PREDICTION OF LOCATOR PIN ERRORS IN FACING PROCESS USING NEURAL NETWORK AND SUPPORT VECTOR REGRESSION

Abstract

Facing is a common process in machining. Facing process error could be the result of a three-pin error in 3-2-1 principle. Measuring pin error is difficult to carry out periodically. This study discusses how to identify pin errors based on the objects deviation after facing using neural network and supporting vector regression. The facing process is carried out on two surfaces. The two processes separately modelled. First case is pin error value prediction based on object deviation at fifteen measuring points. Second case is pin error value prediction based on the deviation of the object at sixteen measuring points and the surface error represented by the pin error in the first case. The performance neural network model for case 1 and case 2 based on R2 score reaches 0.960 and 0.986. The support vector regression model for case 1 and case 2 reaches 0.9485 and 0.9921 of R2 score.

Keywords

• 3-2-1 principle
• facing
• neural network
• pin error
• support vector regression