FDM 3D Printing PLA: Can You Balance Strength and Cost?

Choosing FDM 3D printing settings is not just a matter of appearance. The settings can affect how a printed part responds to tension, how long it takes to produce and how much filament it uses. This study examines those trade-offs for PLA specimens and develops models to predict three practical outcomes.

Key findings

• The study evaluated seven FDM process parameters in a definitive screening design with 17 runs, including a central point.
• Infill density and wall thickness were the most influential factors for maximum tensile force; layer height, print speed and build-plate temperature were also significant.
• Layer height, wall thickness and infill density affected all three responses: force, print time and material consumption.
• Print speed significantly affected maximum tensile force and print time.
• Adjusted R² values were 99.32% for force, 96.80% for time and 93.25% for material consumption.
• Maximum relative errors in confirmation tests were 5.98%, 3.60% and 4.33% for force, time and material, respectively.

Why examine these FDM settings?

Fused deposition modelling (FDM) is an additive manufacturing process that builds a part by depositing filament in successive layers. A setting that reduces production time may also change how deposited lines and layers bond. Increasing infill or wall thickness can require more material. Understanding the connections between mechanical response and production inputs helps engineers and product designers make informed choices when developing a printed part.

The study measures maximum tensile force, not a universal tensile strength value. Maximum tensile force is the highest force recorded while a specimen is pulled in a tensile test; it is affected by the specimen’s geometry as well as the material and print settings. The other two responses are printing time and material consumption, making the work relevant to both mechanical testing and production planning.

What did the study test?

The seven factors were print orientation, layer height, wall thickness, infill density, build-plate temperature, print speed and fan speed. The tested ranges included orientation from 0° to 90°, layer height from 0.1 to 0.2 mm, infill density from 20% to 80%, build-plate temperature from 30 to 90 °C, print speed from 30 to 70 mm/s and fan speed from 60% to 100%.

The study used a definitive screening design (DSD), an experimental approach intended to assess factor effects and support the development of predictive models. The design comprised 17 runs, including a central point. PLA specimens were designed to ISO 527-2 and printed on an Ultimaker S5. UltiMaker Cura was used to generate the G-code, and the specimens used a grid infill pattern.

Tensile tests followed ISO 527-1, using a Shimadzu AGS-X universal testing machine at a test speed of 5 mm/min. Maximum force was recorded with Shimadzu Trapezium-X software. Printing time and material consumption were obtained from UltiMaker Cura. The study analysed the results in Minitab 19, using analysis of variance (ANOVA) and regression analysis.

What did the results show?

For maximum tensile force, infill density and wall thickness were the most influential factors. Layer height, print speed and build-plate temperature were also significant. Within the tested conditions, increasing layer height, wall thickness and infill density was associated with an increase in maximum force. The paper relates higher infill and thicker walls to more material in the specimen, and higher build-plate temperature to improved adhesion between lower layers. Higher print speed was associated with lower maximum force; the paper explains this in terms of reduced adhesion between deposited rasters and layers.

The force model had an adjusted R² of 99.32% and an RMSE of 13.03 N. RMSE, or root mean square error, describes the typical size of prediction errors in the model’s response units. For print time, the initial model was inadequate, so the original times were transformed using the natural logarithm. The resulting model identified print speed, layer height, infill density and wall thickness as significant. Increasing print speed and layer height reduced predicted ln(T), while increasing infill density and wall thickness increased it. The adjusted R² was 96.80%, and the RMSE was 0.056 for the transformed response.

For material consumption, infill density was the strongest factor, followed by wall thickness and layer height. Increasing each of these factors increased predicted material use. This model had an adjusted R² of 93.25% and an RMSE of 0.276 g. These statistics describe the models developed from this study; they should not be treated as universal guarantees for different printers, filament materials or part geometries.

What do the findings mean in practice?

For engineers and designers, the main lesson is that settings should be selected against more than one objective. Increasing infill density or wall thickness may raise maximum tensile force, but it can also increase print time and material consumption. Increasing print speed can reduce predicted print time, but in this study it was also associated with lower maximum force. Layer height matters across all three responses, so it is useful to consider it alongside the other settings rather than optimize it in isolation.

For students and researchers, the paper demonstrates how DSD can connect factor screening with regression modelling. ANOVA helps identify statistically significant factors, and confirmation experiments provide a check on whether models predict outcomes for additional parameter combinations. Three additional specimens were printed using randomly selected settings. The maximum relative errors were 5.98% for maximum force, 3.60% for print time and 4.33% for material consumption.

Limitations and next steps

The findings apply to the PLA material, grid infill pattern, ISO 527-2 specimen geometry and FDM setup used in the work. The study examines seven factors, not every possible printer setting, material or part shape. Its results therefore do not establish how other filament materials or complex product geometries would perform. Model confirmation was carried out on three additional specimens, which is important context when considering predictions beyond the tested conditions.

Future work could test the models with other geometries, materials and settings, although those investigations are not part of this paper. In practice, the models are best treated as estimates within the investigated parameter ranges and checked with appropriate testing before being applied to a different production setup.

Conclusion

The study finds that layer height, wall thickness and infill density affect maximum tensile force, print time and material consumption in FDM-printed PLA specimens. Print speed significantly affects force and time. The paper reports regression models with high adjusted R² values and assesses their predictive accuracy through additional experiments. The paper was published in the NT 2024 proceedings. In a study by Adi Pandžić, PhD, regression models were developed and checked through additional experiments.

Frequently asked questions

Author: Adi Pandžić, PhD in Mechanical Engineering

This article summarises a published research paper; please refer to and cite the original paper. About the author

← All articles