Debottlenecking a Depropanizer That Is Already at 95% of Flood

What the symptoms already tell you. A gradual rise in tray ΔP over 18 months with a loss of separation points to fouling, and fouling usually shows up first as downcomer restriction or as foam stabilisation, not as classic jet flooding. The fact that raising reflux made it worse is diagnostic: more reflux increases both vapour and liquid loads, so it drives the column further into whatever capacity limit it is already at.

Discriminating the three mechanisms

• Jet flooding: ΔP rises steeply and fairly uniformly up the column, entrainment rises, efficiency falls across all trays, and it usually appears first at the top where vapour loads are highest.
• Downcomer flooding: ΔP shows a distinct step with a high downcomer backup, level and reboiler instability, and it appears where liquid load is highest - often lower down. The unused downcomer review is the obvious gap.
• Foaming: erratic, sawtooth ΔP, sudden efficiency collapse, and a clear link to feed contamination. The amine carry-over history makes this credible.

Data to request

• Tray-by-tray ΔP from the historian, plus reboiler duty, reflux, feed rate, pressure and the level bridle indication.
• A gamma scan or tracer test to locate the active section and measure froth height; a downcomer residence-time and weir-loading check against the tray rating.
• A hydraulic rating from the as-built drawings to identify the limiting section (top trays, bottom trays, downcomer, or the reboiler return line).
• Verification of the flow and pressure transmitters; a misreading transmitter can mimic flooding.
• Feed analysis for amine, water and heavies.

Judgement. Do not buy a revamp until the limiting section and mechanism are known - a packing revamp will not fix a fouled feed or a downcomer limitation.

Why more reflux fails. At fixed product specification, raising reflux raises boil-up and condenser duty, so both the vapour and liquid loads on the trays increase. A column already at 95% of flood has almost no hydraulic margin, so reflux pushes it over the limit and the separation falls rather than rises. Reflux can only be traded against throughput; it cannot buy capacity.

Options

• Low-cost, in the three-day outage: optimise the feed tray, clean or replace fouled trays, check and clear downcomers, check the reboiler return and the condenser for air-side fouling, and inspect for amine carry-over sources. This can recover the lost 18 months, but will not deliver 15% on its own.
• Tray modification: fixed valves, increased downcomer area, or high-capacity trays - typically 10–25% capacity gain, moderate capex, possible in a short outage.
• Structured or random packing revamp: largest gain (often 25–40%) but a full turnaround, and the column's pressure-relief case and pressure drop must be revalidated.
• Reduce the load: feed preheat or a feed split to a second column reduces reboiler duty but rarely gives 15% at the same separation.
• Debottleneck the utilities: reboiler, condenser and, critically, the relief valve and flare header.

Recommendation. Take the three-day outage to clean and inspect, gather the gamma scan and hydraulic rating, and use that to decide between a tray modification and packing at the 14-month turnaround. For the export window, only accept throughput that can be shown to sit within a validated flood margin (typically ≤85% of flood) and within the existing relief case.

The hard constraint. The relief valve was sized at the original reboiler duty and the flare header is already tight. No throughput increase is legitimate until the loss-of-cooling relief load, two-phase flow and flare header hydraulics have been re-run and accepted. This may be the real limit on the whole project, not the trays.

Is it a change? Yes. Operating above the validated hydraulic and relief basis is a change even if no hardware is altered, so it must go through the site management-of-change procedure with the same rigour as a physical modification.

Required work

• HAZOP of the column and its relief system, focused on higher vapour and liquid loads, higher reboiler duty, higher condenser duty, and the loss of cooling, reflux failure, power failure and blocked outlet cases at the new rates.
• Relief and flare revalidation: re-run the governing scenario at the increased duty, check for two-phase relief, re-rate the relief valve, and confirm the flare header capacity in the site-wide worst case (including the existing tight scenario).
• Hydraulic re-rating of trays and downcomers, the reboiler circuit, the condenser, the reflux and bottoms pumps, and the rundown to storage.
• Control and alarms: confirm the level, pressure and temperature control loops and the trips are still valid at the new rates; review the high-pressure trip setpoint against the new relief basis.
• Documentation: update P&IDs, operating envelope, procedures, training, and the inspection and RBI plan; notify the insurer and any party required by the site's statutory obligations.
• Verification: commission at stepped rates with a hold point at each step, and confirm the tray ΔP and separation match the prediction before releasing the next step.

Judgement. If the relief case cannot be closed within the export window, the correct professional answer is to decline the 15% uplift, document why, and offer the capacity that the existing relief basis supports.