hILDe

Innovative, cost-effective and highly accurate indication of imbalance and state of charge in vanadium redox flow batteries using AI-supported detection of specific colours

The hILDe project aims to research, develop and test an innovative, cost-effective and highly accurate method for indicating the state of charge (SOC) and imbalance of vanadium redox batteries (VRFB). The term imbalance refers to the imbalance in the state of charge in the anolyte and catholyte caused by the following phenomena:

Air oxidation of V²⁺ in the anolyte
Chemical and electrochemical H₂-formation in the anolyte
Permeation of vanadium ions and water through the cell membrane

The electrochemically active species of a VRFB are present in the form of aqueous electrolytes (catholyte and anolyte). These contain vanadium ions of varying valence, stability and colour in a variety of chemical complexes. During operation, the VRFB is subject to various ageing mechanisms which impair the quality of the electrolyte as an energy carrier and thus its capacity, cycle stability and the service life of the entire VRFB system. For the long-term stable operation of a VRFB, the use of a high-precision sensor/indicator is essential, which, in addition to displaying the current SOC, also provides information on the imbalance of the electrolyte. The project involves the development and functional demonstration of a SOC/imbalance indicator based on the developments, results and findings of the three subsystems:

AI-platform for electrochemical processes
hILDe sensor and
reference measurement system

The highly complex chemistry of vanadium ions, especially in anolytes, overwhelms conventional physicochemical data analysis. Therefore, the hILDe project will use AI, including machine learning and case-based reasoning, to analyse data for determining the SOC and imbalance of a VRFB. The SOC/imbalance indicator resulting from the hILDe project is intended to be highly accurate, cost-efficient and universally applicable in any configuration system. During the project, the developed hILDe overall system will be used and tested in a real VRFB under everyday conditions. To ensure broad utilisation of the results, two workshops will be held with the participants of a “VRFB expert panel”. Members of this expert panel are representatives from commercial companies and research institutions.

Result

The hILDe research project was successfully completed on 30 September 2023. The project goal of developing/researching an innovative, cost-effective and highly accurate indication of the imbalance and state of charge (SOC) of vanadium redox flow batteries (VRFB) using AI-supported detection of specific colours was successfully implemented. The hILDe sensor detects the SOC and imbalance of VRFB optically by evaluating specific wavelengths. The sensor measures the electrolyte in both half-cells (anolyte and catholyte) and then evaluates the results using an AI algorithm. The project result of IZES gGmbH was the development of the hardware for the hILDe sensor. The project partner DFKI worked on the development of the AI algorithm using machine learning and on training/optimising the AI using measurement data sets. In addition to the measurement data sets provided by IZES, data from a reference measurement system (RMS) was also used to train the AI. The RMS was developed by Saarland University.

Project partners