TY  - JOUR
AU  - A Rahul Kumar
AU  - N Karthikeyan
AU  - Dr. L Balasubramanyam
AU  - Dr. K R Yellu Kumar
AU  - P Ramu
PY  - 2026
DA  - 2026/03/15
TI  - An AI-Integrated CAD/CAM Framework for Predictive Surface Metrology in CNC Milling Using Synthetic Data Augmentation
JO  - Global Journal of Engineering Innovations and Interdisciplinary Research
VL  - 6
IS  - 3
AB  - The development of Industry Version 4.0 requires businesses to use Artificial Intelligence (AI) as a
core technology for the Computer Aided Design and Manufacturing (CAD/CAM) systems, to generate
virtual models. Although the traditional CAM systems develop toolpaths based on geometric shapes,
they cannot measure how the cutting operations actually function because they disregard the tool
vibrations and material defects that determine how surfaces will be developed. This study focuses on
the development of an advanced predictive system that uses Artificial Neural Networks (ANN) to predict
surface roughness (Ra) based on three cutting parameters: cutting speed (v), feed rate (f), and depth
of cut (ap). The smart manufacturing field faces its most difficult challenge in the form of the "small
data" problem, as physical experiments require excessive financial resources. We implemented a data
approach that merged real industrial data with synthetic data generated using artificial intelligence
methods. We increased the experimental base from 45 data points to 500 data points, which created
sufficient variance for the development of robust models without needing extensive physical testing.
The supervised learning system achieved a coefficient of determination (R2) score of 0.96, which means
it correctly predicted surface quality 96% of the time. The AI-integrated framework showed a 20%
decrease in prediction errors when tested against standard CAM software systems. The study shows
that virtual metrology components integrated into CAD/CAM systems provide users with manufacturing
design feedback that operates in real time, thus decreasing material waste while reducing the need for
repeated testing in precision engineering processes.
SN  - 3066-1226
UR  - https://dx.doi.org/10.33425/3066-1226.1276
DO  - 10.33425/3066-1226.1276