Thermal Models & Other Heat Leak Paths
Table of Contents
- Lockheed Equation — Scope & Limitations
- Alternative MLI Thermal Conductivity Models
- Other Heat Leak Paths (Beyond MLI)
- Engineering Manufacturing Practices
- Why Material Testing Matters — And How We Can Help
1. Lockheed Equation — Scope & Limitations
The Lockheed equation is the most widely used semi-empirical correlation for estimating the effective thermal conductivity of multilayer insulation (MLI). Its core form is:
q/N = Cs·(T_mean^2.56 / ΔT)^0.76 + Cr·(T_mean^2.63 / ΔT)^0.76 + Cg·(T_mean^0.52 / ΔT)^0.76 × P
where q/N is the heat flux per layer, T_mean is the mean temperature, ΔT is the temperature difference across the blanket, P is the interlayer pressure, and Cs, Cr, Cg are empirically fitted coefficients for solid conduction, radiation, and gas conduction respectively.
What the equation models
The Lockheed equation captures exactly three heat transfer mechanisms within the MLI blanket:
- Radiation — thermal photon exchange between adjacent reflective shields
- Solid conduction — heat conducted through spacer materials and at shield-to-shield contact points
- Residual gas conduction — heat transferred through gas molecules trapped between layers
Key assumptions
- Uniform interlayer pressure throughout the blanket
- Ideal layer contact and spacing
- No penetrations, edges, or boundary effects
- Flat, infinite blanket geometry (no curvature effects)
Validated range
- Vacuum: 10⁻⁵ to 10⁻³ Torr
- Layer density: 20–40 layers/inch
- Boundary temperatures: 20–300 K
What the equation does NOT include
The Lockheed equation is strictly a MLI blanket model. It does not account for:
- Structural supports (G10 rods, Kevlar straps, carbon fiber members)
- Pipe penetrations (fill/drain/vent lines)
- Electrical feedthroughs
- Neck tubes and nozzle connections
- Edge losses at blanket boundaries
- Seam and joint losses
- Fastener thermal bridges
- Compression points where hardware contacts the blanket
The safety factor connection
The 2–5× safety factor commonly applied to Lockheed predictions in real engineering practice exists precisely to account for these unmodeled heat leak paths. It is not a fudge factor for equation inaccuracy — it is a correction for the gap between an idealized flat-blanket model and a real installed cryogenic vessel.
2. Alternative MLI Thermal Conductivity Models
Since the original Lockheed work in the 1970s, multiple research groups have proposed alternative models to improve accuracy, extend the valid range, or provide better physical insight. The table below compares seven notable approaches.
| Model | Year | Approach | Accuracy | Best For | Key Source |
|---|---|---|---|---|---|
| Lockheed Original | 1970s | Semi-empirical (3 terms) | ±30% near room temp, larger error at cryo end | Quick first-order estimates | Lockheed Reports (1970s), ASTM C740 |
| Modified Lockheed (MLE) | NASA MSFC | Adjusted coefficients | Unsatisfactory at low-T end (Hacettepe 2022 validation) | NASA heritage designs | NASA MSFC internal |
| Layer-by-Layer (LBL) Separated Mode | McIntosh et al. | Physical layer-by-layer simulation | Best overall accuracy (validated by Hacettepe 2022) | Detailed thermal design, research | McIntosh & Smith; Hacettepe Univ. thesis (2022) |
| Doenecke Equation | 1980s | Empirical | Similar to LBL, good at cryo end | Cryogenic design | Doenecke (1980s) |
| JPL Cunningham Physical Parameterization | 2015 | Physics-based; empirical coefficients decomposed into physical parameters | Adaptable to different MLI constructions; showed MLI conduction can vary 600× across configurations | Custom MLI configurations | JPL / Cunningham, CEC 2015 |
| NASA IMLI New Empirical (Hamill) | 2024 | Fitted to 37 test sets (IMLI + LBMLI) | Good above 150 K; below 150 K requires 1.5–11× correction factor | IMLI and LBMLI systems | NASA TFAWS 2024, NTRS |
| Improved LBL + ELM Pressure Inversion (Wu & Tan) | 2023 | LBL with actual interlayer pressure predicted by ELM neural network | Error reduced from 89% to 2.77% vs. measured data | High-precision prediction with pressure data | Applied Sciences (MDPI); Xi’an Jiaotong Univ. |
Key findings from model comparisons
- Lockheed wins on simplicity and remains the industry default for first-order estimates, but lacks accuracy at the cryogenic (low-T) end.
- LBL (Layer-by-Layer) is physically the most rigorous approach, modeling each shield-spacer pair individually, but is computationally expensive for system-level analysis.
- JPL/Cunningham model is the most adaptable — by decomposing empirical coefficients into physical parameters (spacer conductivity, contact area fraction, radiation exchange factor), it can be tailored to non-standard MLI constructions. Their key finding: MLI effective thermal conductivity can vary by a factor of 600× across different material combinations and layer densities.
- NASA 2024 (Hamill) model is the newest entry, specifically calibrated for IMLI (Improved MLI) and LBMLI (Layer-By-Layer MLI) configurations, but still requires a 1.5–11× correction factor below 150 K.
- The 2022 Hacettepe University master’s thesis by Toygan Er provides the most comprehensive direct comparison of multiple MLI models against the same experimental dataset — a valuable reference for anyone selecting a thermal model.
References & Further Reading
- NASA TFAWS 2024 (Hamill) — IMLI characterization with 37 test sets: Download PDF
- JPL / Cunningham 2015 — Quantifying MLI thermal conduction via physical parameterization: Download PDF
- Hacettepe University 2022 Thesis — Comprehensive MLI model comparison (Toygan Er): View record
- AIP Layer Density Optimization — ICEC proceedings on layer density effects: Download PDF
- Xi’an Jiaotong Univ. (Wu & Tan 2023) — Improved LBL with ELM pressure inversion: Download PDF
3. Other Heat Leak Paths (Beyond MLI)
The Lockheed equation — and any MLI thermal model — only covers the blanket itself. A real cryogenic vessel has multiple additional heat leak paths that must be calculated independently and added to the MLI load.
| Heat Leak Path | Description | Typical Estimation Method | Engineering Notes |
|---|---|---|---|
| Structural supports (G10 rods, Kevlar straps, carbon fiber) | Conduction through mechanical supports connecting inner and outer vessels | Fourier’s law: Q = k·A·ΔT/L; G10: k ≈ 0.3 W/(m·K) at 77 K | Minimize cross-section, maximize length; G10-CR is the preferred cryogenic grade |
| Pipe penetrations (fill / drain / vent lines) | Conduction along process piping that passes through the vacuum space | 1D fin model or FEA; stainless steel k varies strongly with temperature | Use thin-walled tubing; incorporate bellows loops to absorb thermal contraction |
| Electrical feedthroughs | Conduction through conductor wires powering sensors, heaters, etc. | Sum of k·A·ΔT/L for each conductor | Use manganin or constantan wire (low thermal conductivity); minimize wire gauge |
| Neck tubes | Conduction from the warm end down to the cold vessel nozzle | Integrated k(T) over length: Q = ∫k(T)dT × A/L | Often the dominant heat leak path in small Dewars |
| Edge / seam losses | Heat bypassing MLI at blanket edges and seam lines | Typically 2–5% of the MLI area heat load (rule of thumb) | Stagger seams; use lapped seams, never butt joints |
| Compression points | Local thermal bridges where MLI is compressed by fasteners, clips, or ties | Point contact resistance models | Use low-conductance fasteners (Nylon, PEEK); minimize compression area |
| Multi-layer joints | Heat leak at overlaps between adjacent blanket sections | Typically 10–30% higher than continuous sections | Use minimum overlap of 2–3 inches (50–75 mm); offset joints between layers |
The total heat load equation
Total heat leak = MLI calculated value × safety factor (2–5×) + independently calculated penetration/support loads
This is why a complete thermal analysis of a cryogenic vessel requires far more than just plugging numbers into the Lockheed equation. Every support, every pipe, every wire, and every seam contributes to the total heat load budget.
