Traction Battery SOC Control With Dynamic EKF Gain Scheduling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing battery state of charge (SOC) estimation methods in electrified vehicles are inaccurate, leading to unreliable distance-to-empty (DTE) calculations, which can result in vehicles becoming stranded due to depleted batteries at inconvenient locations.
Innovation Solution
A traction battery controller dynamically selects an estimation gain based on vehicle operating conditions, using an extended Kalman filter (EKF) to optimize SOC estimation speed and accuracy by adjusting the EKF gain according to on-board energy (OBE) and power demand, employing a gain-scheduling strategy to ensure accurate SOC estimation, especially when energy levels are low.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If a high estimation gain is used in the EKF, then the SOC estimation speed is increased, but the accuracy of the SOC estimation decreases
Solution Approach 1:
The patent applies dynamics by making the EKF gain adjustable rather than fixed. The gain is dynamically modified based on the estimated SOC value - using higher gain when SOC is above threshold values (faster estimation) and lower gain when SOC is below threshold values (higher accuracy), thus resolving the contradiction between speed and accuracy across different operating conditions
Solution Approach 2:
The patent changes the EKF gain parameter based on SOC levels. By modifying the gain parameter dynamically according to the estimated SOC and comparing it against threshold values, the system adapts the estimation characteristics to match the current operating condition, achieving both fast and accurate estimation as needed
2Measurement precision
If a low estimation gain is used in the EKF, then the accuracy of the SOC estimation is improved, but the speed of SOC estimation decreases
Solution Approach 1:
The system dynamically adjusts the EKF gain based on the current SOC level. When SOC falls below certain threshold values, the gain is reduced to prioritize accuracy over speed, ensuring reliable estimation when battery charge is low and critical for preventing vehicle stranding
Solution Approach 2:
The EKF gain parameter is changed based on SOC thresholds. The controller compares the estimated SOC against predefined threshold values and modifies the gain parameter accordingly - using lower gain when SOC is low to improve accuracy, and higher gain when SOC is high to maintain faster estimation response
3Device complexity
If a fixed nominal estimation gain is used, then the system complexity is reduced, but the accuracy of SOC estimation under varying operating conditions deteriorates
Solution Approach 1:
Rather than using a complex adaptive observer with continuously varying gain, the patent implements a simpler dynamic approach where the EKF gain is adjusted based on discrete SOC threshold comparisons. This maintains relatively low system complexity while still improving accuracy under varying operating conditions
Solution Approach 2:
The patent modifies the EKF gain parameter based on SOC levels without requiring complex adaptive algorithms. By comparing estimated SOC against threshold values and selecting appropriate gain values, the system achieves improved accuracy across different operating conditions while keeping the control logic relatively simple
Data Source
AI summary
A traction battery controller of an electrified vehicle dynamically selects an estimation gain based on an operating condition of the vehicle and controls the vehicle according to a state (e.g., a state-of-charge (SOC)) of the traction battery estimated from an equivalent circuit model of the traction battery that depends on the estimation gain. The operating condition of the vehicle may be indicative of on-board energy (OBE) or a power demand of the vehicle. In operation, the controller selects the estimation gain to be (i) a high estimation gain during a first operating condition of the vehicle, whereby compared to a nominal estimation gain the battery state is estimated at an increased speed but with lower accuracy and (ii) a low estimation gain during a second operating condition of the vehicle, whereby compared to the nominal estimation gain the battery state is estimated with greater accuracy but at a decreased speed.