ioandanielc commited on
Commit
c549fe7
·
verified ·
1 Parent(s): 0d043a5

Delete README.md

Browse files
Files changed (1) hide show
  1. README.md +0 -136
README.md DELETED
@@ -1,136 +0,0 @@
1
- ---
2
- license: cc-by-4.0
3
- task_categories:
4
- - image-classification
5
- - image-to-image
6
- - tabular-regression
7
- language:
8
- - en
9
- tags:
10
- - additive-manufacturing
11
- - laser-powder-bed-fusion
12
- - melt-pool
13
- - smoothed-particle-hydrodynamics
14
- - keyhole
15
- - generative-machine-learning
16
- - active-learning
17
- - bayesian-optimization
18
- - simulation
19
- pretty_name: SPH Melt Pool Dataset for Single-Track LPBF
20
- size_categories:
21
- - 10K<n<100K
22
- ---
23
-
24
- # SPH Melt Pool Dataset for Single-Track Laser Powder Bed Fusion
25
-
26
- **Paper:** A Simulation-Based Dataset for Melt Pool Dynamics in Single-Track Laser Powder Bed Fusion
27
-
28
- **Authors:** Ioan-Daniel Craciun, Stefan Adami, Felix Dietrich
29
- **Institution:** Technical University of Munich (TUM)
30
- **License:** CC BY 4.0
31
-
32
- ---
33
-
34
- ## Dataset Description
35
-
36
- This dataset contains SPH-simulated melt pool dynamics for single-track laser powder bed fusion (LPBF) experiments on Ti-6Al-4V (TI64). It is designed to enable machine learning research on melt pool geometry, regime classification (conduction vs. keyhole), and generative modeling of subsurface melt pool cross-sections.
37
-
38
- The dataset spans the full process parameter space from stable conduction regimes to hazardous keyhole formation, providing the ground truth needed to learn the boundary between safe and defect-prone process conditions.
39
-
40
- ---
41
-
42
- ## Dataset Structure
43
-
44
- Each experiment corresponds to one simulation run at a unique parameter point. Experiments are named by their parameter values:
45
-
46
- ```
47
- P-{laser_power}_VX-{scan_speed}_LS-{spot_size}_ST-{substrate_temp}_M-TI64_..._H-{hash}/
48
- ├── png_files/ # Cross-sectional images (top, side, front views)
49
- ├── monitor/ # Scalar monitoring time series (.dat files)
50
- ├── gif_files/ # Animated GIFs of melt pool evolution
51
- ├── labels.csv # Per-frame phase labels (timestep, label)
52
- └── *.json # Simulation configuration file
53
- ```
54
-
55
- ### Images
56
-
57
- Three cross-sectional perspectives per timestep, rendered in false color:
58
- - 🔴 Red — melt pool (liquid TI64)
59
- - 🟢 Green — solid substrate (TI64)
60
- - 🔵 Blue — gas atmosphere (Argon)
61
-
62
- | Perspective | Resolution |
63
- |-------------|------------|
64
- | Top (XY) | 256 × 256 px |
65
- | Side (XZ) | 512 × 256 px |
66
- | Front (YZ) | 256 × 256 px |
67
-
68
- ### Phase Labels
69
-
70
- Each frame is assigned one of four mutually exclusive labels:
71
-
72
- | Label | Description |
73
- |-------|-------------|
74
- | `Initial Emptiness` | Domain prior to laser arrival |
75
- | `Forming Phase` | Melt pool nucleation and growth |
76
- | `Convection` | Steady-state Marangoni-driven flow |
77
- | `Keyhole` | Deep vapor depression regime |
78
-
79
- > ⚠️ Labels are auto-initialized via a delta-z heuristic and pending human verification. The `labels.csv` file contains headers only for experiments not yet reviewed.
80
-
81
- ### Scalar Monitoring Data
82
-
83
- Per-timestep `.dat` files in `monitor/`:
84
-
85
- | File | Description |
86
- |------|-------------|
87
- | `maximum-temperature_melt.dat` | Peak melt pool temperature |
88
- | `average-temperature_melt.dat` | Mean melt pool temperature |
89
- | `particle-number_melt.dat` | Proxy for melt pool volume |
90
- | `position-bounds_melt.dat` | Spatial extent (depth, width) |
91
- | `transferred-laser-power_melt.dat` | Absorbed energy per timestep |
92
- | `kinetic-energy_melt.dat` | Marangoni convection intensity |
93
- | `time.dat` / `dt.dat` | Temporal metadata |
94
-
95
- ---
96
-
97
- ## Process Parameters
98
-
99
- | Parameter | Symbol | Min | Max | Unit |
100
- |-----------|--------|-----|-----|------|
101
- | Laser power | P | 76.70 | 249.99 | W |
102
- | Scan speed | v | 0.204 | 0.998 | m/s |
103
- | Laser spot radius | r_s | 45.10 | 89.70 | µm |
104
- | Substrate temperature | T_s | 300.3 | 399.6 | K |
105
-
106
- Sampling: uniform random across the 4D parameter space.
107
-
108
- ---
109
-
110
- ## Intended Uses
111
-
112
- - Melt pool regime classification (conduction vs. keyhole)
113
- - Generative modeling of melt pool morphology
114
- - Melt pool geometry regression (width, depth)
115
- - Initialization for Bayesian optimization-based active learning
116
-
117
- ## Out-of-Scope Uses
118
-
119
- - Multi-track or powder bed LPBF (bare substrate only)
120
- - Materials other than Ti-6Al-4V
121
- - Direct experimental validation without sim-to-real transfer
122
-
123
- ---
124
-
125
- ## Citation
126
-
127
- ```bibtex
128
- @dataset{craciun2026sph,
129
- title = {A Simulation-Based Dataset for Melt Pool Dynamics in Single-Track Laser Powder Bed Fusion},
130
- author = {Craciun, Ioan-Daniel and Adami, Stefan and Dietrich, Felix},
131
- year = {2026},
132
- publisher = {HuggingFace},
133
- url = {https://huggingface.co/datasets/ioandanielc/sph_dataset},
134
- license = {CC BY 4.0}
135
- }
136
- ```