Gas Well Liquid Loading: Automate the Turner Check

Gas well liquid loading Turner check automated with Python

By Saad Iqbal

The pumper calls in at 6 a.m.: casing pressure is climbing, tubing pressure is falling, and the well that used to flow steady at 550 Mcf/d is now surging and dying every few hours. Nobody wants to say it out loud yet, but everyone on the call already knows — this gas well is loading up with liquid. The question is never “is it loading,” it’s “how far below critical are we, and what do we do about it before the well dies completely.” That question has a textbook answer from 1969 that still holds up, and by the end of this tutorial you’ll have it running as a five-line Python check you can drop into any well-review spreadsheet.

Gas well liquid loading happens when the gas velocity in the tubing drops below the speed needed to carry produced liquids to surface. Below that critical velocity, liquid droplets fall back down the tubing instead of being lifted out, accumulate at bottom, and add hydrostatic backpressure against the formation — which chokes gas rate further, which drops velocity further, until the well loads up completely. Turner, Hubbard and Dukler published the still-standard critical velocity correlation for this in 1969, and it remains the first calculation most production engineers reach for when a gas well starts acting up.

Prerequisites

  • Python 3.8 or newer installed locally, or a notebook environment (python.org for a native install, or Jupyter if you prefer notebooks)
  • No external packages required for the core calculation — it uses only Python’s built-in math module
  • Matplotlib installed only if you want to reproduce the sensitivity chart later (pip install matplotlib)
  • Current wellhead flowing pressure and temperature, tubing ID, gas specific gravity, and an estimate of produced liquid gravity and gas-liquid surface tension
  • Basic familiarity with gas well nodal behavior helps but isn’t required — every term below is defined as it’s used

Step 1: Understand what “loading up” actually looks like downhole

Before touching a calculator, picture the two states side by side. In a healthy, unloaded gas well, gas velocity through the tubing is high enough that liquid droplets are entrained in the gas stream and carried continuously to surface — the well produces a steady gas rate with only a trickle of associated liquid. Once velocity drops below critical, droplets start falling faster than they’re lifted. Liquid collects as a column at the bottom of the tubing or above a restriction, and that column’s hydrostatic head adds directly to the flowing bottomhole pressure the reservoir has to overcome. The well doesn’t fail all at once — it shows up first as increasing casing-tubing pressure differential, then intermittent slugging, then heading, and finally a well that dies unless it’s unloaded mechanically (velocity strings, plunger lift, or a swab job).

Why it matters: catching the trend early — while the well is still flowing but trending toward critical — is the difference between planning a velocity string or plunger installation on your schedule versus an emergency workover after the well dies. The Turner check gives you a number to track against, not just a gut feeling from the pressure chart.

Gas well liquid loading comparison showing lifted droplets and liquid fallback

Step 2: Gather the four inputs the Turner equation actually needs

The Turner critical velocity correlation needs surprisingly few inputs, and all four should already be on your well test sheet or in your production database:

  • Flowing wellhead pressure (P), psia — from your last well test or continuous pressure gauge
  • Flowing wellhead temperature (T), °F — same source
  • Tubing inside diameter (d), inches — from your tubing tally, not the nominal OD size
  • Gas specific gravity (γg), air = 1.0 — from your gas composition analysis
  • Produced liquid gravity (γL) and gas-liquid interfacial tension (σ), dynes/cm — condensate typically runs 20 dynes/cm, produced water closer to 60-70 dynes/cm; use the lighter, more conservative liquid if you’re producing both

Why it matters: using nominal tubing OD instead of actual ID is the single most common input error in this calculation — a 2-3/8 inch nominal tubing string can have an actual drift ID anywhere from 1.87 to 1.995 inches depending on weight, and since critical rate scales with the tubing’s cross-sectional area, getting this wrong shifts your answer by 10-15% in either direction.

