Page 35 - IJAMD-1-3

International Journal of AI for
Materials and Design
Prediction of wall geometry for wire arc additive manufacturing

A B

Figure 8. Two walls manufactured with optimal (A) and conventional
(B) parameters

machining depth. The main information is to have enough
matter for machining to get the right geometry and also a
good health matter.
Moreover, the distance between beads was also
optimized based on the BW. Previous studies suggest
1,35
that utilizing three or four beads does not substantially
Figure 7. Response surface plot of Dt depending on the number of layers influence the predictions of height and W for a specific
and predicted incremental height
interlayer dwell time. However, heat accumulation varies
with cumulative energy inputs, indicating the importance
4. Discussion of integrating geometry into the prediction models.
The integration of machine-learning models enhances the These findings align with the results of previous studies
predictive capability of current approaches for multilayer emphasizing the importance of thermal management and
bead geometry. Dt emerges as a critical parameter process parameter optimization in WAAM for producing
influencing microstructure and dimensional stability, as high-quality metal components.
highlighted by Turgut. This study provides empirical Finally, in a related study, Hu et al. investigated the
34
36
evidence demonstrating that varying interlayer dwell prediction of welding parameters for various layer heights
times can substantially influence the final quality of in robotic WAAM. Their model accurately predicted the
parts fabricated using WAAM. These considerations are required parameters, enhancing adaptability and precision.
particularly critical for controlling heat input, which is a Wang and Xue conducted WAAM experiments on 316L
37
pivotal factor in maintaining the structural integrity and stainless steel, maintaining a constant deposition rate
dimensional accuracy of manufactured components. In while varying arc modes. Their results demonstrated that
the same study by Turgu, three samples were fabricated SpeedArc and SpeedPulse manufacturing processes were
through continuous deposition with interlayer dwell times stable and efficient, revealing correlations between arc
of 60 s and 120 s. Findings revealed that the interlayer mode, microstructure, and mechanical properties. Wahsh
dwell time effectively controlled temperature fields, which et al. focused on selecting parameters for the WAAM
38
in turn influenced the microstructural and mechanical process, emphasizing the importance of identifying optimal
properties of the parts. For instance, an increased dwell settings to achieve the desired outcomes. Their study
time resulted in greater hardness and yield strength, offered comprehensive guidelines for enhancing process
highlighting the importance of thermal management in efficiency and part quality. Kumar et al. performed a
17
the WAAM process. Based on the microstructural analysis parametric study and characterization of steel structures
results of our models, microstructures can be predicted. fabricated using WAAM. They identified key parameters
These outcomes emphasize the importance of optimizing influencing mechanical properties and dimensional
interlayer dwell time to achieve the desired material accuracy, also providing a detailed analysis of various
properties. settings. Patel and Savsani utilized a multiobjective
39
Furthermore, the analysis extends to multibead wall improved teaching–learning-based optimization
production, focusing on height and W. The optimal model algorithm to optimize multiple objectives simultaneously.
for predicting wall height incorporates TS, Dt, and their This algorithm demonstrated substantial improvements in
interactions, indicating a parabolic relationship wherein WAAM process optimization. Collectively, these studies
height initially increases with these parameters before highlight the critical role of parameter optimization in
subsequently decreasing. Figure 8 illustrates two wall improving the WAAM process, leading to enhanced
surfaces: one manufactured using optimal parameters and mechanical properties, dimensional accuracy, and overall
another manufactured using conventional parameters. part quality. However, the interlayer dwell time is rarely
This indicates the possibility of predicting W and surface investigated, despite being a key factor in part production.

Volume 1 Issue 3 (2024) 29 doi: 10.36922/ijamd.4285