Machine Learning Prediction of Wear of NAO Brake Pad Using Scaled-Down FMVSS 135 Testing

Brake pad lifespan significantly influences environmental pollution and public health due to non-exhaust particulate emissions. Global vehicle production reached approximately 7 billion units in 2025, with brake systems generating an estimated 50,000 to 100,000 metric tons of wear debris annually, underscoring the need for improved durability and material optimization. This study investigates the capability of a Fully Connected Neural Network (FCNN) to predict brake pad wear and support formulation optimization. Controlled experiments were performed using a simplified FMVSS 135 protocol on a Universal Mechanical Tester (UMT) to simulate realistic braking conditions. Wear loss was measured under varying normal loads, sliding velocities, and interface temperatures. An FCNN was selected for its ability to capture nonlinear interactions between operating conditions and tribological responses. Despite a limited but high-quality dataset, the application of data standardization, regularization, and hyperparameter optimization enabled robust model generalization. The resulting architecture achieved a balance between predictive accuracy and model interpretability. A Taguchi L8 design of experiments was employed to optimize brake pad compositions, significantly reducing experimental time and material consumption compared to conventional approaches. The proposed AI-driven framework demonstrated strong agreement between experimental measurements and FCNN predictions, validating its effectiveness. Its scalable structure further enables seamless integration of additional material and process parameters, which supports iterative brake formulation development in industrial settings. By improving wear and friction property prediction while minimizing empirical testing requirements, this approach also supports sustainable brake material development. In addition, enhanced brake pad durability reduces replacement frequency, waste generation, and associated carbon emissions, contributing to improved environmental and public health outcomes.