Real-Time Battery State Prediction Using Differential Algorithm
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing algorithms for determining the state of rechargeable batteries in vehicles, such as the superposition integral (SI) algorithm, face instability at high sampling rates and are sensitive to initial parameters, leading to numerical anomalies and limited accuracy in predicting state of charge (SOC) and state of power (SOP).
Innovation Solution
The direct differential (DD) algorithm uses a mathematical model based on an equivalent RC circuit to measure current and voltage in real time, treating charging and discharging events separately, and employs a weight recursive least squares method to regress parameters, providing more stable and accurate predictions of SOC and SOP.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the superposition integral (SI) algorithm is used to determine battery state, then the method can predict state of charge (SOC) and state of power (SOP), but the algorithm becomes unstable at high sampling rates and produces numerical anomalies
Solution Approach 1:
The patent changes the mathematical approach from integral-based to differential-based models. The differential equation model directly solves for battery state using time-derivative relationships, which are more stable at high sampling rates than integral methods. This parameter change in the mathematical formulation resolves the instability issue while maintaining prediction accuracy.
Solution Approach 2:
The patent replaces the integral-based mathematical mechanism with a differential equation-based mechanism. This substitution fundamentally changes how the algorithm processes battery data, using differential relationships between voltage, current, and time to achieve more stable and accurate predictions without numerical anomalies.
2Measurement precision
If existing algorithms are used to determine battery state, then SOC and SOP can be predicted, but the algorithms are sensitive to initial parameters leading to limited accuracy
Solution Approach 1:
The patent implements a feedback mechanism where the differential equation model continuously adjusts and refines the battery state estimation based on actual measured voltage and current data. This feedback loop reduces sensitivity to initial parameters by constantly correcting the estimation, leading to more accurate SOC and SOP predictions.
Solution Approach 2:
The patent uses dynamic differential equations that adapt to changing battery conditions in real-time. The model dynamically adjusts its behavior based on the current state of the battery, making it less sensitive to initial parameter values and improving overall prediction accuracy through adaptive estimation.
3Productivity
If high sampling rates are used to improve measurement frequency, then real-time battery monitoring is enhanced, but algorithm stability deteriorates
Solution Approach 1:
The patent changes the mathematical approach from integral-based to differential-based models. The differential equation model directly solves for battery state using time-derivative relationships, which are more stable at high sampling rates than integral methods. This parameter change in the mathematical formulation resolves the instability issue while maintaining prediction accuracy.
Data Source
AI summary
A method of determining and predicting a state of a rechargeable battery device in real time involves measuring a current and a voltage of the rechargeable battery in real time, inputting the measured current and voltage into an algorithm, and applying the algorithm to determine the state of the rechargeable battery. The algorithm includes a first mathematical model based on a direct solution of at least one differential equation characterizing an equivalent RC circuit of the battery as a function of time. The first model generates a plurality of parameters that are usable to determine the state of the battery. The algorithm further includes a second mathematical model configured to regress the parameters over time, and a third mathematical model configured to estimate the state of the battery.