Page 43 - IJAMD-2-3

International Journal of AI for
Materials and Design Biomimetic ML for AFSD aluminum properties

variables (e.g., temperature and displacement) within each The use of the element activation and deactivation
element. An implicit temporal integration approach is technique in the numerical model enables accurate
widely employed to enhance numerical stability in coupled simulation of the dynamic nature of the AFSD process.
thermo-mechanical analyses. Accurate simulation also This method provides valuable insights into the thermal
requires the application of adequate boundary and initial and mechanical behavior of the deposited material, aiding
conditions. in the prediction of material properties, the identification
The activation and deactivation of the elements of possible flaws, and the assessment of overall structural
are governed by the activation function A (t), which performance (Figure 3). Additional simulation views
c
determines whether an element is active or inactive at a and results for other alloy systems are provided in
given time, as shown in Equation III: Supplementary File (Figure S1-S5 and Video S1).
1 if elemente is active at time t. 3.2. Prediction of von Mises stress and logarithmic
At (III) strain in additive friction stir deposited walled
c
0 if elemente is inactive at timme t. structures using GA-coupled ML algorithms
The heat transfer equation for active elements is Von Mises stress and logarithmic strain are critical
modified to incorporate the activation function and heat parameters in analyzing and optimizing the AFSD process.
input from deposition, as given in Equation IV: Von Mises stress is a widely used yield criterion that

T estimates the onset of yielding in materials under complex
c . t . kT QA tH t.( ) (IV) loading conditions. In AFSD, which involves severe plastic
p
c
deformation, high strain rates, and elevated temperatures,
where H (t) is the heat input modulated byA (t). evaluating von Mises stress is essential for determining
c
Similarly, the structural response for active elements is whether the material will yield during deposition. This,
represented by Equation V: in turn, is critical for maintaining strong interlayer
bonding while avoiding material failure. In addition, von
2
u Mises stress offers information about material ductility
)
.( . At f b . t 2 (V) and flow characteristics, which are key to achieving the
c

Figure 3. Thermal and mechanical responses of the alloys during the deposition process
Abbreviations: GRADT: Temperature gradient; LE: Logarithmic strain.

Volume 2 Issue 3 (2025) 37 doi: 10.36922/ijamd.5014