Battery Cell Quality Testing With EIS Filter Mask Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for assessing battery cell quality, particularly lithium-ion batteries, are time-consuming and resource-intensive, as they require measuring electrochemical impedance spectroscopy over a wide frequency range to predict capacity and service life, with existing models failing to represent all observed characteristics and identify new relationships.
Innovation Solution
A quality test system utilizing a filter mask and quality model created through machine learning methods to select a reduced frequency range for analysis, optimizing the quality model to calculate features like charge storage capacity, service life, and self-discharge rate, using techniques such as auto-encoders and dimension reduction to transform complex impedance spectra into meaningful data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If electrochemical impedance spectroscopy is performed over the entire frequency range to predict battery cell quality, then measurement precision is improved, but measurement time increases significantly
Solution Approach 1:
The patent divides the entire frequency range into multiple sub-ranges and uses a filter mask to selectively focus on specific sub-ranges that are most relevant for predicting particular quality features. This segmentation allows the system to capture essential quality information without measuring the complete frequency spectrum, thereby reducing measurement time while maintaining prediction accuracy.
Solution Approach 2:
The patent dynamically adjusts the frequency range parameters based on the specific quality feature being predicted. The filter mask is configured to emphasize frequency sub-ranges that have been identified as most correlated with the target quality parameter, allowing the system to adapt the measurement parameters to the specific prediction task and reduce unnecessary measurements.
2Loss of information
If statistical data analysis and spectral analysis methods are used to derive features from impedance spectra, then existing quality relationships are identified, but new relationships cannot be discovered and all frequency data must be measured
Solution Approach 1:
The patent introduces a filter mask as an intermediary component that processes the impedance spectra before quality prediction. This filter mask acts as a learned selector that identifies and emphasizes the most informative frequency sub-ranges, enabling the system to discover new quality relationships while reducing the amount of data that needs to be processed.
Solution Approach 2:
The system performs preliminary analysis to identify which frequency sub-ranges are most relevant for quality prediction before conducting full measurements. The filter mask is pre-configured or adaptively trained to recognize important frequency regions, allowing subsequent measurements to focus only on these critical sub-ranges and avoid unnecessary data collection.
3Reliability
If regression models are trained using impedance spectra from stress tests to predict quality features, then prediction capability is improved, but the knowledge of which EIS characteristics influence quality remains unresolved
Solution Approach 1:
The filter mask serves as an interpretable intermediary that bridges the gap between complex impedance spectra and quality predictions. By explicitly showing which frequency sub-ranges are most important for each quality feature, the filter mask provides interpretability while maintaining prediction reliability, allowing users to understand which EIS characteristics influence quality without dealing with the full complexity of the original spectra.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach significantly reduces the time and resources needed for battery cell quality assessment, allowing for more efficient and scalable production by focusing on only the relevant frequency ranges, enhancing the precision and speed of quality feature determination.
Implementation Method 1
A more precise and much quicker method for assessing the battery cell quality may use so-called electrochemical impedance spectroscopy (EIS). This measures the response, in the form of a signal strength or pulse response, of a battery cell to dedicated current or voltage excitations over a wide frequency range
Data Source
AI summary
Various embodiments include a method for producing a quality test system executing a quality test model with a filter mask and a quality model to determine a quality feature of a battery cell. The system has an electrochemical impedance spectroscopic unit for capturing test data relating to the battery within a frequency range. The method includes: creating the model; and producing the system. Creating the model includes: capturing spectroscopic learning data; creating the filter mask using a first machine learning method with analysis data from part of the frequency range by consulting the filter mask and creating the model using a second machine learning method. The first and the second learning method are coupled based on the learning data. The first machine learning method creates a filter mask determining the analysis data such that the second machine learning method creates a quality model optimized with respect to maximizing the quality.