Statistical evaluation of PMMA surface roughness

Abstract

There is increased enthusiasm in polymer materials and yet limited research on single point diamond turning of Polymethyl methacrylate (PMMA) used to produce contact lenses. This study is a presentation of a statistical-based PMMA surface roughness prediction and parameter significance. The data utilized is obtained during dry single point diamond turning of PMMA. The experiment has been designed with the Central Composite Design (CCD) method and the Response Surface Methodology (RSM) has been used for surface roughness prediction and evaluation of cutting parameter importance. The surface roughness data used for regression model generation has been acquired manually using a profilometer. The resultant surface roughness (Ra) dataset has been split into training and testing datasets. The accuracy of the model has been determined based on the Mean Absolute Percentage Error (MAPE). Results have indicated that the generated regression model can predict surface roughness with 75.12 % accuracy. Furthermore, the order of parameter importance in decreasing order is as follows: feed rate, cutting speed, and depth of cut. The paired t-test results indicate that the difference between the averages of measured surface roughness and predicted surface roughness is not big enough to be statistically significant. Hence, the technique can be reliably utilized in predicting surface roughness during single point diamond turning of PMMA.