Battery Time-Series Synthesis for Accurate State Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing battery degradation state monitoring methods require application-specific teacher data, increasing time and processing load, and lack efficient data generation for battery evaluation indices.
Innovation Solution
An information processing device generates application-specific teacher data from common data, synthesizing time-series data to estimate battery states and calculate evaluation indices, using interpolation and extrapolation techniques to match feature information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If application-specific teacher data is prepared for each battery application, then estimation model accuracy is improved, but time consumption and processing load increase
Solution Approach 1:
The patent creates a universal common data set that can serve multiple battery applications simultaneously. Instead of preparing separate application-specific data for each use case, the system generates a single common data set that encompasses diverse battery operating conditions, allowing one data set to support multiple estimation models for different applications.
Solution Approach 2:
The patent segments the data generation process into two stages: first generating common data that represents universal battery characteristics, then creating application-specific data by combining this common data with application-specific features. This segmentation allows efficient reuse of common data across applications while maintaining application-specific accuracy.
2Measurement precision
If application-specific teacher data is prepared for each battery application, then estimation model accuracy is improved, but processing load increases
Solution Approach 1:
The patent creates a universal common data set that can serve multiple battery applications simultaneously. Instead of preparing separate application-specific data for each use case, the system generates a single common data set that encompasses diverse battery operating conditions, allowing one data set to support multiple estimation models for different applications.
Solution Approach 2:
The patent merges common data with application-specific data through a combination process. The common data set is combined with application-specific features to generate application-specific teacher data, which reduces the total processing load by reusing common data across multiple applications rather than processing entirely separate data sets for each application.
3Productivity
If common data is synthesized from multiple data sources, then data generation efficiency is improved, but data quality and relevance may deteriorate
Solution Approach 1:
The patent applies local quality by maintaining distinct characteristics for different data components. The common data set preserves universal battery characteristics while application-specific data maintains unique application features. This ensures that each data component retains its specific quality and relevance while contributing to the overall efficiency of data generation.
Solution Approach 2:
The patent performs preliminary data processing and quality validation during the common data generation phase. By pre-processing and validating data quality before combining with application-specific features, the system ensures that the final application-specific data sets maintain high quality while benefiting from the efficiency of common data reuse.
Data Source
AI summary
According to an embodiment, an information processing device includes one or more hardware processors configured to select, from a plurality of common data being time-series data indicating changes in charging and discharging of a battery and associated with mutually different feature information indicating features of the charging and discharging, a plurality of common data associated with the feature information identical to or similar to specified first feature information, and to generate first time-series data for the first feature information by synthesizing the plurality of selected common data.