Step 3: The governing equation

Turner’s correlation gives the minimum gas velocity needed to keep the largest stable liquid droplet entrained, derived from a droplet drag-versus-gravity force balance:

v_g = 1.593 × σ^0.25 × (ρ_L − ρ_g)^0.25 / ρ_g^0.5

where v_g is critical gas velocity in ft/s, σ is gas-liquid surface tension in dynes/cm, and ρ_L and ρ_g are liquid and gas density in lb/ft³. Converting that velocity into a critical gas rate for your specific tubing size uses continuity:

q_c = 3.067 × P × v_g × A / (T × z)

where q_c is critical gas rate in MMscf/d, P is flowing pressure in psia, A is tubing cross-sectional area in ft², T is flowing temperature in °R (°F + 460), and z is the gas compressibility factor at flowing conditions. Gas density itself comes from the real gas law, ρ_g = (P × M) / (z × R × T), where M is the gas’s apparent molecular weight (28.97 × γg) and R is 10.732 (the gas constant in field units). The well is loading up whenever actual gas rate falls below q_c.

Gas well liquid loading Turner check worked example and critical rate

Step 4: Work a real example by hand first

Take a gas well flowing at wellhead pressure P = 800 psia, temperature T = 140°F, through 2-3/8 inch tubing with an actual ID of 2.441 in, producing 0.65 specific gravity gas with condensate (γL = 0.78, σ = 20 dynes/cm) at an actual measured rate of 0.55 MMscf/d.

First, estimate z at flowing conditions using Standing’s pseudo-critical correlations for a low-CO2/H2S gas: pseudo-critical temperature Tpc = 168 + 325(0.65) − 12.5(0.65)² ≈ 373°R, pseudo-critical pressure Ppc = 677 + 15(0.65) − 37.5(0.65)² ≈ 671 psia. That gives reduced properties Tpr = 600/373 ≈ 1.61 and Ppr = 800/671 ≈ 1.19, which the Papay correlation resolves to z ≈ 0.87. Gas density then works out to ρ_g = (800 × 28.97 × 0.65)/(0.87 × 10.732 × 600) ≈ 2.69 lb/ft³, and liquid density is 62.4 × 0.78 ≈ 48.7 lb/ft³.

Plugging into Turner’s equation: v_g = 1.593 × 20^0.25 × (48.7 − 2.69)^0.25 / 2.69^0.5 ≈ 5.47 ft/s. Tubing cross-sectional area for a 2.441 in ID is A = (π/4)(2.441/12)² ≈ 0.0325 ft². Critical rate is then q_c = 3.067 × 800 × 5.47 × 0.0325 / (600 × 0.87) ≈ 0.80 MMscf/d. Since the well’s actual rate of 0.55 MMscf/d is below that 0.80 MMscf/d threshold, this well is confirmed to be loading up — consistent with exactly the symptoms the pumper described on the morning call.

Step 5: Automate it in Python so you can run it in seconds

Hand-calculating z-factor and critical rate for one well is fine once. Running it across a 40-well liquid-loading watchlist every week is not something you want to do by hand. Here’s the full check as a Python function, using the Papay direct correlation for z so there’s no iteration required:

import math

def z_factor(P, T, gamma_g):
    """Papay (1968) direct z-factor correlation using Standing's
    pseudo-critical property correlations. P in psia, T in Rankine."""
    Tpc = 168 + 325 * gamma_g - 12.5 * gamma_g**2
    Ppc = 677 + 15 * gamma_g - 37.5 * gamma_g**2
    Tpr, Ppr = T / Tpc, P / Ppc
    return (1 - (3.52 * Ppr) / 10**(0.9813 * Tpr)
              + (0.274 * Ppr**2) / 10**(0.8157 * Tpr))

def turner_check(P, T_F, d_in, gamma_g, gamma_L, sigma, q_actual_mmscfd):
    """Returns (critical velocity ft/s, critical rate MMscf/d, is_loading)."""
    T = T_F + 460                      # Rankine
    A = math.pi / 4 * (d_in / 12) ** 2  # tubing area, ft^2
    z = z_factor(P, T, gamma_g)
    rho_g = (P * 28.97 * gamma_g) / (z * 10.732 * T)   # lb/ft^3
    rho_L = 62.4 * gamma_L                              # lb/ft^3
    v_g = 1.593 * sigma**0.25 * (rho_L - rho_g)**0.25 / rho_g**0.5
    q_c = 3.067 * P * v_g * A / (T * z)
    return v_g, q_c, q_actual_mmscfd < q_c

v_g, q_c, loading = turner_check(
    P=800, T_F=140, d_in=2.441, gamma_g=0.65,
    gamma_L=0.78, sigma=20, q_actual_mmscfd=0.55
)
print(f"Critical velocity: {v_g:.2f} ft/s")
print(f"Critical rate:     {q_c:.3f} MMscf/d")
print(f"Loading up:        {loading}")

