Knowledge Base
Choked flow (also called critical or sonic flow) happens when gas velocity reaches the speed of sound at a pipe restriction. Once you choke, decreasing downstream pressure does not increase flow. That sounds theoretical until you find your PSV is not passing what you sized it for, or your flare header throttles at the T-joint you added during a revamp.
Quick rule of thumb: if P2/P1 < 0.528 for air or diatomic gas (gamma = 1.4), you are choked. For natural gas (gamma ~= 1.30), the critical ratio is ~0.546. For steam (gamma ~= 1.30), same.
For any gas flow design review, I check three numbers before anything else:
| Check | What it tells you | Limits |
|---|---|---|
| Mach at pipe outlet | Is the main line approaching choking? | Target M < 0.3 design; M < 0.5 short-term; M < 0.7 absolute max |
| Mach at any restriction (reducer, valve, tee) | Localized choking that degrades downstream performance | Same — M < 0.3 preferred |
| Critical pressure ratio at each dP element | Will this element choke and decouple flow from downstream pressure? | If P2/P1 < (2/(gamma+1))^(gamma/(gamma-1)), choked |
The critical ratio r_c = (2/(gamma+1))^(gamma/(gamma-1)) is a function of gamma only:
| Gas | gamma (typical) | Critical P2/P1 |
|---|---|---|
| Monatomic (He, Ar) | 1.67 | 0.487 |
| Diatomic (H2, N2, O2, air) | 1.40 | 0.528 |
| Natural gas (CH4-rich) | 1.30 | 0.546 |
| Propane, butane | 1.13 | 0.576 |
| Steam (saturated) | 1.30 | 0.546 |
| Steam (superheated high-T) | 1.33 | 0.540 |
If the pressure drop across the line is more than ~10% of inlet absolute pressure, compressibility matters. Use Crane TP-410 isothermal Eq. 1-6 or adiabatic Eq. 1-8. The Weymouth / Panhandle equations are shortcuts for long pipeline networks.
Gas expansion in a pipe (especially across a restriction) is closer to isothermal over long distances but locally adiabatic. Temperature drops reduce sonic velocity: a = sqrt(gamma · R · T / M). A sonic line that looked OK at 100 degF can choke at -40 degF downstream of a JT valve.
Crane TP-410 has tabulated K and L/D values. A single gate valve fully open is roughly L/D = 8, a ball valve is 3, a 90-deg elbow standard radius is 30. Add them up before computing friction.
API 520 Part II recommends PSV outlet piping be sized so built-up backpressure < 10% of set for conventional, < 50% for balanced-bellows. Exceed that and the valve chatters or fails to reclose — worse than losing capacity.
For natural gas (MW 18, gamma 1.30) at 100 degF:
a = sqrt(1.30 * 32.174 * 1545.3 * 560 / 18)
a ~= 1430 ft/s
Design for v < 400 ft/s (Mach 0.28) to stay in the safe subsonic region.
| Situation | Design Target | Why |
|---|---|---|
| Process piping, continuous service | Mach <= 0.3 | Erosion, noise |
| Flare header (main) | Mach <= 0.5 | API 521 §7.3.3 |
| Flare sub-header laterals | Mach <= 0.6 | Short duration only |
| Flare tip exit | Mach <= 0.2 | Flame stability per API 537 |
| PSV body (inside valve) | Sonic by design | Required for standard API 520 sizing |
| PSV outlet tailpipe | Mach <= 0.7 at peak | Prevents standing wave / backpressure spikes |
| Compressor suction | Mach <= 0.1 (50 ft/s typ) | NPSH-like requirement for centrifugals |
| Instrument air supply piping | 100 ft/s max | Noise and rider-valve issues |
© 2026 Inflection Point Engineering, LLC. All rights reserved. The content of this page — including calculation methods, reference data, written analysis, interactive tools, and source code — is the intellectual property of Inflection Point Engineering, LLC and is protected under applicable copyright, trademark, and trade secret laws. Unauthorized reproduction, redistribution, modification, or derivative use in whole or in part is prohibited without prior written consent.
Disclaimer. This material is provided for informational and educational purposes only and does not constitute professional engineering advice. Calculations, reference data, and methodologies are based on published standards and accepted engineering practice but are not a substitute for engineering judgment, site-specific analysis, or review by a licensed Professional Engineer. Inflection Point Engineering, LLC makes no warranties, express or implied, regarding the accuracy, completeness, or fitness for a particular purpose of any content presented here, and shall not be liable for any direct, indirect, incidental, or consequential damages arising from its use. Users assume all risk associated with applying this content to real-world design, operations, or decisions.
© 2026 Inflection Point Engineering, LLC. All rights reserved.