An effort towards predicting the mechanical and tribological performance of FDM-printed PLA using ANN and regression models

This study investigates the effects of three fused deposition modelling (FDM) process parameters, build orientation (0°, 45°, and 75°), infill density (40%, 70%, and 100%), and layer height (0.1, 0.2, and 0.3 mm), on the tensile strength and specific wear rate of polylactic acid (PLA). A three-level full-factorial design comprising 27 parameter combinations was employed, with tensile and wear testing conducted according to ASTM D638 and ASTM G99, respectively. Two single-output artificial neural network (ANN) models with a 3-5-1 architecture were compared with multiple linear regression (MLR) and quadratic response surface methodology (RSM) models using a 19/8 hold-out split and five-fold cross-validation. Input and target scaling for the ANN was fitted exclusively to each training subset. The highest tensile strength in the original factorial dataset was 59.666 MPa at 0° orientation, 100% infill density, and 0.1 mm layer height. The lowest specific wear rate in the original factorial dataset was 17.932 × 10 −4 mm 3 /(N·m) at 45° orientation, 100% infill density, and 0.1 mm layer height. Build orientation had the largest main effect on tensile strength, whereas infill density had the largest main effect on wear. Quadratic RSM achieved pooled out-of-fold coefficients of determination (R 2 ) of 0.982 and 0.942, compared with 0.946 and 0.892 for ANN. These results favour quadratic RSM for predicting condition-level mean responses within the investigated domain. Contour-based screening identified a candidate printing window of 0°–15° orientation, 92%–100% infill, and 0.10–0.11 mm layer height for high tensile strength and low specific wear rate. Three additional experiments within this window gave mean tensile strengths of 55.93–60.01 MPa and mean specific wear rates of 19.38 to 19.73 × 10 −4 mm 3 /(N·m). The selected contour regions target TS > 50 MPa and specific wear rate <22.5 × 10 −4 mm 3 /(N·m).

Authors

Institutions

Publication Details

Journal
Journal of Thermoplastic Composite Materials
Published
2026-11-21
DOI
https://doi.org/10.1177/08927057261511981
Primary Topic
Additive Manufacturing and 3D Printing Technologies
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

An effort towards predicting the mechanical and tribological performance of FDM-printed PLA using ANN and regression models

Yogesh Shrivastava, Dalvir Singh, Mayur Pratap Singh
Journal of Thermoplastic Composite Materials
Additive Manufacturing and 3D Printing Technologies
article

An effort towards predicting the mechanical and tribological performance of FDM-printed PLA using ANN and regression models

Yogesh Shrivastava, Dalvir Singh, Mayur Pratap Singh
article en

Abstract

This study investigates the effects of three fused deposition modelling (FDM) process parameters, build orientation (0°, 45°, and 75°), infill density (40%, 70%, and 100%), and layer height (0.1, 0.2, and 0.3 mm), on the tensile strength and specific wear rate of polylactic acid (PLA). A three-level full-factorial design comprising 27 parameter combinations was employed, with tensile and wear testing conducted according to ASTM D638 and ASTM G99, respectively. Two single-output artificial neural network (ANN) models with a 3-5-1 architecture were compared with multiple linear regression (MLR) and quadratic response surface methodology (RSM) models using a 19/8 hold-out split and five-fold cross-validation. Input and target scaling for the ANN was fitted exclusively to each training subset. The highest tensile strength in the original factorial dataset was 59.666 MPa at 0° orientation, 100% infill density, and 0.1 mm layer height. The lowest specific wear rate in the original factorial dataset was 17.932 × 10 −4 mm 3 /(N·m) at 45° orientation, 100% infill density, and 0.1 mm layer height. Build orientation had the largest main effect on tensile strength, whereas infill density had the largest main effect on wear. Quadratic RSM achieved pooled out-of-fold coefficients of determination (R 2 ) of 0.982 and 0.942, compared with 0.946 and 0.892 for ANN. These results favour quadratic RSM for predicting condition-level mean responses within the investigated domain. Contour-based screening identified a candidate printing window of 0°–15° orientation, 92%–100% infill, and 0.10–0.11 mm layer height for high tensile strength and low specific wear rate. Three additional experiments within this window gave mean tensile strengths of 55.93–60.01 MPa and mean specific wear rates of 19.38 to 19.73 × 10 −4 mm 3 /(N·m). The selected contour regions target TS > 50 MPa and specific wear rate <22.5 × 10 −4 mm 3 /(N·m).

Journal of Thermoplastic Composite Materials
Galgotias University (IN), Galgotias College of Engineering & Technology (IN)
Additive Manufacturing and 3D Printing Technologies
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon