A New Roundness Error Evaluation Method.

Roundness error is one of the important geometric error indicators, which directly affects the assembly accuracy and service life of mechanical parts. Aiming at the problems of the traditional non-linear model for evaluating the region with the smallest roundness error, the evaluation efficiency is low when the amount of extracted data is large, etc., a support vector machine-based evaluation method for the region with the smallest roundness error is proposed. The sampling points on the roundness are mapped to the high-dimensional space, so that the problem in the non-linear space is converted to the linear space, which greatly reduces the complexity of the algorithm. By comparing the roundness error evaluation results of the least squares method, simplex algorithm and support vector machine, the accuracy and feasibility of the evaluation method for the minimum roundness error region based on the support vector machine is verified, and it is found that it is processing huge The efficiency of data extraction can be applied to online roundness error evaluation in the production process. The research results have a certain promotion effect on improving the digital measurement and online detection technology of geometric errors of parts.