PSI - Issue 68

5

Ilia Nikitin et al. / Procedia Structural Integrity 68 (2025) 24–31 I. Nikitin et al. / Structural Integrity Procedia 00 (2025) 000–000

28

(a)

(b)

Fig. 2. The scheme of heat conduction problem for selective laser melting (SLM).

However, the boundary conditions in this case are mathematical complex, and the numerical problem is difficult due to the moving phase boundary. In the present work, the heat conduction problem is formulated in terms of enthalpy, as proposed by White (1982). This formulation allows for the implementation of a comprehensive calculation algorithm without focusing on the phase transition boundary. A variety of laser parameters are investigated. The problem was solved using an explicit scheme, and the relationship between laser beam parameters and the geometry of the solidified tracks was obtained for both single and multiple tracks. Based on the results of the numerical simulation, such local zones were interpreted as inhomogeneous material with different mechanical properties. The re-melt zone is subjected to additional heat treatment that breaks a primary boundary between grains and leads to a coarse microstructure, resulting in elevated mechanical properties. The non melt zone has weak links between neighboring volumes of material and associated with voids or material with a low elastic modulus. Such microstructural features can be simulated by local volumes of material with high or low elastic moduli. Solving the heat conduction problem provides a prediction for initial defect distribution in the SLM material and its nature. These results are then used for predicting the fatigue life of the SLM material. 3. Numerical algorithms The numerical algorithm for fatigue life prediction is based on the solution to the elastic problem. The coefficients (9) – (10) are calculated based on the corresponding equivalent stress (11) and (12). The value of the damage function at a given time layer n+1 and at a given spatial node k is calculated using equation (13), based on the value of the damage function and the coefficients calculated at the previous time layer n . (13) The increment of number of cycles ∆ is limited by the conditions of stability and convergence for the chosen difference scheme. ( ) ( ) !"#! $ ! % ! ! ! % ! ! ! " ! # ! $ $ $ ! ! " " ! # $ % # $ % &  + = # # # # ( &  &  &  &  ) * ) * !

!

( ) ! #

" !

$ & & )

%   *

!

#

(14)

! " ! " ! # #

!

!

" !

#

(

=

$

$

( ) ! #

" !

!

"

$