state simulations obtained by running the full CFD model at different inlet manifold mass flow rate samples. This look-up-table is used to build a series of mid-cell and end-cell LPV ROMs that are used to approximate the thermal behavior of the battery module. The use of divide-and conquer system modeling combined with the LPV ROM will provide a good approximation to the thermal behavior of the battery module while significantly reducing the time required to generate the training data.