Model accuracy
Saved per test
In potential annual savings
The customer, a leading global manufacturer, designs world-class critical safety systems including brake pads and shock absorption technology for automotive firms around the world. The material scientists and chemical formulators use their domain knowledge and experience to formulate chemical compositions to design the brake pads with friction coefficients specific to the automobile requirements.
Their Research & Development team invests significant time and money to test and determine friction performance levels of the brake pad material. Therefore, this is a complex process that requires a great deal of testing and engineering.
They wanted to implement an automated solution to predict the friction coefficient and material coefficient of a brake pad in order to reduce the number of iterations in designing a brakepad while also reducing the resources wasted in testing.
Python
Google's BigQuery
CatBoost
Google Cloud Platform
Quantiphi developed a simulation environment with a virtual material testing framework to conduct the tests live without wasting too many resources. The virtual testing solution enables chemical formulators to design materials on the desktop and run confirmatory tests only for the most promising material compositions; allowing preemptive screening of sub optimal brake pads and identification of optimal designs.
Quantiphi also developed a web-based user interface for predicting the friction coefficient of the brake pad enabling chemical formulators to submit input features and get corresponding friction coefficient of the designed materials.
