Clutch Characteristic Curve Learning for Stable Torque Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing clutch control systems face issues with excessive slip and impact due to unreflected changes in clutch torque characteristics caused by wear, thermal deformation, and friction coefficient changes, leading to inadequate control based on a predetermined characteristic curve.
Innovation Solution
A method for adjusting the clutch characteristic curve through learning, which involves obtaining torque-stroke learning values, calculating convergence values, determining new characteristic curve values based on difference values, and updating the curve to suppress transient response changes while maintaining the overall shape.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the clutch characteristic curve is adjusted rapidly to reflect current clutch state, then control precision is improved, but transient response instability occurs
Solution Approach 1:
The patent implements dynamic adjustment of the clutch characteristic curve by continuously learning and updating torque-stroke relationships based on actual clutch operation. The system adapts the characteristic curve in real-time to reflect current clutch state, resolving the contradiction between rapid adaptation and transient stability through controlled dynamic updates.
Solution Approach 2:
The patent employs feedback mechanisms by calculating convergence values based on actual clutch torque and stroke measurements, then using these feedback values to adjust the characteristic curve. This closed-loop feedback ensures precise control while maintaining stability through iterative refinement rather than abrupt changes.
2Measurement precision
If the characteristic curve is updated frequently to maintain accuracy, then control accuracy is improved, but system complexity increases
Solution Approach 1:
The patent implements a self-learning control system that automatically updates its own characteristic curve based on measured clutch behavior. The system performs self-adjustment through automated convergence value calculations and curve updates, eliminating the need for external manual calibration and reducing operational complexity while maintaining high accuracy.
Solution Approach 2:
The patent calculates convergence values in advance based on learned torque-stroke relationships, preparing adjustment parameters before actual control actions are needed. This preliminary calculation of characteristic curve updates allows the system to maintain accuracy while simplifying real-time control decisions.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present invention relates to a method for adjusting a clutch characteristic curve, characterized by including steps of obtaining a torque-stroke learning value for adjusting the clutch characteristic curve, calculating a convergence value for each control point of the clutch characteristic curve, calculating a difference value between the convergence value and a characteristic curve value for each control point, and determining a new characteristic curve value of each control point according to whether a maximum value of the calculated difference values exceeds a preset reference value.