Convective Heat Transfer Correlations

Convective Heat Transfer Correlations

Overview

Pure-Python implementations of the classical pipe/duct flow correlations plus a thin wrapper around ht.boiling_nucleic.Rohsenow:

• Dittus-Boelter: simplest turbulent pipe correlation, broad use.
• Gnielinski: turbulent / transitional pipe, wider Pr range, more

accurate than Dittus-Boelter for moderate-Pr fluids.

• Sieder-Tate: turbulent pipe with viscosity correction (good for oils,

high-viscosity service).

• Laminar pipe (constant Tw or constant q).
• Churchill-Chu: natural convection from a vertical plate (laminar +

turbulent unified).

• Rohsenow pool boiling: screening estimate via ht.

Prerequisites

1. uv available.
2. On first use, the script writes LICENSE_NOTIFICATION.txt.

When to Use

• Estimating h for one side of a heat exchanger before passing to

heat-exchanger-sizing.

• Sizing a steam tracing or insulation problem with natural convection.
• Screening pool boiling heat transfer for a kettle reboiler.

Don't use for

• Full shell-and-tube exchanger design — use heat-exchanger-sizing.
• Plate exchangers — vendor-specific correlations apply.
• Falling-film, agitated vessel, or specialty geometries — those have their

own correlations and are not in this skill.

Utility Scripts

• uv run scripts/h.py dittus-boelter --Re 50000 --Pr 4.5 --k 0.6 --Dh 0.025 --output /tmp/db.json
• uv run scripts/h.py gnielinski --Re 50000 --Pr 4.5 --k 0.6 --Dh 0.025 --output /tmp/gn.json
• uv run scripts/h.py sieder-tate --Re 50000 --Pr 4.5 --k 0.6 --Dh 0.025 --mu-bulk 0.001 --mu-wall 0.0007 --output /tmp/st.json
• uv run scripts/h.py laminar --regime constant-T --k 0.6 --Dh 0.025 --output /tmp/lam.json
• uv run scripts/h.py natural --Ra 1e9 --Pr 0.7 --k 0.026 --L 1.0 --output /tmp/nc.json
• uv run scripts/h.py boiling --T-sat 373.15 --T-wall 383.15 --P 101325 --k-l 0.68 --rho-l 958 --rho-g 0.6 --sigma 0.059 --cpl 4217 --dHvap 2257000 --mu-l 0.00028 --output /tmp/boil.json

Procedure

1. Decide the flow regime: Re, Pr, geometry, heated/cooled.
2. Pick the matching correlation:

• Pipe, turbulent (Re > 10000), simple fluid: Dittus-Boelter.
• Pipe, transitional or wide Pr (oils, gases): Gnielinski.
• Pipe, large viscosity gradient at the wall: Sieder-Tate.
• Pipe, laminar (Re < 2300): laminar formulas with the proper BC.
• Surface, natural convection: Churchill-Chu (geometry-dependent).
• Pool boiling: Rohsenow as screening; site-specific surface factors

dominate.

1. Compute h = Nu * k / Dh.
2. Cross-check with a second correlation; flag the spread.

Pitfalls

• Using Dittus-Boelter for Re = 5000. Outside its validity range.
• Using Dittus-Boelter for Pr = 200. Outside its validity range.
• Forgetting that Pr changes with temperature; use a properly averaged

film temperature.

• Treating the "heating" exponent (0.4) as universal; use 0.3 for cooling.
• Using a single h for the entire tube length when entrance effects are

significant (short tubes, low Re); Nu varies along the entrance length.

• Ignoring the viscosity correction (mu_bulk / mu_wall)^0.14 for oils.
• Using Rohsenow with the wrong Csf surface factor (it varies by 2x for

real surfaces); the ht wrapper uses the canonical Csf for water/copper.

Fallback Strategies

• If ht is not installed, the single-phase correlations are pure Python

and work without it.

• For specialty boiling (forced convective, flow film boiling), this skill

has only the screening Rohsenow form; surface a request to the user for a different correlation.

Verification

• Run the listed script with representative inputs and an --output file when a deterministic calculation is available.
• Confirm the JSON result contains ok: true, expected units, and no unhandled warnings.
• Check result magnitudes against the stated assumptions, references, and a hand calculation or known operating range before reporting them.

References

• references/correlation_summary.md — equation forms and ranges.
• Incropera & DeWitt, Fundamentals of Heat and Mass Transfer.
• Kakac, Liu, Heat Exchangers Selection, Rating, and Thermal Design.
• ht documentation: https://ht.readthedocs.io/

Anti-Patterns

• Picking a correlation by name rather than by validity range.
• Reporting h without stating the correlation used.
• Computing tube-side h only and ignoring the shell-side (it usually

dominates).

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: jskherman
• Source: jskherman/engg-skills
• License: Apache-2.0

Install and usage instructions live in the source repository linked above.