understanding gas flow rate conversion

• How do I understand SCFM? What does it mean physically?

Standard Cubic Feet per Minute (SCFM) is physically a measurement of mass flow normalized at a 'standard' condition.  SCFM tells you how much gas (in mass) is moving per minute, expressed as the volume that mass would occupy at standard conditions.

However, note that SCFM is not literally a mass flow unit – it’s still a volume (cubic feet per minute) but at imaginary standard conditions. So you must specify what “standard” means every time. Different industries use different standard reference points.   There is this hysterical table on wikipedia  shown below that shows many of the current possible options. Always ask what temp/pressure/humidity the SCFM value is based on.  Humidity would only apply to air likely b/c if you are considering other gases they would be pressurized dry sources.

Most common for industrial piping application is Temperature: 60 °F (15.6 °C), Pressure: 14.7 psia (1 atm), Humidity: Usually 0% RH (dry gas basis) but always check.

• How to interpret SCFM for real components/systems?   

If you’ve got a flow curve in SCFM vs ΔP for air, N₂ , other fluid (compressor, regulator, valve, filter) and you want to estimate the curve for another gas (argon, helium, etc.), you can do it—but the accuracy depends heavily on what type of component it is and what regime you’re in (subcritical vs choked, laminar vs turbulent, real-gas effects, etc.).

Below is the engineer’s “least-wrong” playbook.

The big idea

Most vendor “SCFM vs ΔP” curves are really representing:

• an effective flow coefficient of the device (Cv/Cg/Kv or an internal equivalent),
  
  

• plus assumptions about gas properties (MW, k, Z),
  
  

• and sometimes hidden assumptions about upstream pressure, temperature, and standard conditions.
  
  

So the right way is usually:

1. Convert the curve to a gas-agnostic coefficient (Cv/Cg or equivalent), then
   
   

2. Recompute flow for the new gas using compressible gas equations.
   
   

When you can’t do that, you use scaling laws as an approximation.

Case A — Valves / Regulators / Filters (ΔP devices)

Best method (recommended): back-calculate an equivalent Cv/Cg curve

If the curve provides inlet pressure P1, outlet pressure P2, temperature, and gas (air/N₂), you can:

1. For each point on the curve, compute an “effective Cv” (or Cg) using the standard compressible valve equations.
   
   

2. Assume the same Cv vs valve position / lift / opening applies to other gases.
   
   

3. Recalculate Q for the new gas at the same P1, P2, T.
   
   

Pros: physically grounded; handles choked/subcritical properly
Cons: you need enough info to solve the valve equation

If you only have SCFM vs ΔP but no P1, you’re missing a major state variable and will be guessing.

Quick-and-dirty scaling when you only have SCFM vs ΔP

You’ll see two common regimes:

1) Subcritical compressible flow (not choked)

A decent first-order scaling is:

Q∝1GQ \propto \frac{1}{\sqrt{G}}Q∝G​1​

where GGG is specific gravity relative to air (≈ MW / 28.97 for ideal gases at same T).

So to go from gas A to gas B at the same P1, ΔP, T:

QB≈QAGAGB=QAMWAMWBQ_B \approx Q_A \sqrt{\frac{G_A}{G_B}} = Q_A \sqrt{\frac{MW_A}{MW_B}}QB​≈QA​GB​GA​​​=QA​MWB​MWA​​​

Example (N₂ → Ar)
MW: N₂=28.0, Ar=39.95

QAr≈QN22839.95≈0.84 QN2Q_{Ar} \approx Q_{N2}\sqrt{\frac{28}{39.95}} \approx 0.84\,Q_{N2}QAr​≈QN2​39.9528​​≈0.84QN2​

So an Ar curve will sit ~16% lower in SCFM than N₂ at the same ΔP (roughly).

2) Choked (sonic) flow through restrictions

Choked gas mass flow scales similarly:

m˙∝kMW⋅P1T\dot m \propto \sqrt{\frac{k}{MW}} \cdot \frac{P_1}{\sqrt{T}}m˙∝MWk​​⋅T​P1​​

Volumetric at standard conditions:

Qstd∝m˙ρstd∝m˙⋅1MWQ_{std} \propto \frac{\dot m}{\rho_{std}} \propto \dot m \cdot \frac{1}{MW}Qstd​∝ρstd​m˙​∝m˙⋅MW1​

So in choked flow, MW and k both matter, and SCFM scaling can deviate from the simple √MW rule.

Practical takeaway:

• Helium (low MW, high k) can blow your estimate up.
  
  

• CO₂ (MW 44, k ~1.3, real-gas Z) can be way off near high pressure.
  
  

Limitations / inaccuracies for valves/regulators/filters

Here’s what bites you:

1. Unknown P1/P2 (absolute pressures)
   ΔP alone is not enough for compressible gas. Same ΔP at 30 psia vs 300 psia is totally different flow.
   