4. Engineering Manufacturing Practices
Real-world MLI performance depends heavily on installation quality and manufacturing practices. The following factors bridge the gap between lab-measured performance and field performance.
Installation Factor (2–5× Safety Margin)
Laboratory MLI measurements are taken under ideal conditions: flat coupon, uniform layer pressure, no penetrations, no edge effects. Real installations involve:
- Curved surfaces (cylinders, spheres, complex geometries)
- Seams, overlaps, and cutouts for penetrations
- Compression points at fastener locations
- Potential handling damage during installation
Typical safety factor values:
- 2× — well-controlled aerospace installations with documented quality procedures
- 3–5× — industrial cryogenic vessels with standard workshop practices
Important: The safety factor does not replace proper design — it accounts for installation variability only.
Layer Density Zoning
Different areas of a vessel may require different layer densities:
- Flat surfaces: Standard density (20–30 layers/inch, 8–12 layers/cm)
- Curved surfaces: Higher density (30–40 layers/inch, 12–16 layers/cm) to maintain full coverage under compression
- Critical areas: May require localized additional layers
Key practice: Record the actual installed density for each zone — this is what goes into the thermal calculation, not the nominal specification.
Quality Factor (Q-factor)
The Q-factor quantifies the gap between ideal and installed performance:
Q = q_actual / q_ideal
- Industrial cryogenic vessels: Q typically 2–5
- Aerospace applications: Q typically 1.2–2
Track Q-factor across projects to build institutional knowledge and improve future designs. A decreasing Q-factor trend indicates improving manufacturing capability.
Interlayer Pressure & Outgassing
The actual pressure inside the MLI blanket is higher than what the vessel’s vacuum gauge reads, because:
- Materials within the MLI (spacers, adhesives, tags) continue to outgas
- Gas must diffuse through perforations or around edges to reach the vacuum pump
- Local pressure gradients exist across the blanket thickness
Engineering responses:
- Adsorbent placement strategy: Position activated carbon or molecular sieves near the warm boundary, where outgassing originates
- Perforated vs. solid MLI: Perforated layers allow gas to escape more easily but may compromise edge protection and particulate retention
- Hold time consideration: Vacuum degrades over time; design for end-of-life vacuum, not beginning-of-life
Adsorbent Selection
| Adsorbent | Best For | Notes |
|---|---|---|
| Activated carbon | N₂, O₂, H₂O (broad spectrum) | Most common choice for general cryogenic service |
| 5A molecular sieve | Preferential CO₂ and H₂O adsorption | Use when CO₂ is a specific concern |
| Placement | Near warm boundary | Where outgassing originates and desorption capacity is highest |
Critical rule: Adsorbent capacity must cover the entire service life (hold time requirement) of the vessel, not just the initial pump-down.
Acceptance Testing
The customer’s acceptance test method should be the basis for any thermal performance guarantee.
- Calorimetric test: Measure boil-off rate of a known cryogen → calculate total heat leak
- Preferred approach: System-level test (complete vessel) over coupon test, because it captures all heat leak paths including installation effects
- Guarantee basis: The thermal guarantee should reference the same test method and conditions that will be used for acceptance
5. Why Material Testing Matters — And How We Can Help
The Lockheed equation and its variants are estimation tools — they provide useful first-order predictions, but actual thermal performance depends on installed conditions, material batch quality, vacuum integrity, and workmanship. For critical applications, we recommend characterizing thermal performance using project-representative samples under conditions that replicate your operating environment.
Every set of validated test data strengthens the knowledge base of our industry. As more measured results become available, predictive models improve, standards become more rigorous, and the entire cryogenic insulation field advances.
At East Far Cryo Materials, we are committed to close material collaboration throughout this process. Whether you need material samples for testing support, are exploring next-generation insulation solutions through joint R&D, or require industrial-scale production capacity when your design moves from prototype to volume deployment — we are positioned to support you at every stage, from lab bench to factory floor.