Page 117 - MSAM-4-3

Materials Science in Additive Manufacturing Sustainable manufacturing composite material optimization

and process monitoring. In this study, a twin-screw 3D printer to dynamically optimize process parameters,
extruder was utilized to manufacture TPU 95A filaments, thereby minimizing the need for manual parameter
with real-time energy consumption tracked using a power tuning. By combining data-driven AI models and expert-
meter. This allows sustainability measures to be closely driven MCDM techniques, this study presents a logical
tied into the MCDM framework. Experimental data from and sustainable decision-making approach for FDM-
mechanical testing and energy monitoring served as based impeller production, with enhanced process efficacy,
inputs to the Fuzzy AHP-TOPSIS model, with each metric material utilization, productivity, and environmental
carefully analyzed to determine the optimum process performance. The architecture of the implemented ANN
conditions. By applying the hybrid MCDM approach, model is illustrated in Figure 6, illustrating the data flow
optimal trade-offs among mechanical strength, energy from FDM parameters to predicted mechanical and
efficiency, and print accuracy were achieved, facilitating energy-related outputs.
high-performance, sustainable manufacturing of impellers.
The Fuzzy AHP-TOPSIS model was integrated with AI 3. Results and discussion
regression techniques, making it more appropriate for Mechanical performance, energy consumption, and
handling a vast array of impeller geometries and materials. AI-MCDM optimization of FDM-based 3D-printed TPU
The model incorporates real-time sensor feedback from the 95A impellers were comprehensively examined. All aspects
of the process, from filament extrusion to 3D printing
and mechanical testing, were meticulously monitored
for energy consumption to evaluate overall sustainability.
Optimization of FDM printing parameters, such as layer
thickness, infill density, shell thickness, and print speed,
was performed using AI-based predictive modeling and
MCDM methods, namely, Fuzzy AHP and TOPSIS. The
mechanical properties of the impellers were evaluated
through tensile, flexural, and compressive tests, while wear
resistance was assessed using a pin-on-disc tribometer. The
findings indicate that increasing shell thickness and infill
density enhances wear resistance and mechanical strength,
but also increases energy consumption. Therefore, an
optimal trade-off strategy must be identified to achieve
sustainable manufacturing. The test results are presented
in Table 1.
Figure 4. Scatter plot comparing artificial intelligence-predicted and The mechanical characteristics of TPU 95A impellers
experimentally measured compression strength
were determined based on the infill density and shell
thickness from the tensile, flexural, and compression
strength tests. Sample S1, with a layer thickness of 0.1 mm,
infill density of 20%, and a shell thickness of 0.8 mm,
exhibited poor mechanical properties, characterized by

Figure 5. Scatter plot comparing artificial intelligence-predicted and Figure 6. Architecture of the artificial neural network model used
experimentally measured wear rates in this study

Volume 4 Issue 3 (2025) 9 doi: 10.36922/MSAM025200033