Input File Reference¶

All simulation parameters are defined in code_base/inputfile/input_param.txt. The file uses Fortran namelist format.

File Structure¶

The input file has two sections:

Geometry block (free-format, line-by-line)
Namelist blocks (&name ... / format)

Geometry Parameters¶

The geometry is defined by zones in each direction. Each zone has a length, number of control volumes, and power-law exponent for grid stretching.

nzx                          ! number of x-zones (max 7)
xzone(1), xzone(2), ...     ! length of each zone (m)
ncvx(1), ncvx(2), ...       ! number of CVs in each zone
powrx(1), powrx(2), ...     ! power-law exponents (1=uniform, >1=cluster at end, <0=cluster at start)

Same pattern repeated for y (nzy) and z (nzz).

Example (current setup):

Direction	Zones	Lengths	CVs	Exponents	Total
x	1	4.0 mm	400	1 (uniform)	400 cells
y	1	4.0 mm	400	1 (uniform)	400 cells
z	2	0.5 mm + 0.2 mm	10 + 40	-1.5, 1	50 cells

The z-direction uses two zones: a substrate zone (0.5 mm, coarse, clustered toward top) and a powder/build zone (0.2 mm, fine, uniform). Total domain: 4 mm x 4 mm x 0.7 mm.

Grid Stretching

Power-law exponent controls cell clustering:

powr = 1: uniform spacing
powr > 1: cells cluster toward the end of the zone
powr < 0: cells cluster toward the start of the zone (e.g., -1.5 clusters z-cells toward the substrate-powder interface)

Namelist Blocks¶

`&process_parameters` — Surface Laser Source¶

Parameter	Unit	Description
`alaspow`	W	Surface laser power (set to 0 when using volumetric source)
`alaseta`	-	Surface laser absorption efficiency
`alasrb`	m	Laser beam radius (1/e^2)
`alasfact`	-	Gaussian factor (typically 2 for 1/e^2 definition)

`&volumetric_parameters` — Volumetric Heat Source¶

Parameter	Unit	Description
`alaspowvol`	W	Volumetric laser power
`alasetavol`	-	Volumetric absorption efficiency
`sourcerad`	m	Source radius (Gaussian 1/e^2)
`sourcedepth`	m	Source penetration depth

Note

Either surface (alaspow) or volumetric (alaspowvol) source is used. Set the unused one to 0. The volumetric source distributes heat as:

\[q(x,y,z) = \frac{P \cdot \eta \cdot f}{\pi r_s^2 \cdot d_s} \exp\left(-\frac{f}{r_s^2}\left[(x-x_b)^2 + (y-y_b)^2\right]\right)\]

`&material_properties` — Primary Material (C=1)¶

Parameter	Unit	Description	Example
`dens`	kg/m^3	Solid density	8440
`denl`	kg/m^3	Liquid density	7640
`viscos`	Pa*s	Dynamic viscosity (liquid)	0.007
`tsolid`	K	Solidus temperature	1563
`tliquid`	K	Liquidus temperature	1623
`tboiling`	K	Boiling/vaporization temperature	2650
`hlatnt`	J/kg	Latent heat of fusion	290000
`acpa`	J/(kg*K^2)	Solid specific heat coefficient (quadratic term)	0.2441
`acpb`	J/(kg*K)	Solid specific heat coefficient (linear term)	338.59
`acpl`	J/(kg*K)	Liquid specific heat (constant)	709.25
`thconsa`	W/(m*K^2)	Solid thermal conductivity coefficient (linear term)	0.0155
`thconsb`	W/(m*K)	Solid thermal conductivity coefficient (constant term)	5.0435
`thconl`	W/(m*K)	Liquid thermal conductivity (constant)	30.078
`beta`	1/K	Thermal expansion coefficient (for buoyancy)	5e-5
`emiss`	-	Surface emissivity (for radiative loss)	0.3
`dgdtp`	N/(m*K)	dg/dT — thermal Marangoni coefficient	-3.8e-4

Specific Heat Model

Solid specific heat is temperature-dependent: $c_p(T) = a \cdot T + b$ where acpa = $a$, acpb = $b$. The enthalpy-temperature curve is:

Solid ($T \leq T_s$): $H = \frac{a}{2}T^2 + bT$
Mushy ($T_s < T < T_l$): linear interpolation between $H_s$ and $H_l$
Liquid ($T \geq T_l$): $H = H_l + c_{p,l}(T - T_l)$

`&powder_properties` — Powder Layer¶

Parameter	Unit	Description	Example
`layerheight`	m	Powder layer thickness	0.04e-3
`pden`	kg/m^3	Powder density (accounts for porosity)	4330
`pcpa`	J/(kg*K^2)	Powder specific heat (quadratic)	0.2508
`pcpb`	J/(kg*K)	Powder specific heat (linear)	357.7
`pthcona`	W/(m*K^2)	Powder conductivity (linear)	0
`pthconb`	W/(m*K)	Powder conductivity (constant)	0.995

Powder properties apply to cells in the top layerheight of the domain that have not yet been melted (solidfield <= 0.5).

