Soft-Sensing Model on the Roughness of Machining Surface under the Numerical Control and its Application

Abstract

In order to make sure a high-accuracy and fast- speed survey, a Soft-sensing model for the roughness of machining surface was built based on the support vector machines using rotate speed n, feed peed vf, and depth of cutting as independent parameters, taking groups of actual machining experiment data as samples.The allowable error ε and the positive aligned c and the kernel function parameter r were optimized by an adaptive genetic algorithm. After being optimized 300 steps, the following results can be gained through the training, testing and application. The average relative error tended to saturation training was 4.0%; the test error was less than 2.6%; the average relative error between the Soft-sensing value for the roughness of machining surface under the numerical control and the test value of the profile and roughness tester for the SV-C3000 super surface of was ranging from 0.4% to 1.25%.