Parallel Battery System Current Prediction via Backward Model
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing battery systems in electric vehicles with multiple chemical types of batteries connected in parallel face challenges in predicting how these batteries will respond to demanded currents due to their different dynamic characteristics and impedances, which affects the system's ability to output the required current efficiently.
Innovation Solution
A method and system that use a backward-looking model to predict battery responses, considering state of charge (SOC) dynamics and internal impedance, to accurately determine how batteries of different chemical types will operate together to meet a demanded current command, including predicting SOC changes, open circuit voltage, internal resistance, output currents, and terminal voltage responses.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If multiple chemical types of batteries are connected in parallel to increase system capacity and flexibility, then the battery system can provide better performance and adaptability, but predicting the response of each battery to demanded current becomes more complex due to different dynamic characteristics and impedances
Solution Approach 1:
The patent segments the battery system into individual battery units, each with its own dynamic characteristics and impedances. By treating each battery separately in the prediction model rather than as a homogeneous group, the system can accurately capture the unique response of each chemical type while maintaining overall system management.
Solution Approach 2:
The patent incorporates dynamic parameters including state of charge (SOC) and internal impedance that change with operating conditions. The prediction model adjusts these parameters in real-time based on demanded current and battery state, allowing accurate prediction despite varying chemical characteristics across different battery types.
2Ease of operation
If a simple battery model is used for prediction, then computational speed and ease of operation are improved, but accuracy in predicting battery responses including SOC dynamics and internal impedance changes is reduced
Solution Approach 1:
The patent replaces complex iterative mechanical/electrical simulation methods with a direct algebraic prediction approach. By using closed-form equations that directly calculate battery responses from demanded current and initial state, the system achieves both computational efficiency and prediction accuracy without requiring iterative solving procedures.
Solution Approach 2:
The model dynamically updates key parameters such as state of charge and internal impedance based on predicted current draw and battery chemistry characteristics. This allows the prediction process to maintain high accuracy by adapting parameters to current operating conditions while keeping the computational framework relatively simple.
Data Source
AI summary
A powertrain having a battery system including at least two batteries of different chemical types connected in parallel is operated according to predicted responses of the batteries to a demanded current command for the battery system to output a demanded current. The battery responses are predicted directly from the demanded current using a backward-looking model of the battery system.