TY  - JOUR
A1  - Bandyopadhyay, Sankhyabrata
A1  - Fabian, Matthias
A1  - Li, Kang
A1  - Sun, Tong
A1  - Grattan, Kenneth TV
Y1  - 2023///
UR  - http://dx.doi.org/10.3390/batteries9100508
N1  - © 2023 by the Authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
EP  - 16
ID  - discovery10196826
N2  - Real-time monitoring of the state of charge (SOC) of the batteries used in a wide variety of applications is becoming increasingly important, especially given the impetus by the current targets towards ?net-zero?. In this research, an advanced approach was used involving fiber Bragg grating (FBG)-based sensors that were developed and implemented for the measurement of the key parameters required to ensure optimum battery performance. In this work, one of the biggest challenges to assess (and then map) the data from the sensor system developed is tackled in order to better understand the key parameters of the battery in an efficient and improved way. It is well known that the relationship between the changes in the resonance wavelength of the FBGs used in the sensor system, arising due to change in the electrical parameters of the battery, is complex and dependent on several different factors. In this work, this effect was evaluated by coupling the sensor data to a data-driven regression model approach that was developed for the measurement of the SOC of the batteries used, and this was obtained directly and conveniently from the FBG data. In this comprehensive study, FBG-based sensors were fabricated and then installed onto the battery, which then was subjected to a range of charging?discharging cycles, following which the electrical parameters of the battery were estimated from recorded data using a black-box machine learning (ML) model. Data-driven regression algorithms were employed for the training of the black-box model. The efficiency of the estimation of the SOC of the battery from the FBG-based sensor data was found to be high, at 99.62% (R2 values of Estimated SOC and True SOC line), creating a very satisfactory result for this key measurement. Thus, the work shows the robustness of the FBG-based sensor system combined with the neural network algorithm as an effective way to evaluate the electrical parameters of the battery, which is particularly important, as no physical/electrochemical/electrical model of the system is thus required.
KW  - fiber Bragg grating; optical sensors; battery condition monitoring; state-of-charge estimation; machine learning algorithms
AV  - public
IS  - 10
PB  - MDPI
VL  - 9
TI  - Fiber-Bragg-Grating-Based Sensor System to Measure Battery State of Charge Based on a Machine Learning Model
JF  - Batteries
ER  -