Battery Charge Profile Segmentation Using ML Section Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional charging/discharging equipment lacks the ability to accurately segment and allocate identification numbers to charge/discharge control sections in battery activation processes, making it difficult to diagnose and analyze battery cells effectively.
Innovation Solution
A method and apparatus that utilize a machine learning model, such as a decision tree, to automatically assign section classification information to time-series data of battery charge/discharge profiles during the activation process, enhancing segmentation and diagnosis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional charging/discharging equipment is used for battery activation, then the basic charge/discharge control can be performed, but the equipment cannot accurately segment and allocate identification numbers to charge/discharge control sections
Solution Approach 1:
The patent introduces a profile analysis apparatus as an intermediary system that receives charge/discharge profiles from conventional equipment and performs segmentation analysis. This mediator contains a database with reference profiles and a processor that compares actual profiles against references to automatically allocate section identification numbers, thereby achieving precise segmentation without modifying the original charging equipment
Solution Approach 2:
The patent replaces the mechanical/manual segmentation approach with an automated information processing system. The profile analysis apparatus uses computational methods to compare charge/discharge profiles, identify characteristic points, and automatically assign identification numbers to control sections, substituting manual analysis with algorithmic processing to achieve higher precision
2Reliability
If no section classification information is allocated to charge/discharge profiles, then the equipment operation is simple, but it becomes difficult to extract features and diagnose battery cells
Solution Approach 1:
The patent performs preliminary segmentation and allocation of identification numbers to charge/discharge control sections before the actual battery diagnosis process. By pre-processing the charge/discharge profiles and organizing them into structured sections with identification numbers, the system prepares the data in advance, making subsequent feature extraction and diagnosis more reliable and efficient
Solution Approach 2:
The profile analysis apparatus acts as an intermediary between the raw charge/discharge data and the diagnosis system. It processes the raw profiles to extract structured section information and identification numbers, providing cleaned and organized data to downstream diagnosis applications, thereby improving diagnosis reliability without requiring complex integration in the diagnosis system itself
3Measurement precision
If manual segmentation setting is performed to determine product quality, then some diagnosis accuracy can be achieved, but the process is time-consuming and requires precise manual configuration
Solution Approach 1:
The patent implements a self-service segmentation system where the profile analysis apparatus automatically performs the segmentation task by comparing charge/discharge profiles against reference profiles in its database. The system autonomously identifies control sections and allocates identification numbers without requiring manual intervention, thereby achieving diagnosis precision equivalent to manual methods while eliminating the time loss associated with manual configuration
Solution Approach 2:
The system performs segmentation analysis in advance before product release decisions are made. By pre-processing the charge/discharge profiles and allocating section identification numbers ahead of time, the system prepares diagnostic information proactively, eliminating the need for time-consuming manual segmentation when diagnosis is actually needed
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Provided is a method and apparatus for battery charge/discharge profile analysis. The method includes training a machine learning model using a plurality of training charge/discharge profiles as a training dataset, wherein each training charge/discharge profile includes training section classification information, and the training section classification information is a dataset in which an identification number of any one of a plurality of charge/discharge control sections performed in a sequential order in an activation process is allocated to each time index; inputting a target charge/discharge profile acquired through the activation process of a battery cell to the machine learning model; and acquiring target section classification information for the input target charge/discharge profile from the machine learning model. The target section classification information is a dataset in which the identification number of any one of the plurality of charge/discharge control sections is allocated to each time index of the target charge/discharge profile.