DCT Clutch Torque Learning via Angular Acceleration Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Dual Clutch Transmission (DCT) systems face challenges in predicting clutch torque variations due to inconsistencies in engine rotation inertia and drag characteristics, which affect shifting sensation and accuracy.
Innovation Solution
A method and apparatus for learning clutch torque in DCT systems that involve judging shifting events, controlling torque transmission between clutches, detecting learning values based on engine angular acceleration, and compensating clutch torque using slip control torque, while considering vehicle driving states and shifting conditions to improve shifting sensation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If T-S curve shape and touch point are predicted using engine operating conditions and clutch rotation speed, then clutch torque characteristics can be estimated, but prediction accuracy deteriorates due to variation in engine rotation inertia and drag characteristics
Solution Approach 1:
The patent implements feedback by detecting actual clutch torque during shifting operations and using this detected value to correct and update the predicted T-S curve characteristics. The controller continuously compares predicted torque with actual torque and adjusts the prediction model accordingly, thereby compensating for variations in engine rotation inertia and drag characteristics that affect prediction accuracy.
Solution Approach 2:
The patent changes the parameters used for torque prediction from static engine operating conditions to dynamic parameters that include detected actual clutch torque values. By updating the T-S curve prediction based on real-time detected torque and adjusting engine rotation inertia parameters during learning operations, the system adapts to varying engine conditions and maintains accurate torque prediction across different operating scenarios.
2Measurement precision
If learning is performed frequently to improve touch point prediction, then prediction accuracy improves, but system complexity and control burden increase
Solution Approach 1:
The patent performs learning operations at predetermined timing during normal shifting sequences without requiring separate dedicated learning modes. By utilizing regular shifting opportunities to update clutch torque characteristics, the system achieves continuous improvement of prediction accuracy while avoiding the complexity of dedicated learning procedures or additional hardware components.
3Ease of operation
If engine torque is assumed consistent for T-S curve prediction, then prediction process is simplified, but shifting sensation deteriorates due to torque inconsistency
Solution Approach 1:
The patent uses feedback from actual clutch torque detection to continuously update the T-S curve prediction model. This allows the system to maintain simple prediction calculations while compensating for engine torque inconsistencies through real-time correction based on detected actual torque values, thereby preserving both computational simplicity and shifting quality.
Data Source
AI summary
An apparatus and method of using the apparatus for learning a clutch torque of a Dual Clutch Transmission (DCT) includes judging whether shifting begins, controlling torque transmission of a coupling-side clutch and a release-side clutch to cross each other while following a coupling-side target clutch torque when shifting begins, and calculating and storing a learning value, after the controlling has begun, using a function determined by the relationship between an average value of an engine angular acceleration and a coupling-side shifter input-shaft angular acceleration, engine rotation inertia, and a torque compensated to the torque transmission of the coupling-side clutch for reducing a slippage of the release-side clutch by feedback control during a torque handover interval.