In this paper, a model based sensor fault detection strategy is proposed for series connected Li-Ion cells. A state space model of the equivalent circuit of each cell is first derived and a Luenberger observer is designed for residual generation, which plays a crucial role in the fault detection process. Since the residual is influenced by the fault's magnitude, it may result in subtle changes that are challenging to identify. Therefore, a Quickest Change Detection (QCD) approach is presented to further process the residual and ensure sensor malfunctions are not missed. The Cumulative Sum algorithm is then used to solve the QCD problem. Simulation results are then presented to verify the feasibility of the proposed method.
Model Based Change Detection Approach For Sensor Fault Identification in Battery Packs
2023-08-28
1861005 byte
Conference paper
Electronic Resource
English
A correlation based fault detection method for short circuits in battery packs
Online Contents | 2017
|METHODS AND CIRCUITRY FOR FAULT DETECTION AND AUTOMATIC EQUALIZERS FOR BATTERY PACKS
European Patent Office | 2019
|