The ANN analysis and Taguchi method optimisation of the brake pad composition

Abstract

In this study, the Taguchi optimisation technique was utilised to establish the composition of an asbestos-free brake pad. As a result of optimisation outputs, 18 various compositions were obtained. The produced specimens according to outputs were subjected to friction assessment and screening tests to evaluate their average friction coefficient and friction surface temperature. By assessing the results, the composition of brake pads, which can be used in vehicle brakes, was determined. The experimental results were also examined by variance analysis to demonstrate effect factors. Furthermore, when the test results were evaluated in terms of the friction coefficient, wear rate and temperature of the brake disc, a high coefficient of friction, a moderate wear rate and an acceptable temperature of the brake disc were obtained in sample 13. The composition of sample 13, which has the highest friction coefficient, is 3 g brass powder, 9 g cashew, 9 g carbon fibre, 9 g copper dust and 9 g graphite. Lastly, the average friction coefficients of brake pads were estimated through artificial neural network (ANN) analysis. With respect to the results, it was discovered that the generated ANN model promises a high ratio estimation capacity.