davide
d62dfd13a8
- add skill package and SKILL.md with AM workflow, guardrails, and output structure - add technical reference corpus (DfAM, fatigue, defects, process parameters, compliance, cost) - add materials-db.json with polymer/metal data, roughness/post-processing ranges, and selection guides - add CLI tools: select_material.py and postprocess_route.py for material ranking and post-processing route generation
5.1 KiB
5.1 KiB
Metal AM Alloys
Primary database:
materials-db.json— contains all structured data (UTS, YS, elongation, density, heat treatment, accuracy, shrinkage, applications, warnings) for all AM metal alloys. This file provides decision context and critical notes that cannot be structured in JSON.
How to use the JSON database for metals
To select a metal alloy:
1. Filter by T_max_service (e.g. >300°C → titanium or superalloys)
2. Filter by process (LPBF / EBM / Binder Jetting)
3. Compare strength-to-weight ratio (UTS/density) if weight is critical
4. Check biocompatible for medical applications
5. Read heat_treatment — complexity and cost impact lead time
6. Read warnings — some alloys have mandatory requirements
Critical notes on heat treatment (DO NOT ignore)
Heat treatment is part of the process, not optional
- AlSi10Mg: Stress relief BEFORE removing from build plate. Without it: distortion and cracking.
- Ti-6Al-4V: Stress relief 650°C mandatory. HIP mandatory for biomedical.
- 17-4PH: H900 aging (480°C/1h) MANDATORY. AS-BUILT properties are ~40% of H900.
- IN718: Full solution + double aging cycle mandatory. Plan weeks in advance.
- IN625: Simpler — stress relief only. No precipitation hardening.
Universal LPBF sequence (do not deviate)
- Stress relief → 2. Build plate removal → 3. Support removal → 4. HT/HIP → 5. Machining → 6. Inspection
Selection by strength-to-weight ratio
| Alloy | UTS/density (MPa·cm³/g) | Note |
|---|---|---|
| AlSi10Mg | ~160 | Excellent for lightweight structures |
| Scalmalloy | ~195 | Best Al available in AM |
| Ti-6Al-4V | ~225 | Aerospace benchmark |
| IN718 | ~135 | High density — justified by elevated temperatures |
| 17-4PH | ~155 | High-strength stainless steel |
Lot-to-Lot Variability and Property Scatter
Inter-lot variability in metal AM is higher than in forged material — and is often underestimated during the design phase. Do not design to the mean value from data tables: use P10 values (10th percentile) or apply an explicit knockdown factor.
Typical variability by alloy (CoV = coefficient of variation)
| Alloy / Condition | UTS CoV | Fatigue CoV | Primary source |
|---|---|---|---|
| Ti-6Al-4V LPBF as-built | 5–10% | 20–35% | Variable micro-porosity between builds |
| Ti-6Al-4V LPBF HIP + machined | 2–4% | 8–15% | HIP drastically reduces scatter |
| AlSi10Mg LPBF as-built | 8–15% | 25–40% | Highly sensitive to powder moisture and O₂ |
| 316L LPBF as-built | 4–7% | 15–25% | Ductile → low UTS scatter, moderate fatigue |
| 17-4PH LPBF H900 | 5–9% | 15–25% | Depends on aging cycle: temperature control critical |
| IN718 LPBF (full HT) | 5–8% | 18–28% | Variable carbide distribution between builds |
| PA12 SLS | 6–12% | 20–30% | Fresh/recycled powder ratio critical |
| PA12 FDM | 15–25% | 30–50% | Anisotropy + filament moisture |
Source: aggregated literature (Sames 2016, Lewandowski 2016, Gu 2012, EOS datasheets). Fatigue CoV is always >> UTS CoV — fatigue is far more sensitive to localized defects.
Effect of powder reuse on properties (LPBF metals)
| No. of powder reuses | UTS variation | Elongation variation | Porosity variation | Recommended action |
|---|---|---|---|---|
| 0–5 | baseline | baseline | baseline | None — normal use |
| 5–10 | −1 to −3% | −5 to −10% | +0.02–0.05% | Monitor PSD and chemical composition |
| 10–20 | −3 to −8% | −10 to −20% | +0.05–0.15% | Mandatory coupon testing for structural applications |
| > 20 | Unpredictable | Unpredictable | > 0.2% | Replace powder; unacceptable risk |
Parameters to monitor for powder:
- PSD: D10, D50, D90 — deviation > 15% from baseline → sign of degradation
- Satellite content: > 10% → increased gas porosity risk
- Chemical composition (O₂, N₂ especially for Ti): oxygen increase > 0.02% → reduced elongation
- Flowability (Hall flow): > 30 s/50g → risk of non-uniform distribution
Recommended knockdown factors for design
For robust design, apply knockdowns to nominal table values:
| Application | Knockdown on UTS | Knockdown on fatigue limit |
|---|---|---|
| Functional prototype | −5% | −15% |
| Structural (FS ≥ 2.0) | −10% | −20% |
| Fatigue-critical (FS ≥ 1.5) | −10% | −30% |
| Aerospace / biomedical (certified) | Use values from coupons on the same build plate | Use coupon values + B-basis statistics |
B-basis (statistics): value guaranteed at 90% with 95% confidence. For structural aerospace this is the reference value — not the mean. Requires a minimum of 30 samples to calculate.
Notes on Metal Binder Jetting process
- Shrinkage ~20% linear during sintering — always compensate in CAD (not in slicer)
- Post-sinter tolerances ±0.3–0.5mm vs ±0.05–0.1mm for LPBF
- No structural supports during printing (like SLS) → geometric freedom
- Ceramic setters for sintering on cantilevered geometries
- Post-sinter HIP recommended for critical structural applications
- Economically competitive for volumes >30–50 parts compared to LPBF