Torque Anomaly Detection in Electric Vehicles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current torque monitoring systems in electric and hybrid vehicles fail to accurately detect abnormal differences and variations between the driver's torque request and the torque produced by the engine, leading to potential safety issues such as untimely acceleration, loss of traction, and over-braking, which are not effectively addressed by existing tolerance threshold-based methods.
Innovation Solution
A system and method that construct torque templates with upper and lower limits based on the driver's torque request and average percentage error, allowing for the detection of untimely acceleration, loss of traction, and over-braking by filtering the driver's torque request and verifying consistency between the requested and estimated torques during quasi-static phases, thereby reducing false detections and improving safety.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a fixed tolerance threshold is used to detect torque anomalies, then the detection method is simple, but the detection accuracy is poor and false detections occur during dynamic phases
Solution Approach 1:
The system dynamically adapts the torque template based on the detected phase (quasi-static or dynamic). During quasi-static phases, a tolerance-based template is used, while during dynamic phases, a learned template from normal operation is used. This dynamic adaptation resolves the contradiction by allowing high detection accuracy without requiring complex real-time calculations during all operating conditions.
Solution Approach 2:
The system changes the detection parameters (torque template) based on the operating phase. The template is constructed differently for quasi-static phases (using tolerance thresholds) versus dynamic phases (using learned normal behavior patterns). This parameter change allows the system to maintain high detection accuracy across different operating conditions while keeping the overall system manageable in complexity.
2Reliability
If monitoring is performed during all phases including dynamic variations, then comprehensive coverage is achieved, but false detections increase due to normal torque variations
Solution Approach 1:
The monitoring system segments the operating phases into quasi-static and dynamic phases. During quasi-static phases, anomaly detection is performed using tolerance-based templates. During dynamic phases, the system either uses learned templates or reduces monitoring intensity. This segmentation resolves the contradiction by applying appropriate detection strategies to each phase, maintaining reliability without generating false alarms from normal dynamic variations.
Solution Approach 2:
The system periodically learns the normal torque patterns during dynamic phases and uses this learned information to distinguish normal variations from actual anomalies. By periodically updating the understanding of normal behavior, the system can comprehensively monitor all phases while filtering out false detections caused by routine dynamic variations.
3Measurement precision
If a dynamic torque template is constructed based on learned normal phases, then detection accuracy during dynamic phases improves, but the system complexity and computation time increase
Solution Approach 1:
The system performs preliminary learning of normal torque patterns during dynamic phases when no anomalies are present. This learned information is stored and reused during subsequent dynamic phases, avoiding the need for complex real-time analysis during each dynamic phase. This preliminary action resolves the contradiction by achieving high detection precision through offline learning rather than online computation.
Solution Approach 2:
The system creates a copy or model of normal torque behavior during learning phases and uses this model to detect anomalies during operation. Instead of performing complex real-time analysis, the system compares current torque against the pre-established model, significantly reducing computational complexity while maintaining high detection precision.
4Ease of manufacture
If tolerance thresholds are used for torque comparison, then the implementation is straightforward, but the system cannot detect abnormal variations in torque difference
Solution Approach 1:
The system uses feedback from the torque monitoring to dynamically adjust the detection strategy. When the torque difference exceeds the tolerance threshold, the system triggers further analysis using the learned torque template. This feedback mechanism resolves the contradiction by maintaining simple implementation through tolerance thresholds while enhancing reliability through adaptive follow-up detection when anomalies are suspected.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
System for monitoring the torque supplied by the motor of a motor vehicle, notably of an electric or hybrid vehicle. The system comprises a means of determining the driver's request for torque, a means of estimating the torque produced by the propulsion unit, a means (1) of filtering the driver torque request which means is connected at input to a means (2) of determining the consistency between the torque request and the torque produced comprising a means (3) of determining the consistency of the positive torque and a means (4) of determining the consistency of the negative torque each one comprising a means (5, 7) of determining the phases of quasi-static variation of the request and a means (6, 8) of detecting torque anomalies which are able to construct an acceptable torque template and to determine that a torque anomaly has occurred as a function of the torque estimate, the torque request and the acceptable torque template.