Battery State Estimation Using Adaptive Time Intervals and LSTM
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing battery state estimation methods are ineffective in accurately determining battery state information regardless of charging speed, leading to inefficiencies in both fast and slow charging scenarios.
Innovation Solution
A processor-implemented method and apparatus that acquire current and voltage information, determine time interval values based on current integration, and use a recurrent neural network (RNN) with long short-term memory (LSTM) to generate feature vectors and estimate battery state, ensuring consistent feature vector length across charging speeds.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional battery state estimation methods are used, then the system is simple to implement, but the accuracy of battery state information determination deteriorates under varying charging speeds
Solution Approach 1:
The patent applies dynamics by making the time interval values adaptive rather than fixed. The processor determines different time interval values between feature points based on the actual charging conditions and current integration requirements. This dynamic adjustment allows the system to maintain accurate battery state estimation across varying charging speeds without requiring an overly complex fixed-structure approach, resolving the contradiction between measurement precision and device complexity.
2Measurement precision
If fixed time intervals are used for voltage sampling, then the processing is simple, but the feature vector length varies with charging speed reducing estimation accuracy
Solution Approach 1:
The patent changes the parameter of time interval values from fixed to variable based on current integration. By determining time interval values that satisfy specific current integration conditions rather than using fixed intervals, the system maintains consistent feature vector lengths across different charging speeds. This parameter change enables accurate battery state estimation while the processor efficiently handles the variable intervals through systematic determination methods.
3Adaptability or versatility
If current integration-based time interval determination is implemented, then consistent feature vector length is achieved across charging speeds, but the calculation complexity increases
Solution Approach 1:
The patent implements feedback by using current integration values as a basis for determining subsequent time interval values. The system continuously monitors current information, calculates integrations over determined intervals, and uses these integration results to determine the next time interval values. This feedback mechanism ensures consistent feature vector lengths across different charging speeds while managing calculation complexity through an iterative, condition-based determination process.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A processor-implemented method of estimating a state of a battery includes acquiring current information and voltage information of a battery; determining time interval values based on the acquired current information such that current integration values corresponding to the time variation values satisfy a condition; determining voltage values corresponding to the determined time interval values in the acquired voltage information; and determining state information of the battery based on the determined voltage values.