Optimisation of Clutch Disc Friction Material Using a Multi-Layer Perceptron Artificial Neural Network

Area of Science:

• Materials Science
• Mechanical Engineering
• Tribology

Background:

• Clutch discs are critical components in manual transmissions, requiring specific friction properties.
• Variations in friction material composition affect performance and durability.
• Optimizing friction materials is essential for enhancing vehicle performance and longevity.

Purpose of the Study:

• To analyze and optimize clutch disc friction materials from different manufacturers.
• To develop an artificial neural network model for predicting and improving friction material performance.
• To identify key chemical elements influencing friction coefficient and temperature development.

Main Methods:

• Microstructural and chemical characterization using Scanning Electron Microscopy (SEM) and Energy-Dispersive X-ray Spectroscopy (EDX).
• Tribological testing to measure friction coefficients against cast iron under simulated conditions.
• Development of an artificial neural network (ANN) model using Easy NN software (V 14) with chemical composition as input and friction/temperature data as output.

Main Results:

• Identified significant correlations between specific chemical elements and friction material performance.
• The ANN model with three hidden layers effectively predicted friction coefficient and temperature.
• Optimization using the ANN model aimed to minimize friction coefficient variations and temperature fluctuations.

Conclusions:

• The study successfully optimized clutch disc friction material composition through an AI-driven approach.
• Key chemical elements were identified as crucial for achieving desired tribological properties.
• The developed ANN model provides a valuable tool for designing advanced friction materials for manual transmissions.