Battery State Estimation Using Sensor Offset Noise
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing battery state estimation methods, such as the dual adaptive extended Kalman filter, do not accurately reflect the offset and variance of current and voltage sensors, leading to reduced accuracy in estimating battery state of charge (SOC) and state of health (SOH).
Innovation Solution
A battery state estimating apparatus and method that calculates voltage and current offsets and variances, and uses these values to determine an offset noise matrix and a variance noise matrix. These noise matrices are then used to calculate system noise, which is applied to a recursive filter to more accurately estimate battery state information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a conventional dual adaptive extended Kalman filter is used for battery state estimation, then the estimation process can be performed recursively, but the accuracy of SOC and SOH estimation is reduced due to not considering sensor offset and variance
Solution Approach 1:
The patent applies preliminary action by calculating the offset and variance of current and voltage sensors in advance during a rest period (when battery current is zero or near-zero). These pre-calculated parameters are then used in the subsequent recursive filtering process to improve SOC and SOH estimation accuracy without adding complexity to the real-time estimation algorithm.
Solution Approach 2:
The patent changes the parameters of the Kalman filter by incorporating dynamically calculated offset and variance parameters that reflect actual sensor characteristics. Instead of using fixed or theoretically assumed noise parameters, the system adapts the process noise covariance matrix Q and measurement noise covariance matrix R based on empirically determined sensor offset and variance, thereby improving estimation accuracy.
2Measurement precision
If sensor offset and variance are not considered in the filtering process, then the estimation algorithm remains simple, but the accuracy of SOC and SOH estimation is reduced
Solution Approach 1:
The system applies self-service by automatically calculating its own sensor offset and variance parameters during normal operation. The battery management system uses periods when the battery is at rest (zero current) to self-diagnose and determine the actual offset and variance of the current and voltage sensors, eliminating the need for external calibration equipment or manual measurement procedures.
Solution Approach 2:
The patent implements feedback by continuously monitoring the relationship between measured voltage/current and the battery model predictions. During rest periods, the system compares the open-circuit voltage with the model's predicted voltage to detect sensor offset, and monitors voltage fluctuations to estimate sensor variance. This feedback information is then fed back into the Kalman filter parameters to improve subsequent estimation accuracy.
3Reliability
If the system uses preset noise parameters in the recursive filter, then the implementation is straightforward, but the estimation accuracy does not reflect actual sensor performance
Solution Approach 1:
The system performs preliminary calculation of noise parameters during rest periods when the battery is not charging or discharging. By calculating offset and variance in advance during these idle periods, the system prepares accurate noise parameters for the subsequent active operation phases, ensuring reliable estimation without complicating the real-time processing during charging/discharging.
Solution Approach 2:
The patent applies periodic action by repeatedly calculating sensor offset and variance at scheduled rest periods throughout battery operation. Instead of attempting to calculate these parameters continuously during charging/discharging (which would be computationally intensive and less accurate), the system periodically updates the noise parameters during natural rest intervals, maintaining reliability while managing complexity.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
The present disclosure has an advantage of more accurately estimating the battery state by adding system noise to a recursive filter used to estimate the battery state. According to an embodiment of the present disclosure, since the parameters used in an extended Kalman filter are corrected in consideration of the offset and variance of the current sensor and the voltage sensor, the state of the battery can be more accurately estimated.