When gas expands through a well test choke, it cools down. This temperature drop — the Joule-Thomson effect — is one of the most critical phenomena in surface well testing. Get it wrong and you may find your lines plugged with hydrates minutes into a test that took days to prepare.
This guide explains the physics behind the effect, how the cooling is calculated in practice, how it connects to hydrate risk, and what engineers can do about it.
The Joule-Thomson (JT) effect describes the change in temperature of a real gas when it expands through a restriction at constant enthalpy — that is, with no heat exchange with the surroundings and no work done. In a well test choke, the wellstream passes from high wellhead pressure to a lower downstream pressure across a small orifice. That pressure drop drives the cooling.
The magnitude of the cooling depends on the Joule-Thomson coefficient (µJT), defined as:
For most natural gas compositions at typical well test conditions (pressures below 5,000 psia, temperatures between 100°F and 300°F), µJT is positive — meaning the gas cools as pressure drops. Heavier components (C3+, CO2) generally increase the coefficient; lean dry methane has a lower coefficient.
The JT coefficient is not constant. It varies with pressure, temperature, and fluid composition. Using a fixed average value across the full pressure drop introduces significant error — especially at high wellhead pressures where the coefficient changes most rapidly.
The cooling across a choke can range from a few degrees Fahrenheit for a modest pressure drop on a lean gas to over 60°F for a rich condensate or a large pressure differential. The relationship is approximately:
But because µJT changes along the expansion path, rigorous calculation requires integrating the isenthalpic flash across the pressure step — which is exactly what a physics-based simulator does. The table below shows the cooling calculated for four illustrative gas scenarios:
| Fluid | Wellhead P (psia) | Downstream P (psia) | ΔP (psia) | ΔT at 150°F upstream (100–200°F) |
|---|---|---|---|---|
| Lean dry gas (C1 93%) | 1,000 | 300 | 700 | −34°F (−28 to −42) |
| Rich gas / condensate (C1 72%, C7+ 3%) | 2,000 | 300 | 1,700 | −67°F (−65 to −69) |
| Rich gas, high pressure | 3,000 | 500 | 2,500 | −70°F (−68 to −71) |
| Gas with 15% CO2 | 2,000 | 400 | 1,600 | −83°F (−68 to −102) |
These are model results, not field measurements: they come from Olimetra's rigorous isenthalpic flash (Peng-Robinson) for the illustrative compositions shown, with the range in parentheses covering upstream temperatures of 100 to 200°F. Oil-rich streams (GOR below roughly 2,000 scf/bbl) are not tabulated: their cooling depends on the gas that evolves and the model's uncertainty is larger there. The actual cooling for a specific well depends on the complete fluid composition, the starting temperature, and the exact pressure profile across the choke.
Hydrates are ice-like crystalline structures that form when water combines with light hydrocarbons (C1–C4) or CO2 at elevated pressures and low temperatures. The choke outlet — the coldest point in the surface flow string — is precisely where hydrate conditions are most likely to be met.
The critical check is simple in concept: if the outlet temperature after JT cooling falls below the hydrate formation temperature at the downstream pressure, hydrate risk exists. In practice the calculation is less simple because:
Using the gravity of the whole wellstream instead of the gravity of the gas. The Katz, Towler–Mokhatab and Motiee correlations are functions of gas gravity, not of composition. For a gas condensate, the wellstream's molecular weight is dominated by the heavy end and overstates the gas gravity. In a demonstration well, using the wellstream gravity (0.83) instead of the gas gravity (0.70) moved the hydrate temperature at 620 psia from 57.9°F to 62.9°F. The error here is conservative, but it hides the real uncertainty. These correlations are also valid only for sweet gas of gravity 0.55 to 1.0; H2S and CO2 raise the hydrate temperature, and outside that range a single number should not be trusted.
Not all expansion cools the gas. The Joule-Thomson coefficient falls as pressure rises and changes sign at very high pressure, where the gas heats as it expands. In Olimetra's Peng-Robinson model, a lean gas (93% methane) expanding from 10,000 to 8,000 psia warms slightly — about +0.3 to +2°F for upstream temperatures between 150 and 450°F — while the same gas expanding from 10,000 to 5,000 psia already cools by about 6 to 9°F. This means:
A rigorous isenthalpic flash using the Peng-Robinson EOS captures this behavior, including the sign reversal. Rule-of-thumb approximations do not.
The choke outlet temperature should be part of the pre-job engineering, not a surprise in the field. With wellhead pressure, temperature, fluid composition, and expected flow rate as inputs, the JT cooling and hydrate margin can be calculated in advance.
If hydrate risk exists, continuous injection of thermodynamic inhibitor (methanol or MEG) upstream of the choke is the standard mitigation. The required concentration depends on the temperature depression needed — which itself depends on the JT cooling, the hydrate formation temperature, and the desired safety margin. Under-sizing the injection rate means the inhibitor concentration at the choke outlet may not be sufficient.
The choke outlet temperature is the inlet temperature for the downstream separator and heater (if any). If the JT cooling is larger than expected, the separator may receive fluid at a lower temperature than designed, affecting gas-liquid equilibrium and liquid carryover. For tests with heaters, the duty must be sufficient to bring the fluid to the required separator inlet temperature.
In the field, the temperature immediately downstream of the choke body is a real-time indicator of JT cooling. If the measured temperature is significantly colder than calculated, it may indicate a higher GOR than expected, a different fluid composition, or a larger pressure drop than planned. Significant deviation from the pre-job calculation is a signal to re-evaluate hydrate risk in real time.
A frozen choke body — visible as ice or frost on the external surface — is a direct visual indicator of severe JT cooling. By the time ice forms externally, the internal temperature has been well below freezing for some time and hydrate formation inside the line has likely already begun.
Olimetra WTS calculates the choke outlet temperature using a rigorous isenthalpic flash based on the Peng-Robinson equation of state. The calculation integrates the enthalpy change across the full pressure drop rather than applying a fixed JT coefficient, which captures the change of sign near the inversion point that a fixed coefficient cannot.
The choke temperature model is checked against six reference-simulation cases (two rich, sour fluids), with a mean absolute deviation of about 0.8°F on the temperature change across the choke; the largest deviation, about 3.5°F, is in an ultra-high-pressure case. The hydrate formation temperature is the higher of two published gas-gravity correlations (a deliberately conservative envelope) (Towler–Mokhatab 2005 and Motiee 1991), evaluated with the gravity of the gas phase in equilibrium at the hydrate temperature rather than that of the whole stream, with a correction for H2S and CO2. Where the gas gravity falls outside 0.55–1.0, or there is no gas phase, Olimetra reports "not evaluated" instead of a number; expect an uncertainty of several degrees Fahrenheit either way. Olimetra then compares the choke outlet temperature against the hydrate curve at the downstream pressure — flagging risk and computing the required inhibitor injection rate in the same run.
The result is a single calculation that covers JT cooling, hydrate risk assessment, and inhibitor sizing simultaneously — with no need to run separate tools or transfer numbers between spreadsheets.
Olimetra WTS runs the full surface well test network — choke, separators, PSV, flare and more — in a single physics-based simulation. Checked against reference simulation cases.
See plans and pricing →