A hybrid analytical and data driven framework for optimizing radially grooved wet clutch geometry.

Area of Science:

• Mechanical Engineering
• Fluid Dynamics
• Tribology

Background:

• Drag torque in wet clutches arises from relative motion of disks within an oil film.
• Minimizing drag torque is crucial for improving efficiency in wet clutch systems.

Purpose of the Study:

• To optimize the geometry of radially grooved wet clutches for drag torque reduction.
• To develop and validate data-driven models for predicting drag torque.

Main Methods:

• An analytical model was employed to generate datasets for single- and multiphase flow conditions.
• Computational Fluid Dynamics (CFD) validated the analytical model's accuracy (max 8% deviation).
• Artificial neural networks and a genetic algorithm were used for optimization.

Main Results:

• Optimized geometries achieved drag torque reductions of at least 70% across four cases.
• Case 3 showed a 97% reduction in peak drag torque (0.92 to 0.022).
• Parametric study revealed disk distance as the most impactful parameter (86% reduction), groove angle the least (2% reduction).

Conclusions:

• Data-driven optimization effectively minimizes drag torque in radially grooved wet clutches.
• Geometric parameters significantly influence drag torque, with disk spacing being a key factor.
• The developed models offer an efficient alternative to computationally expensive numerical simulations.