`&numerical_relax` — Solver Parameters¶

Parameter	Unit	Description	Example
`maxit`	-	Max iterations per time step	60
`delt`	s	Time step size	2e-5
`timax`	s	Total simulation time	0.09
`urfu`	-	Under-relaxation factor for u-velocity	0.7
`urfv`	-	Under-relaxation factor for v-velocity	0.7
`urfw`	-	Under-relaxation factor for w-velocity	0.7
`urfp`	-	Under-relaxation factor for pressure	0.7
`urfh`	-	Under-relaxation factor for enthalpy	0.7

`&boundary_conditions` — Thermal Boundaries¶

Parameter	Unit	Description	Example
`htci`	W/(m^2*K)	Convection coefficient on x-faces (west/east)	5
`htcj`	W/(m^2*K)	Convection coefficient on y-faces (north/south)	5
`htck1`	W/(m^2*K)	Convection coefficient on bottom face	5
`htckn`	W/(m^2*K)	Convection coefficient on top face	5
`tempWest`	K	Far-field temperature at west boundary	293
`tempEast`	K	Far-field temperature at east boundary	293
`tempNorth`	K	Far-field temperature at north boundary	293
`tempBottom`	K	Far-field temperature at bottom boundary	293
`tempPreheat`	K	Initial/preheat temperature	293
`tempAmb`	K	Ambient temperature (for radiation)	293

The top surface uses combined convection + radiation: $$q_{loss} = h(T - T_{amb}) + \epsilon \sigma (T^4 - T_{amb}^4)$$

`&output_control` — Output Settings¶

Parameter	Unit	Description	Example
`outputintervel`	steps	VTK output frequency (every N time steps)	50
`case_name`	-	Case identifier (sets output directory name)	'testcase'
`toolpath_file`	-	Path to toolpath file (.crs)	'./ToolFiles/B26.crs'
`species_flag`	-	Enable species transport (0=off, 1=on). When `1`, secondary-material properties are read from `solver_species/inputfile/input_param_species.txt`	0
`predict_flag`	-	Enable field prediction by integer-cell shift (0=off, 1=on)	0

When predict_flag=1, the solver shifts enthalpy, velocity, and pressure fields by an integer number of cells in the scan direction before entering the iteration loop on heating steps (laser on, melt pool present). This provides a better initial guess and reduces iterations by approximately 38% during heating.

`&adaptive_mesh` — Adaptive Mesh¶

Parameter	Type	Description	Default
`adaptive_flag`	integer (0 or 1)	Enable adaptive mesh	0
`amr_dx_fine`	real (m)	Uniform fine-zone cell size	10.0e-6
`amr_init_half`	real (m)	Half-width of the initial fine-zone square (centred on the first laser-on waypoint)	50.0e-6
`amr_buffer_ratio`	real	Fine-zone half-extent / pool half-extent on each side. ≥ 1.0; pick > 1.2 to keep the bbox out of the geometric coarse ring	1.5
`amr_growth_ratio`	real	Geometric growth factor per coarse cell outside the fine zone	1.15
`amr_max_cell`	real (m)	Cap on coarse cell size; warning if exceeded	1.0e-4
`amr_max_ncv`	integer	Hard cap on per-axis interior cells (drives X/Y allocation)	250
`remesh_interval`	integer	Regrid every N timesteps	20

Behaviour when adaptive_flag=1:

The top-of-file X/Y zone block is ignored. ncvx(1) / ncvy(1) are overridden to amr_max_ncv at startup so the field arrays are sized for the worst case. The Z block always applies (Z is never remeshed).
Initial fine zone is a 2 × amr_init_half square centred on the first toolpath row with laser_on ≥ threshold.
After every remesh_interval steps, the asymmetric fine-zone half-extents become max(amr_init_half, alen_{tail|head|lo|hi} × amr_buffer_ratio) (clamped to the domain). This tracks the long-tail / short-head shape of moving-laser pools without truncating either side.
Outside the fine zone, coarse cells grow geometrically at amr_growth_ratio per cell. If a cell exceeds amr_max_cell the writer emits one [amr] WARNING line per remesh; consider raising amr_max_ncv or shrinking the domain.

Tuning notes. With amr_dx_fine = 10 μm, amr_max_ncv = 250, and a 2 mm domain, a steady-state pool of 800 μm × 200 μm is comfortably inside the fine zone with amr_buffer_ratio = 1.5. Drop amr_buffer_ratio to ~1.05 only if the CA-coupling writer's clip warning is silent — otherwise the bbox lands in the geometric coarse ring and assert_axis_uniform aborts.

`&mechanical_params` — Mechanical Solver¶

Parameter	Type	Description	Default
`mechanical_flag`	integer (0 or 1)	Enable mechanical solver	0
`mech_interval`	integer	Solve mechanical every N thermal steps	10
`mech_output_interval`	integer	Write mechanical VTK every N mechanical solves	5
`mech_mesh_ratio`	integer	FEM grid coarsening ratio vs thermal grid	2

When mechanical_flag=1, the EBE FEM solver computes residual stress and deformation from the thermal field. The FEM grid is coarsened from the thermal grid by mech_mesh_ratio (e.g., 200x200 thermal with ratio=2 gives 100x100 FEM nodes).

Parallel mode: use bash run.sh <case> <thermal_threads> <mech_threads> to run thermal and mechanical as separate processes. See Mechanical Solver for details.

Input File Reference¶

File Structure¶

Geometry Parameters¶

Namelist Blocks¶

&process_parameters — Surface Laser Source¶

&volumetric_parameters — Volumetric Heat Source¶

&material_properties — Primary Material (C=1)¶

&powder_properties — Powder Layer¶

&numerical_relax — Solver Parameters¶

&boundary_conditions — Thermal Boundaries¶

&output_control — Output Settings¶

&adaptive_mesh — Adaptive Mesh¶

&mechanical_params — Mechanical Solver¶