Battery Module SOC Estimation Using Composite Probability Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing battery management systems require extensive sensor networks and computational resources to accurately monitor and estimate the state of charge (SOC) of multiple battery cells, which increases costs and affects performance.
Innovation Solution
A method and system using a composite probability variable model and Kalman filter operation, leveraging a microcontroller unit (MCU) and neural processing unit (NPU) co-processing to estimate SOC based on voltage measurements of select battery cells, reducing the need for full cell monitoring.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If voltage measurement sensors are connected to all battery cells for full monitoring, then measurement precision of SOC is improved, but device complexity and cost increase
Solution Approach 1:
The patent divides the battery module into multiple segments (battery cells) and selectively monitors only specific segments (first and last cells) rather than all cells. This segmentation approach allows SOC estimation with reduced sensor coverage, maintaining acceptable accuracy while lowering system complexity and cost.
Solution Approach 2:
The patent introduces a probabilistic model and Kalman filter algorithm as intermediaries to infer the states of unmonitored battery cells based on measurements from monitored cells. This intermediary computational approach compensates for the reduced physical sensor coverage, enabling accurate SOC estimation without requiring sensors on every cell.
2Measurement precision
If voltage measurement data from all battery cells is processed, then SOC estimation accuracy is improved, but computational resource usage increases
Solution Approach 1:
The patent extracts and processes only the essential voltage measurement data from selected battery cells (first and last cells) rather than processing data from all cells. This selective data extraction reduces the computational burden while maintaining sufficient information for accurate SOC estimation through the probabilistic model and Kalman filter.
Solution Approach 2:
The patent applies partial action by processing voltage data from only a subset of battery cells (specifically the first and last cells in series) rather than all cells. This partial monitoring approach, combined with probabilistic inference, achieves acceptable SOC accuracy with significantly reduced computational resource consumption.
3Device complexity
If selective voltage measurement of first and last battery cells is used, then device complexity is reduced, but measurement precision of SOC may deteriorate
Solution Approach 1:
The patent introduces a probabilistic model and Kalman filter as computational intermediaries that infer the states of unmonitored battery cells based on measurements from the first and last cells. This intermediary inference mechanism compensates for the limited physical sensor coverage, maintaining SOC estimation accuracy despite reduced sensor network complexity.
Solution Approach 2:
The patent changes the approach from direct measurement of all cell parameters to probabilistic inference of unmonitored cell parameters based on monitored cells. By transforming the problem from complete direct measurement to partial measurement with statistical inference, the system achieves comparable accuracy with reduced complexity.
Data Source
AI summary
The present disclosure relates to a method of estimating a state of charge of a battery module, comprising: obtaining, by a microcontroller unit, a composite probability variable model associated with a plurality of battery cells included in a particular battery module, receiving, by the microcontroller unit, voltage measurement data of a first battery cell and voltage measurement data of a second battery cell of the plurality of battery cells included in the particular battery module, and estimating, by the microcontroller unit and/or a neural processing unit, an SOC of the particular battery module via a Kalman filter operation based on the composite probability variable model, the voltage measurement data of the first battery cell, and the voltage measurement data of the second battery cell.