Global Tire Manufacturer using Predictive Analytics for Product Design
Modern tires are made up of multiple layers of synthetic rubber, natural rubber, fabric and wire, along with carbon black and other chemical compounds. Different materials, combinations of materials, and tread shapes and styles offer different benefits. One could be better in the winter, another in the summer. One offers better fuel mileage, another less noise. This recipe of materials, as well as the production process, which offers another set of variables such as temperature at different stages, ultimately defines the resulting product. Previously, materials engineers from this major tire manufacturer would create and test numerous options when they were creating a new product. They would then select those options coming closest to the tire with the desired properties. But in the end, this system, with a collection of tests and measurements for several thousand options, is complex and costly.
The firm set up a test with a RapidMiner-based predictive model that predicts the properties of the results based on the recipe. Analysts can now predict how a tire will behave based on the recipe, and can then focus on the most promising prototypes and actually prototype only those.