- Stage of development
-
Tested at lab scale
- Intellectual property
-
Trade Secret
- Intended collaboration
-
Licensing and/or co-development
- Contact
-
Noelia GranadoVice-presidency for Innovation and Transferngranado@icv.csic.escomercializacion@csic.es
- Reference
-
CSIC/NG/002
Additional information
#Energy
#ICT
#Energy Storage
#Batteries
#Artificial Intelligence
#Machine Learning
State of Charge and Health for Batteries using ML and Electrical Impedance Profiles
Battery Analysis Pipeline: Data processing and AI model training using in situ impedance spectroscopy vectors.
- Market need
-
The rapid growth of electric vehicles, renewable energy storage, and portable electronics has increased the demand for accurate battery monitoring. Current battery management systems rely on simplified models that struggle to capture aging, degradation, and complex electrochemical behavior, leading to uncertainty in State of Charge (SoC) and State of Health (SoH) estimation. This uncertainty negatively impacts safety and performance. Electrochemical Impedance Spectroscopy provides valuable diagnostic information that encoding physiochemical states of batteries. There is a clear need for standardized and scalable methods to transform raw impedance data into actionable insights.
- Proposed solution
-
ML-driven data processing pipeline designed specifically for Electrochemical Impedance Spectroscopy analysis. The platform automatically cleans, structures, and extracts relevant features from raw impedance data. Machine learning models are trained to accurately predict State of Charge and State of Health across unique battery chemistries and operating conditions.
This approach transforms complex impedance datasets into actionable diagnostics. The pipeline is scalable and easily integrable into existing battery management systems and research workflows.
- Competitive advantages
-
- Rapid and High prediction accuracy for SoC and SoH.
- Purpose-built for EIS data – not generic ML, optimized specifically for impedance spectroscopy.
- Scalable architecture – handles large datasets and real-time diagnostics.
- Modular design – easy integration into existing BMS and laboratory workflows.