Battery Modeling Using Phase-Based Data Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing battery models face challenges in accurately estimating state of charge (SOC), state of health (SOH), and state of function (SOF) due to the complexity of acquiring accurate parameters, particularly in real-time applications, as most models require extensive computation and specific knowledge of the battery type.
Innovation Solution
A method and apparatus that utilize processing circuitry to classify battery information as effective or ineffective data based on the phase of the battery cycle, identifying parameters for a circuit model using condition number measurements, and generating accurate battery state estimations by separating the charging/discharging cycle into distinct regions for parameter identification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If electrochemical model is used to describe chemical reactions inside battery, then accuracy of battery modeling is improved, but computation complexity increases making it difficult for real-time applications
Solution Approach 1:
The patent segments the battery cycle into distinct phases (charging phase, discharging phase, rest phase) and applies different modeling approaches to each phase. The electrochemical model is used only during rest phases when accuracy is critical and computation can be performed offline, while simplified models are used during active charging/discharging phases for real-time applications.
Solution Approach 2:
The patent applies different modeling qualities to different operational conditions. High-fidelity electrochemical modeling is applied locally during rest phases where the battery state is stable and accurate parameter identification is most valuable, while lower-complexity models are used during dynamic phases where real-time performance is prioritized.
2Device complexity
If analytical model with flat discharge profiles is used, then computation complexity is reduced, but applicability is limited to large lead-acid batteries only
Solution Approach 1:
The patent creates a universal modeling framework that can handle multiple battery types (lithium-ion, lead-acid, nickel-based) and multiple operational phases (charging, discharging, rest) within a single system. The framework adapts the appropriate model complexity and parameters based on battery type and phase, making it versatile across different applications.
Solution Approach 2:
The patent transitions from static flat discharge profiles to dynamic phase-based modeling that adapts to actual battery operating conditions. The model dynamically selects appropriate mathematical representations based on whether the battery is charging, discharging, or at rest, improving applicability to various battery types and operating scenarios.
3Measurement precision
If model-based algorithms are used for SOC/SOH/SOF estimation, then estimation accuracy can be improved, but accurate parameter identification becomes difficult
Solution Approach 1:
The patent performs preliminary parameter identification during rest phases when the battery is not actively charging or discharging. During these periods, the battery reaches a stable equilibrium state, allowing accurate open-circuit voltage measurements and simplified parameter extraction before the battery enters dynamic operating phases where estimation algorithms are applied.
Solution Approach 2:
The patent leverages the battery's own natural behavior during rest phases to self-identify parameters. By utilizing the spontaneous relaxation of the battery chemistry at rest, the system can automatically determine parameters without requiring external excitation signals or complex identification procedures, making the process easier and more accurate.
4Quantity of substance
If all battery data is used for parameter identification, then more data is available for modeling, but ineffective data reduces the accuracy of parameter identification
Solution Approach 1:
The patent extracts only the effective portions of battery data for parameter identification. By separating charging/discharging phases from rest phases and selecting only data from rest phases for certain parameter identifications, the system removes ineffective data that would otherwise degrade the accuracy of parameter estimation while maintaining sufficient data quantity for robust modeling.
Data Source
AI summary
A method and an apparatus for battery modelling. The method includes acquiring battery information from a sensor attached to a battery, the battery information including at least a terminal voltage and a load current; classifying, the battery information as effective data or ineffective data based on a phase of a battery cycle during which the battery information is acquired; identifying one or more parameters of the circuit model associated with the battery based on the effective data; and generating an estimation of a state of the battery using the circuit model having the one or more parameters identified using the effective data. Further, a circuit model is identified using effective data.