Russian scientists have created an optical AI sensor that monitors the charge level of vanadium redox flow batteries in real time. It uses the data obtained to adjust the battery's operation and increase its service life. The development was created within the walls of NUST MISIS.
Our sensor takes into account changes in the concentrations of active components and self-calibrates during operation. It uses AI to analyze battery parameters and adjust the operating mode. As a result, the error in determining the state of charge is reduced, since the influence of side reactions and the accumulation of decomposition products in the electrolyte is eliminated.
Vanadium redox flow batteries are one of the most promising methods of energy storage, which surpasses existing lithium batteries in many parameters. Their use is limited by the fact that developers have not yet had the required methods for accurately determining the charge level.
Alexey Kuzin, an employee of the Photonic Gas Sensors Laboratory at NUST MISIS, explained that it is difficult to accurately calculate the remaining charge in vanadium batteries. According to him, traditional methods, such as coulometry or measuring the voltage in an open circuit, can accumulate errors with each charge and discharge.
Russian scientists drew attention to significant changes in the refractive index of the electrolyte during operation. Based on this idea, they created a nanochip with microchannels and a photonic ring resonator that constantly monitors the optical properties of the electrolyte associated with changes in the concentration of vanadium ions.
The data collected by the chip is processed using a machine learning system that determines the charge level and adjusts the battery operation if necessary. Experiments have confirmed the high accuracy of the system.
