Case Study: Largest Brake Pad Manufacturing Firm predicts friction classes with Quantiphi's machine learning solution

Friction prediction for Materials using Machine Learning

Largest Brake Pad Manufacturing Firm partnered with Quantiphi for a virtual testing materials initiative to predict friction for brake pad material compositions. The R&D team was spending significant time and money running repeated physical tests and adjusting formulations, with around 2,000 features in messy, unstructured historical data making it difficult to identify the right inputs for accurate friction prediction.

Quantiphi implemented a machine learning solution using Python, R, and GCP, applying random forest with grid search, extreme gradient methods, and dynamic time warping to pre-determine friction classes. Quantiphi’s framework helped the customer achieve 76% precision, enabling the R&D team to screen test cases earlier, reduce iterations, and lower development costs.