Winter Letdown Hydrate Risk: Simulation Says 9 °C, the Plant Says 18 °C
A gas terminal in Hong Kong takes rich gas at 92 barg and 32 °C and lets it down to 28 barg through a single Joule–Thomson valve ahead of the metering skid. The gas runs 3.6 mol% CO2, 1.1 mol% C3+ and 60 ppm H2S, and is water-saturated at the inlet. There is no upstream dehydration. Over three winters the unit tripped five times on hydrate blockage downstream of the valve; each time the valve outlet was measured at 16–19 °C. The steady-state simulation left behind by your predecessor used Peng–Robinson with default binary interaction parameters, a dry lean design gas, and predicted hydrate formation only below 9 °C. Operations were therefore told that hydrate risk was "covered by low ambient conditions". Send-out is now to rise 20%, so the valve will pass more gas and the downstream meter run and strainer will see higher velocities. The only inhibitor on site is a drum of methanol held for emergency use. You must issue a written winter operating basis before the uplift is accepted.
1. Reconcile the simulation with the five winters of field evidence. What is the most likely explanation, and which specific model and data checks would you complete before issuing an operating basis?
Read the worked answer
Key thermodynamic point first. With free water present (the gas is water-saturated and there is no dehydration), the hydrate dissociation temperature at a given pressure is fixed by the gas composition, not by the water content. So a poor water-solubility model is a weak suspect; composition and the actual temperature–pressure path are the strong suspects.
Ranked checks
- Trace heavies. 1.1 mol% C3+ raises the hydrate curve by several degrees relative to the dry lean design gas. Ask for the worst winter chromatograph, not the design basis, and re-run with it.
- CO2. At 3.6 mol% CO2 can dominate hydrate behaviour and forms hydrate at warmer conditions than methane. Check whether the tool used a methane-only correlation or actually handled CO2 hydrate.
- EOS and mixing rules. Re-run with an independent hydrate program or a different EOS; if the answer moves materially, the default Peng–Robinson kij set is not fit for this duty.
- The cold spot. The model probably used the thermodynamic valve outlet temperature only. Plot the whole path - valve, strainer, meter run, control valve - against the hydrate curve; each pressure drop adds cooling, and the lowest pressure in the meter run, not the nominal 28 barg, is the governing point.
- Back-calculate the five events. Take the recorded flows, inlet conditions and outlet temperatures and see which model reproduces 16–19 °C. If no model does, the field data or the assumed composition is wrong.
Judgement. Do not issue a revised basis from an unvalidated model. Until a model reproduces the five events within about 2 °C, the site has no defensible operating envelope.
2. Set out the mitigation options and recommend one. Include the basis for sizing the inhibitor and the operating hazards that come with it.
Read the worked answer
Options
- Continuous methanol. Lowest capex, quickest to implement. Dose from the inhibitor curve at the coldest point and the highest water rate. The trap with a 20% uplift is that water carry-over rises roughly with gas rate, so the inhibitor demand rises 20% too. Methanol is toxic and flammable (flash point near 11 °C), needs good atomisation and injection-point mixing, cannot be regenerated, and will follow the condensate and sales gas. Acceptable as an interim campaign measure, not as the permanent answer for a sustained uplift.
- MEG with regeneration. Regenerable, better for continuous duty, but contactor and regenerator capex, inventory and corrosion control are real burdens.
- Upstream TEG dehydration. Removes the root cause and also cuts corrosion. Size to a water dew point with a margin below the coldest metal temperature, typically 10 °C or more.
- Heat or reduce ΔP. A feed/product heater or a two-stage letdown with interstage reheat keeps the path above the hydrate curve without chemicals, at an energy cost, and does nothing on depressurisation.
Recommendation. Treat methanol plus a documented operating envelope as the interim for this winter, and commit to dehydration (or two-stage letdown with reheat if dehydration is impractical on this plot) as the permanent solution. Whichever is chosen, require a hydrate margin of at least 5 °C at the coldest point on the coldest design day.
Hazards to manage now. Methanol fire and toxicity, closed-drain and exposure controls, a HAZOP of the injection package, low-flow alarm and totaliser, and confirmation that methanol carry-over does not breach the sales-gas or effluent limits.
3. Define the operating envelope and the safeguards you would require, and state what would trigger a further management-of-change review.
Read the worked answer
Envelope. Build it from the validated hydrate curve, not from a single number: a plot of permissible flow against inlet temperature, bounded by a maximum ΔP across the letdown path and a minimum permissible temperature at each cold point, each with a 5 °C margin. Publish it as a table for the winter months and put it in the operating manual and the DCS graphics.
Safeguards
- Temperature transmitter at the valve outlet and at the meter run, with low alarm and a high-high trip that closes the JT valve; assess the trip's integrity (SIL) rather than assuming it.
- Methanol injection flow indication with low-flow alarm and totaliser, and an analyser for water dew point at the inlet once dehydration is installed.
- Written response to a suspected blockage: isolate, depressurise, never apply steam to a liquid-full isolated section, and never use hammers.
- Winter checklist covering cold starts, purging and drying after maintenance.
Change triggers. Any of the following re-opens the case: throughput above the validated envelope, a composition change (especially more CO2 or heavies) beyond the agreed band, a new injection point, a change of valve trim or meter run, or a failure of the low-temperature trip. Each is a management-of-change item requiring a HAZOP, updated P&IDs and cause-and-effect, updated procedures, and a check of the relief and flare cases.