Friction material design for brake pads using database

Stochastic methods were used for the analysis of test data related to the design of brake pad compositions. The friction and wear test data were analyzed by MRA (Multiple Regression Analysis) and the effect of %volume of the components on the friction and wear characteristics was obtained quantitatively in the form of the standard regression coefficients. The analytical results concerning the effects on component materials on friction and wear characteristics correspond well to the knowledge we have gained though experiences. It was also found that the standard regression coefficients, particularly those for the wear are sensitive to the experimental conditions. This finding corresponds to the lower reproducibility of the results of the wear test. The data were then used to design the composition of brake pads. Optimization was conducted using a GA (Genetic Algorithm). The predicted friction data of the designed brake pads agreed well with the test data. The calculation time is about 1/700 of that by an alternative random search method.