Battery Temperature Estimation Using Neural Network Impedance Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for estimating battery temperature require state of charge as an input parameter and rely solely on impedance magnitude, making them less robust and specific to individual batteries.
Innovation Solution
A computer-implemented method using an artificial neural network to estimate battery temperature based on series of electrical impedance measurements at different frequencies, eliminating the need for state of charge as an input and utilizing both real and imaginary parts of impedance values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional methods use impedance magnitude and state of charge as input parameters, then temperature estimation can be performed, but the method becomes less robust and specific to individual batteries
Solution Approach 1:
The patent extracts and removes the state of charge parameter from the input requirements, relying solely on impedance measurements (both magnitude and phase) across multiple frequencies to estimate temperature. This extraction of unnecessary parameters simplifies the input requirements while maintaining or improving estimation robustness through the use of phase information.
Solution Approach 2:
The patent transitions from using only impedance magnitude (one-dimensional information) to utilizing both impedance magnitude and phase (two-dimensional complex impedance information). This dimensional expansion provides more comprehensive battery state information, improving temperature estimation robustness without adding external sensor requirements.
2Measurement precision
If temperature sensors are used to measure battery temperature, then direct temperature measurement is achieved, but the system complexity and cost increase
Solution Approach 1:
The patent replaces physical temperature sensors (mechanical/thermal measurement devices) with an electrical measurement-based temperature estimation system. By substituting thermal sensors with electrical impedance measurements and neural network processing, the system achieves temperature estimation without direct thermal contact or additional sensor hardware.
Solution Approach 2:
The patent introduces impedance phase information as an intermediary parameter that indirectly reveals temperature information. Instead of directly measuring temperature with sensors, the system uses phase data from impedance measurements as a mediator to infer temperature through neural network processing.
3Measurement precision
If only impedance magnitude is used for temperature estimation, then the method is simpler to implement, but the estimation precision and battery-specific adaptation are reduced
Solution Approach 1:
The patent expands from one-dimensional impedance magnitude measurements to two-dimensional complex impedance measurements by incorporating phase information. This additional dimensional information from the phase component enhances temperature estimation precision by providing more comprehensive electrical characteristics of the battery.
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
A computer-implemented method and a temperature estimating system for estimating a temperature (T) of an electrochemical battery (14), including: providing (S12) a series of electrical impedance measurements of an electrochemical battery (14), each electrical impedance measurement being measured at a respective measurement frequency (fs), the series being ordered according to the respective measurement frequencies; and determining a temperature of the electrochemical battery (14) using artificial neural network means (28) configured to receive as inputs a series of electrical impedance values, wherein a series of electrical impedance values is provided to the artificial neural network means (28), the series of electrical impedance values corresponding to the provided series of electrical impedance measurements, wherein the artificial neural network means (28) receives and processes the provided series of electrical impedance values to generate therefrom an output signal representing a temperature (T) associated with the electrochemical battery (14).