# >> Critical velocity: 5.47 ft/s
# >> Critical rate:     0.801 MMscf/d
# >> Loading up:        True

Notice the function returns the exact same 5.47 ft/s and 0.80 MMscf/d we worked by hand above — that match is your first sanity check that the code is right. From here it takes one more loop to run this against every well in a CSV export from your production database and flag anything where q_actual < q_c, turning a manual well review into an automated watchlist that refreshes every time new test data lands.

It’s also worth seeing how sensitive that critical rate is to wellhead pressure, since pressure is usually what’s declining as a well ages toward loading:

Gas well liquid loading Python code and verified Turner check output

As flowing pressure drops with reservoir depletion, gas density at wellhead drops with it — and since critical velocity scales inversely with the square root of gas density, the well needs a higher and higher actual rate just to stay above critical. This is exactly why gas wells that flowed comfortably for years start loading up as they mature, even with no change in tubing size or fluid properties.

If you’re also running nodal analysis in Python on the same well, the Turner critical rate is a natural second curve to overlay on your VLP plot — anywhere your operating point sits left of critical on the x-axis is a loading well by definition. And once a well is confirmed loading, the typical next step is sizing artificial lift; our ESP affinity laws calculator and plunger or velocity-string sizing follow directly from the critical rate you just calculated here.

Expected result and how to verify it

Running the function above on the worked example should return a critical velocity of 5.47 ft/s and a critical rate of 0.80 MMscf/d, flagging the well as loading since its actual 0.55 MMscf/d sits below that threshold. As a sanity check, critical velocity for typical gas well conditions (500-2,000 psia, condensate present) almost always falls between 3 and 8 ft/s — if your function returns something wildly outside that range, check your pressure units (psia, not psig) and temperature units (Rankine, not Fahrenheit) first, since those are the two most common sources of an order-of-magnitude error.

Common pitfalls

  1. Using tubing OD instead of actual drift ID. Pull the actual ID from your tubing tally or the manufacturer’s drift table — nominal size alone can be off by 5-15% depending on weight and grade.
  2. Forgetting the psia/psig and °F/°R conversions. The Turner equation and the z-factor correlation both require absolute pressure and absolute temperature — plugging in gauge pressure or Fahrenheit directly silently produces a plausible-looking but wrong answer.
  3. Using produced water’s surface tension when condensate is also present. Condensate’s lower surface tension (≈20 dynes/cm versus ≈60-70 for water) gives a lower, more conservative critical velocity — using water’s higher tension when both fluids are present can make a loading well look falsely healthy.

Gas Well Liquid Loading Field Checklist

  • Recalculate gas well liquid loading after every stable well test.
  • Trend gas well liquid loading margin against tubing pressure and rate.
  • Investigate gas well liquid loading when casing pressure rises as tubing pressure falls.
  • Validate gas well liquid loading inputs at flowing conditions.
  • Use gas well liquid loading screening before selecting deliquification.
  • Review gas well liquid loading after choke or compression changes.
  • Document gas well liquid loading uncertainty alongside the Turner result.

The Turner check won’t tell you which artificial lift method to install or exactly when the well will die — but it turns “the well feels like it’s loading” into a specific number you can track test over test, which is usually all the lead time you need to plan the fix on your terms instead of the well’s.

Saad Iqbal Avatar

About the author

Saad Iqbal

Petroleum Engineer · Well Intervention & Stimulation Specialist

Saad Iqbal is a petroleum engineer and well intervention and stimulation specialist with more than a decade of field experience in hydraulic fracturing, coiled tubing, CSG, tight sandstone and shale developments. He explores practical AI, automation and data-driven engineering for safer, smarter upstream operations.