Hybrid nanophotonic-microfluidic sensor integrated with machine learning for operando state-of-charge monitoring in vanadium flow batteries

Journal of Energy Storage. 2025. Vol. 111. Article 115349.

Vlasov V., Aleksei Y. Kuzin, Florya I., Buriak N., Chernyshev V., Golikov A., Krasnov L., Mikhailov S., Pugach M., Nasibulin A., An P., Vadim V. Kovalyuk, Gregory N. Goltsman, Gorin D.

This research presents an advanced method for measuring State of Charge (SoC) in Vanadium Redox Flow
Batteries (VRFB) using Refractive Index (RI) combined with Machine Learning (ML). The study comprised of
three primary phases: ex-situ measurements, in-operando measurements, and ML model training. Initially, tests
were conducted using a hybrid photonic-microfluidic sensor on simulated solutions mimicking specific VRFB
SoCs. Subsequently, in-operando measurements were performed during cyclic processes within an experimental
VRFB. Finally, utilizing the experimental data, an ML model was trained to accurately predict SoC by analyzing
spectral characteristics. This study illustrates the potential of RI-based VRFB SoC measurement methods over the
long term and addresses technological gaps by establishing a platform for precise SoC prediction with minimal
cycle data.

Keywords: refractive index, Photonic integrated circuit, Machine learning, Battery management system, State of Charge, Vanadium redox flow batteries