.png){kind=icon}
.png){kind=icon}
Return
State of Charge Estimation using a Nonlinear Autoregressive with Exogenous Input Neural Network
.png){kind=icon}
PRE
AI
P
F
J
E
S
DOI：10.1109/POWERTECH55446.2023.10202982.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Among different methods that can be used to estimate the State of Charge of a battery in a dynamic profile, the data-driven method based on Neural Network (NN) is one of the most promising. This work investigates for dynamic automotive applications a time-recurrent Neural Network, such as Nonlinear Autoregressive with Exogenous Input Neural Network, to analyze its performance using two different battery datasets. Initially, tests that comprehend current, voltage, and temperature as input data have been performed. In the same step, a sensitivity analysis for three different timesteps has been investigated. In the following step, interpolation and extrapolation processes run on different battery temperature cycles have been considered. The third phase of this work introduced noise in input data to replicate a real-life case with its problems and finally, in the last step, it has been provided a generalization of the process, where the test datasets were different from the training ones. Throughout all the steps, NARX-NN shows good performances together with relatively low computational time.
Keywords:
Bayesian Optimization
Lithium-Ion Battery
Machine Learning
Neural Network
Nonlinear AutoRegressive with eXogenous input
NARX
Noise
State of Charge
Journal
I
IF:
0
Papers:
29
Citations:
0
Organization
Cited Papers