   

2. Choking threshold differs by k
   Critical pressure ratio depends on k. Different gas ⇒ choke onset shifts.
   
   

3. Real-gas Z
   At elevated pressure (common in cover gas and bottle-fed systems), Z≠1, especially for CO₂, but also noticeable for N₂/Ar at high pressure. That changes density and flow.
   
   

4. Temperature changes (JT + expansion)
   Across regulators especially: gas cools during throttling. T is not constant unless you’re lucky or heat transfer is strong.
   
   

5. Viscosity matters for filters / laminar regions
   Filters often have mixed regimes:
   
   

• At low flow: ΔP ~ μQ (viscosity dominates)
  
  

• At higher flow: ΔP ~ ρQ² (inertial dominates)
  
  

Different gas changes μ and ρ differently, so a single scaling factor won’t fit the whole curve.

6. Acoustics / noise trim / multi-stage regulators
   Some designs behave very nonlinearly with different gases.
   
   

Case B — Compressors (especially displacement compressors)

Compressors are not just ΔP devices. Their curves are messy because the machine physics matters.

Two different “ratings” you might have:

1) Inlet volumetric capacity (ACFM at suction)

For a positive displacement compressor, the swept volume per rev is basically fixed, so:

• ACFM at suction is roughly fixed (minus slip / re-expansion)
  
  

• SCFM changes with inlet density
  
  

If you keep suction P,T the same and swap gases, the ACFM is similar, but mass flow changes with MW and Z.

2) SCFM delivery rating (converted to standard conditions)

If the vendor states “SCFM” for air, they’ve already implied standard density for air.

To estimate for another gas, it depends on what limit dominates:

• power limit (horsepower)
  
  

• temperature limit
  
  

• speed limit
  
  

• machinery valve losses
  
  

• motor amps
  
  

A common first-pass assumption:

• compressor is power-limited, and efficiency doesn’t change too much
  
  

Then:

Power∼m˙⋅Δh\text{Power} \sim \dot m \cdot \Delta hPower∼m˙⋅Δh

For ideal gas compression:

Δh∝kk−1RT[(P2P1)(k−1)/k−1]\Delta h \propto \frac{k}{k-1}RT \left[\left(\frac{P_2}{P_1}\right)^{(k-1)/k}-1\right]Δh∝k−1k​RT[(P1​P2​​)(k−1)/k−1]

So different gas ⇒ different k and R ⇒ different power per unit mass.

Meaning:

• Helium often looks “easy” volumetrically but can be “weird” due to leakage/clearance and high k.
  
  

• Argon vs nitrogen is usually a modest shift.
  
  

In practice: compressor curves translated between gases without vendor support can be off by ±10–30%, sometimes worse.

Compressor conversion limitations

• Slip/leakage depends on gas viscosity and density
  
  

• Valve dynamics (reed valves etc.) change with gas speed of sound, density
  
  

• Cooling and discharge temperature depend on k
  
  

• Choked flow in internal ports can appear with lighter gases
  
  

If you need tight accuracy, you want vendor performance for that gas, or you model the compressor.

Recommended workflow (what I’d do on an argon cover gas system)

1. Get/assume operating envelope: P1, P2 (or ΔP + P1), T, and expected flow range.
   
   

2. For valves/regulators/filters:
   
   

   • Convert vendor air/N₂ curve into Cv (or equivalent) point-by-point.
     
     

   • Recompute curve for argon using compressible equations with k, MW, Z.
     
     

   • Apply a correction band: ±10% for clean valve trim, ±20% for filters/regulators unless validated.
     
     

3. For compressors:
   
   

   • Identify whether the curve is suction ACFM, discharge SCFM, or “free air delivery”.
     
     

   • If you can, convert to mass flow vs pressure ratio and then re-map with the new gas properties.
     
     

   • Otherwise, treat it as order-of-magnitude and seek vendor confirmation.
     
     

Rule-of-thumb error bars

These are blunt but realistic:

• Simple orifice/valve trim, subcritical, same P1/T: ±5–15%
  
  

• Regulator performance curve (with cooling, droop, choked transitions): ±10–25%
  
  

• Filters (unknown regime mix): ±15–35%
  
  

• Compressors (different gas, same machine): ±10–30% (can be worse for He/H₂/CO₂)
  
  

• Air compressors

  • For Positive Displacement (PD) compressors (screw, piston), the displacement is volumetric. If the machine is sealed and safe for the gas, a 100 ACFM air compressor will move roughly 100 ACFM of Nitrogen. However, leak rates (slip) vary with viscosity and density.

  • A particularly dangerous confusion occurs when converting SCFM for centrifugal compressors (dynamic machines). Unlike PD compressors, which physically trap and squeeze gas, centrifugal compressors accelerate gas to generate pressure. The energy imparted to the gas is measured in Head (feet or meters), which depends on the impeller tip speed.

• Regulators/valves/filter