Prediction Of Surface Roughness In Additively Manufactured Samples

Artificial intelligence (AI) and additive manufacturing (AM) are excellent and revolutionary technologies. The most recent technology for producing objects through layer-over-layer deposition is called additive manufacturing (AM). Based on the thermal analysis of bonding formation in the 3D printed parts, the mechanism model of surface roughness was established. Most of the printing parameters, including infill density (ID), printing speed (PS), nozzle temperature (NT), and layer height (LH), deposition road width, and printing platform temperature, were considered to ameliorate the surface morphology of printed parts. The main objective of this research work is to predict surface roughness in additively manufactured processes in PLA+ polymer material by using different machine learning algorithms like linear regression, support vector machine (SVM), and two ensemble learning techniques: – Xtreme gradient boosting (XGBoost) and random forest regressor and also characterization of the material . Taguchi’s Design of the Experiment was used to make L25 orthogonal array sample datasets, compare all the machine learning algorithms to see which one has the best model-fit accuracy. The machine model works on five key input parameters that influence layer geometries: layer height (LH), infill density (ID), printing speed (PS), and nozzle temperature (NT) with a 00 raster angle. By applying all ML algorithms, random forest regression is the best model, which gives 94.85% accurate results in the datasets with a minimum mean squared error of approx. 0.3756 and a maximum r2_score of approx. 0.92154. XRD shows the PLA+ material is semi crystalline material and the peak is about to 180 with 2𝜃.