Debutanizer Flooding After a Throughput Revamp

What the three symptoms say together

Delta P. The column is at 0.51 bar against 0.32 bar at design, yet reboiler steam is fixed at 4.2 t/h, so the vapor boilup has not changed much. The extra pressure drop is therefore liquid head, not vapor rate. That points to liquid holdup and downcomer backup, not to a higher vapor load.

Drum level. A 20-minute cycle is the signature of a surging, flooded section: liquid accumulates until the downcomer seal breaks or the froth collapses, then dumps into the drum, and the cycle repeats. A level-controller tuning problem would not produce a simultaneous delta P rise.

Gamma scan. Holdup is roughly doubled on trays 6-9 and very low on 26-31. The high-holdup zone is where the flooding starts; the trays below are starved of liquid because the flooded section is holding up the downflow. Because the flooded zone is above the feed, this is rectifying-section flooding, not stripping-section flooding.

Why the rectifying section first. The revamp raised both vapor and reflux traffic, and the rectifying section has the highest liquid-to-vapor ratio in the column. The reflux is also returned subcooled, which adds liquid to the top trays, so that section reaches its downcomer limit first.

Why C5 gets overhead. Two mechanisms act together: froth entrainment carries heavy key up to the condenser, and the flooded zone loses effective stages, so the C4/C5 separation degrades. The slip is intermittent because the flooding is cyclic, which is exactly what the drum-level cycling shows.

The high delta P is the consequence, not the cause. The column is not dirty and the trays are probably not damaged; it is operating beyond the hydraulic capacity of the top section.

Moves, in priority order

1. Cut reboiler steam to about 3.6 t/h (-15%). This is the fastest way to reduce the vapor load driving the froth. Watch the bottoms temperature and the stripping-section delta P. The risk is that the bottoms loses C4 and goes off-spec, so cut in steps and hold each step for 15 minutes before cutting again.
2. Cut feed to about 90 kmol/h. This reduces both the liquid and vapor traffic in one move. The commercial risk is real, but one off-spec parcel is cheaper than a three-day shutdown.
3. Hold the pressure at 12.0 barg, or lower it by 0.5 bar if the condenser approach allows. Lower pressure reduces vapor density and helps, but do it after the steam cut; moving two variables at once makes the response unreadable.
4. Leave the reflux ratio where it is. The reflex to raise reflux to fix the C5 slip is wrong here: reflux adds liquid to the already flooded rectifying section and makes it worse.
5. Put the reflux drum level on manual if it is driving the reflux flow in a cycle. A level controller chasing a surging drum will amplify the disturbance instead of damping it.

What not to do

• Do not increase reboiler steam to push the C5 back down; that raises the vapor load and deepens the flood.
• Do not try to fix the specification by raising column pressure; higher pressure increases vapor density and makes the flooding worse.
• Do not speed up the drum level controller. The cycle is hydraulic, not a tuning problem.
• Do not open the feed further in the hope of steadying the column.

The delta P should fall within 30-60 minutes of the steam cut. If it does not, either a tray or downcomer is damaged or the gamma scan is being misread, and you should plan to shut down rather than ship.

Why lowering pressure buys capacity

Vapor density is proportional to absolute pressure at constant temperature and compressibility: rho_v = PM/(ZRT). Going from 13.0 bar abs (12.0 barg) to 12.0 bar abs is a 7.7% reduction in vapor density, so at the same mass boilup the vapor volumetric flow through the trays falls by about 7.7%. Tray capacity is set by the F-factor, u times the square root of rho_v, so the vapor load on the downcomers falls by roughly 4% even before the volumetric effect on the downcomer is counted. In addition, the relative volatility of C4/C5 rises as pressure falls, so the same tray count does more separation, which directly attacks the C5 slip.

Two constraints that can make it infeasible tonight

1. Condenser capacity and cooling water temperature. The condenser is already at an 8 C approach. At 1 bar lower, the C4 condensing temperature falls by roughly 8-10 C, so the required cooling water temperature may fall below what the tower can deliver in the evening. You would lose pressure control and the column would drift. Check the condenser outlet temperature against the actual cooling water supply curve before moving the set point.
2. Downstream pressure requirement. The overhead C4 has to be condensed and pumped to the next unit or to storage. If the receiver or the product pump needs a minimum suction pressure, lowering the column pressure may violate the downstream battery limit even though the column itself is happy.

Secondary checks: the reboiler temperature falls, so the steam pressure may need to be reduced to avoid over-boiling; and the PSV set points and pressure controller range on the column must still be valid at 11 barg.