Kriging versus Bezier and regression methods for modeling and prediction of cutting force and surface roughness during high speed edge trimming of carbon fiber reinforced polymers

摘要

Although CFRP materials are gaining popularity, their processing is still poorly understood. Accurate and repeatable models related to the cutting mechanism of CFRPs remains challenging. This work represents an investigation into the use of parametric models as an alternative approximation technique for modeling cutting force and surface roughness during high speed edge trimming of CFRP. Three predictive models were described and compared, some of which have been rarely used in this application. The three developed models were evaluated in terms of their accuracy and efficiency using new set of data (external data) and four evaluation indices. Result shows that Bezier models outperform regression models and can be used as alternative for modeling and prediction of machining responses. Although kriging model predict the internal data accurately, it breaks down when using external data. The new selected model can help the manufacturing industries for monitoring the cutting process of CFRPs. (C) 2019 Elsevier Ltd. All rights reserved.