Page 110 - MSAM-4-3

Materials Science in Additive Manufacturing Sustainable manufacturing composite material optimization

intricate geometries, and low material waste. FDM entails the FDM printer, providing valuable insights into energy
passing a thermoplastic filament through a heated extruder efficiency across various process parameters. This approach
nozzle, where it is melted and pushed layer-wise onto a facilitates the identification of optimal printing conditions
build platform. As every new layer is deposited, it melts that minimize power consumption while maintaining
into the one below it, building up gradually to the shape structural integrity and mechanical performance. Although
of the desired 3D part. Important process parameters, AI has been utilized extensively for process automation and
including layer thickness, infill density, printing speed, and defect prediction, limited research has utilized real-time
nozzle temperature, significantly influence the mechanical monitoring of energy consumption within AI-MCDM
performance, surface finish, and energy efficiency of the platforms for sustainable impeller production.
printed object. FDM is widely applied in multiple industries, This study proposes a theoretical AI-assisted MCDM
such as aerospace, automotive, biomedical, consumer approach of FDM process parameter optimization in
products, and rapid prototyping, due to its capability to the production of composite impellers. This approach
produce light, personalized, and functionally graded parts. systematically analyzes the effect of significant FDM
Regardless of its extensive utilization, maintaining uniform parameters – including layer thickness, infill density,
mechanical quality and energy-efficient production nozzle temperature, and print speed – on mechanical
remains a challenge, especially for geometrically complex properties, energy efficiency, and material usage. ML
components such as impellers. The efficiency of an algorithms, combined with fuzzy AHP and TOPSIS,
1,2
impeller is a key factor influencing energy consumption were employed to support multi-objective decision-
and operational performance; therefore, optimizing its making. Mechanical characterization through tensile,
manufacturing processes to enhance mechanical strength, flexural, wear, and compression testing, along with
minimize material wastage, and improve energy efficiency scanning electron microscopy analysis, validates the
is of significant interest. However, the optimization of FDM optimized parameters. Real-time power monitoring
process parameters is typically conducted using trial-and- enables quantitative assessment of energy consumption,
error methods, which are inefficient with complex trade- reinforcing the sustainability dimension of the process. The
offs in mechanical performance, energy consumption, results could benefit sustainable AM practices, optimizing
and sustainability. This limitation underscores the need production performance and environmental footprint in
for artificial intelligence (AI) and machine learning (ML)- fluid-handling applications.
based methods in optimizing FDM-based manufacturing
of impellers. These advances have facilitated data-driven 1.1. AI-driven optimization of FDM
process optimization, allowing real-time decision- The convergence of ML and AI with AM has significantly
making and predictive modeling of AM. Studies have advanced process optimization, in-process monitoring,
demonstrated that mechanical properties can be predicted and defect detection, especially for FDM, the most common
with reasonable accuracy using ML algorithms based on AM process. FDM is valued for its cost-effectiveness,
process parameters, thereby improving print quality and simplicity, and ability to create complex geometries
minimizing defects. 3-5
with minimal material wastage. However, optimizing
9
In addition, MCDM techniques, such as fuzzy analytic FDM-produced parts remains challenging due to the
hierarchy process (AHP) and the Technique for Order of interdependencies between different process parameters,
Preference by Similarity to Ideal Solution (TOPSIS), have such as infill density, cooling rate, nozzle temperature, print
been used extensively to compare and rank various process speed, and layer thickness. These parameters collectively
conditions based on different performance measures, such influence mechanical strength, surface finish, energy
as mechanical integrity, energy efficiency, and material utilization, and material consumption. Conventional
efficiency. While AI and ML have made considerable optimization methods, including empirical modeling
6-8
advances, limited research works have utilized AI-based and trial-and-error, are non-adaptive, making it even
10
MCDM platforms to optimize FDM processing for the more challenging to balance sustainability, production
production of impellers. With the increasing demand for efficiency, and quality. AI optimization techniques – such
sustainable, lightweight, and high-performance impellers, as artificial neural networks (ANNs), convolutional neural
there is a need for a holistic AI-based optimization strategy networks (CNNs), decision tree algorithms, support
to address the United Nations’ Sustainable Development vector machines (SVMs), and reinforcement learning (RL)
Goals (SDGs). Minimizing energy consumption during – have demonstrated significant promise in optimizing
printing is a critical aspect of sustainability in the FDM- FDM processes by enabling predictive modeling, real-
based production of impellers. In this study, a power meter time monitoring, and adaptive control. 11-13 Zhou et al.
14
was utilized to record real-time energy consumption of provided a comprehensive review of AI applications in

Volume 4 Issue 3 (2025) 2 doi: 10.36922/MSAM025200033