Battery Pack Sorting Using Non-Cycling Classification Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for testing microbatteries are time-consuming and can damage the batteries, making them unsuitable for industrial applications, especially for 'Li-free' microbatteries, as they require complete cycling and are sensitive to humidity, leading to premature aging and rendering tested batteries non-marketable.
Innovation Solution
A rapid testing method using automatic learning to classify batteries as functional or defective without cycling, employing a classification model trained on measurements such as open circuit voltage, current profiles, and electrochemical impedance spectroscopy, allowing for stress similar to final use without degrading the batteries.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If complete charge/discharge cycling is performed to sort microbatteries, then functional reliability is improved, but testing time increases significantly and batteries are damaged
Solution Approach 1:
The patent applies preliminary action by performing a partial charge (formation charge) before sorting, which prepares the battery in a controlled state without completing full cycling. This preliminary step allows subsequent rapid classification based on voltage relaxation characteristics, achieving reliable sorting in seconds rather than hours or days of complete cycling
Solution Approach 2:
The patent uses partial action by applying a formation charge of only 10-50 μAh (partial charging) instead of complete cycling. This partial charge is sufficient to activate the battery and elicit measurable voltage relaxation responses that reveal functional characteristics, enabling rapid sorting without the time-consuming and damaging effect of full charge/discharge cycles
2Reliability
If complete cycling is performed for sorting, then functional defects are detected, but battery components age and interfaces degrade
Solution Approach 1:
The method performs a preliminary formation charge to activate the battery and reveal functional defects through voltage relaxation analysis, avoiding the need for complete cycling that would cause aging. This preliminary step is sufficient to detect defects while preserving battery lifespan for subsequent market use
Solution Approach 2:
The patent treats the formation charge as a disposable, low-cost activation step that consumes minimal energy (10-50 μAh) and causes no lasting damage. This contrasts with complete cycling which permanently degrades the battery. The formation charge serves its purpose of revealing defects and can be discarded without affecting the battery's operational life
3Productivity
If rapid testing methods are used to reduce testing time, then productivity is improved, but measurement precision and reliability decrease
Solution Approach 1:
The patent uses feedback by measuring the voltage relaxation response after formation charge and comparing it against reference profiles of known good and defective batteries. This feedback mechanism enables rapid classification with high accuracy, achieving both high productivity (seconds per battery) and high measurement precision through pattern recognition of the relaxation curves
Solution Approach 2:
The method applies partial action by using only the voltage relaxation response following a brief formation charge, rather than waiting for complete cycling. This partial measurement approach captures sufficient information about battery functionality to enable accurate sorting, achieving high precision with minimal testing time and maximum productivity
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
Enables quick sorting of microbatteries without damaging them, eliminating the need for threshold definitions and accommodating multiple dimensions, thus reducing testing time and ensuring functional components are not mistakenly eliminated.
Implementation Method 1
a battery, in other words an electrochemical device capable of storing electrical energy
Implementation Method 2
the ions (for example lithium ions, Li+) migrate between the negative and positive electrodes
Data Source
Figure 1
Figure 2~3
Figure 4
AI summary
A computer-implemented method (101) for machine learning a battery classification model into two categories: functional or defective, the method comprising the steps of: - Acquiring (110), on a set of batteries of the same type, a group of measurements characteristic of the operation of a battery, - Carrying out (111) a complete cycle of each battery of said set and measuring at least one curve from a charging curve or a discharging curve of the battery, - Determining (111), for each battery, a label of belonging to the functional or defective category by comparing at least one measured curve to a reference curve characterizing the proper operation of the battery, and - Carrying out (112) a supervised training of a battery classification model according to the two categories from said groups of measurements and said label of belonging to one of the two categories.