Parameters Optimization of Lebesgue Sampling-based Fault Diagnosis and Prognosis with application to Li-ion Batteries

Wuzhao Yan, Bin Zhang, and Marcos Orchard
Submission Type: 
Full Paper
phmc_16_001.pdf758.13 KBSeptember 7, 2016 - 12:12am

Lebesgue sampling-based fault diagnosis and prognosis (LS-
FDP) is developed with the advantage of less computation
requirement and smaller uncertainty accumulation compared
with the traditional Riemann sampling-based FDP (RS-FDP).
The accuracy and precision of LS-FDP are influenced by the
diagnostic and prognostic models, and need to be studied.
The propagation of the fault could show significant behavior
compared with the model due to the imperfect of the model
and the variance of the system mechanism. The predicted re-
sults can’t represent the real remaining useful life (RUL) in
all application cases, though a random noise term is included
in the model. The parameters in the models are treated as
time-varying and unknown, they need to be adjusted online
to accommodate the variance of mechanism, operation con-
dition, and environment in the real cases. In this paper, a
recursive least square (RLS) based method with a forgetting
factor is employed to optimize the diagnostic and prognostic
models in LS-FDP. The design and implementation of opti-
mization of LS-FDP based on a particle filtering algorithm
are illustrated with experimental results on Li-ion batteries
to verify the performances of the proposed approach. The
experimental results show that the accuracy of results of op-
timized LS-FDP is improved on current state estimation and
long term prediction.

Publication Year: 
Publication Volume: 
Publication Control Number: 
Submission Topic Areas: 
Model-based methods for fault detection, diagnostics, and prognosis
Submitted by: 

follow us

PHM Society on Facebook Follow PHM Society on Twitter PHM Society on LinkedIn PHM Society RSS News Feed