AI Modelled Clutch Operation For Automobiles

Automatic clutch engagement is essential for the modern powertrain. Most manufacturers utilize torque converters with a peak efficiency of 96% at lower speeds. But this rapidly reduces to 80 -85 % at higher speeds. The high inertia of torque converters causes this sharp reduction in peak efficiency. Clutches can efficiently replace the torque converter, but the human interventions in clutch operation make it less attractive. The recent advances in Artificial Intelligence (AI) can significantly reduce the human interventions in the clutch operation and make clutches a worthy candidate to be used as a replacement for torque converters. This paper proposes an AI-enabled clutch mechanism composed of a feed-forward neural network to determine the operation time for a mechanism that consists of a stepper motor attached to a hydraulic actuator via a rack and pinion which moves the diaphragm spring of the clutch. The proposed AI-enabled model is trained using different training methods, and the performance is evaluated using the R-value