Battery SOH Estimation Using Segmented Partial-Charge Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current battery state estimation methods for electric vehicles lack accuracy and convenience in monitoring the health and remaining capacity of batteries, particularly during partial charging and discharging, which can lead to degradation patterns being difficult to recognize.
Innovation Solution
A battery state estimation apparatus that acquires sensing data including voltage, current, temperature, pressure, and humidity data, divides it into segments based on a predetermined time interval, generates segment vectors, maps them to a lower-dimensional feature space, and applies these vectors to a learning model to estimate the state of health (SOH) and remaining capacity of the battery, using pre-estimated information and learning parameters for improved accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If battery state estimation is performed using traditional methods, then the monitoring function is provided, but the estimation accuracy is insufficient particularly during partial charging and discharging
Solution Approach 1:
The patent segments the battery sensing data into multiple fixed-time segments, extracting key features from each segment to form a comprehensive feature vector. This segmentation approach enables accurate capture of degradation patterns during partial charging and discharging cycles, resolving the contradiction between estimation accuracy and reliability under varying charge conditions.
Solution Approach 2:
The patent transforms the segmented battery data into a multi-dimensional feature space by extracting multiple features (mean, standard deviation, maximum, minimum values) from voltage, current, and temperature across different time segments. This dimensional transformation enhances the ability to distinguish degradation patterns, improving both estimation accuracy and monitoring reliability simultaneously.
2Measurement precision
If detailed sensing data is collected and processed, then estimation accuracy improves, but processing complexity increases
Solution Approach 1:
The patent extracts essential features (mean, standard deviation, maximum, minimum) from each segmented portion of the sensing data, discarding redundant information while retaining the most informative characteristics. This extraction process maintains high estimation precision by focusing on key degradation indicators while significantly reducing processing complexity compared to analyzing complete raw datasets.
3Ease of operation
If real-time monitoring is implemented, then user convenience is enhanced, but computational load increases
Solution Approach 1:
The patent performs preliminary segmentation and feature extraction on the sensing data, organizing it into structured feature vectors before applying the learning model. This preliminary processing reduces the computational burden during real-time SOH estimation, enabling continuous monitoring with reduced energy consumption by pre-processing data into compact, informative representations.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A battery state estimation apparatus includes a sensing data acquirer configured to acquire sensing data of a battery. The apparatus further includes a state of health (SOH) information estimator configured to divide the sensing data based on a predetermined time interval, extract a segment comprising a predetermined number of items of the divided sensing data, and estimate SOH information of the battery based on the segments.