Battery Capacity Estimation from CCV Waveforms Using Neural Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for estimating the full charge capacity of secondary batteries, such as the AC-IR method, are costly, require specialized equipment, and involve complex measurements and expert knowledge, limiting their accuracy and speed.
Innovation Solution
A battery management system utilizing a trained neural network that estimates full charge capacity based on CCV waveforms obtained during constant-current charging and discharging, allowing for quick and accurate estimation without the need for specialized instruments or complex calculations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the AC-IR method is used to estimate full charge capacity, then measurement precision is improved, but device complexity and cost increase
Solution Approach 1:
The patent uses image processing techniques to convert CCV waveform data into visual representations that can be analyzed by neural networks. This copying approach transforms complex electrical measurement data into a format that can be processed by standard computing devices, eliminating the need for specialized measurement equipment while maintaining estimation accuracy.
Solution Approach 2:
The patent replaces specialized electrical measurement equipment (frequency response analyzers, potentiogalvanostats) with a standard computer system equipped with a neural network. The neural network processes CCV waveform data through image processing algorithms, substituting complex mechanical/electrical measurement systems with a software-based solution that achieves comparable or superior precision.
2Measurement precision
If the AC-IR method is used to estimate full charge capacity, then measurement precision is improved, but ease of operation deteriorates
Solution Approach 1:
The neural network system performs automatic estimation of full charge capacity without requiring expert intervention. The system self-calibrates and processes measurements autonomously, eliminating the need for operators to have specialized knowledge of AC-IR methods, temperature control, or circuit constant analysis. Users simply input CCV waveform data and receive automated results.
Solution Approach 2:
The patent transforms the measurement approach by changing from direct electrical parameter analysis (requiring expert interpretation) to image processing of CCV waveforms. This parameter transformation makes the measurement process accessible to general users while maintaining high precision through the neural network's automated analysis capabilities.
3Speed
If the AC-IR method is used to estimate full charge capacity, then measurement speed is improved, but productivity deteriorates
Solution Approach 1:
The neural network system is designed to handle multiple measurement tasks simultaneously - it can process CCV waveforms from multiple batteries in batch, perform temperature compensation, and provide comprehensive battery assessments. This universal system replaces multiple specialized measurement setups, improving overall productivity while maintaining the speed benefits of AC-IR methodology.
4Ease of operation
If simplified measurement methods are used to estimate full charge capacity, then ease of operation is improved, but measurement precision deteriorates
Solution Approach 1:
The system performs preliminary processing of CCV waveform data by converting it into image formats and pre-processing the visual data before neural network analysis. This preliminary action simplifies the operational steps required by users while ensuring that the neural network receives optimally prepared data, thereby maintaining high measurement precision despite the simplified user interface.
Solution Approach 2:
The patent introduces image processing as an intermediary step between simple CCV measurement and complex neural network analysis. This intermediary transformation converts raw electrical data into visual representations that are easier to handle and process, bridging the gap between simplified operation and high-precision measurement through a intermediate data format.
Data Source
AI summary
A battery management system includes a control device and a storage. The storage stores at least one trained neural network. The trained neural network includes an input layer that accepts input data that represents a numeric value for each pixel in an image where a prescribed CCV waveform (a CCV charging waveform or a CCV discharging waveform) of a secondary battery is drawn in a region constituted of a predetermined number of pixels, and when input data is input to the input layer, the trained neural network outputs a full charge capacity of the secondary battery. The control device estimates the full charge capacity of a target battery by inputting input data obtained for the target battery into the input layer of the trained neural network.