Distillation Column Troubleshooting

When a column stops making spec, the instinct is to chase reflux. Often the answer is somewhere else entirely. Here is a structured approach, in the order I actually use it.

Before You Touch Anything

Before changing a single setpoint, gather five pieces of data. If any are unavailable, the column will mislead you.

Current vs design conditions: pressure, overhead temp, bottom temp, reflux ratio, feed rate, feed composition. Lay design and current side by side.
Trend data for the last 72 hours: look for the inflection point where the column went off-spec. Was it gradual or a step change? The shape tells you the cause class.
Pressure drop (dP) across the tower: per tray or total. Critical diagnostic. Compare to design and historical normal dP.
Gamma scan or differential pressure profile by section if available. A scan can be done in 4-6 hours and saves days of guesswork.
Feed composition: pulled from lab in the last 24 hours, not last month. Crude diet change is the most common root cause of "random" column problems.

Adjusting reflux without a composition trend and a dP profile is like tuning a PID loop by feel. You might get lucky; you will not converge.

Diagnosis by Symptom: Decision Tree

Symptom 1: Overhead product off-spec (too heavy)

Reflux low? Check reflux flow vs design L/V. If L/V is low, increase reflux first — it's the cheap fix.
Condenser capacity limited? Check condenser dP, CW inlet/outlet temps. A fouled condenser shows up as overhead pressure rising with reflux increase.
Feed zone flooding? dP profile will show a spike at the feed tray. Caused by two-phase flashing in feed pipe or over-subcooled reflux.
Downcomer flooding? Increased tower dP combined with product quality loss. Tower is hydraulically over its limit; cut feed rate.
Weeping below the feed tray? dP below design in the stripping section. Cut reboiler first, then check for vapor bypassing.

Symptom 2: Bottoms product off-spec (too light)

Reboiler duty low? Check steam flow / condensate temperature / hot oil flow. Fouled reboiler is common after 6-12 months.
Feed quality too high? Unexpected vapor fraction entering tower from upstream flash drum heaters. Check feed enthalpy calc.
Stripping trays damaged? Happens after upsets, water slug, or cavitation events. Requires column entry to confirm.
Vapor bypass through manhole or support ring? Post-turnaround failure mode — gaskets not installed.

Symptom 3: Pressure unstable, swings >2-3 psi

Condenser duty fluctuating? CW supply temp changing, or air cooler variable speed drive hunting. Check the ambient.
Overhead accumulator level control cycling? Retune or switch to cascade on reflux.
Vent/pressure control valve oversized? Common on revamp towers — install characterization trim or split-range.
Subcooling in the condenser? Pressure drops suddenly when cold shock happens. Adjust CW outlet temp target.

Symptom 4: High tower dP, capacity declining

Foaming? Look for unstable level on sumps + high dP combined with product quality collapse. Add antifoam; check for amine/solid contamination in feed.
Fouling on trays? Progressive dP rise over weeks/months. Common services: amine regenerator (heat-stable salts), sour water stripper (ammonium bisulfide), FCC main fractionator (coke fines).
Jet flooding (spray flood)? Vapor velocity above design. Check for feed composition shift to heavier/colder giving more vapor traffic.
Downcomer backup? Typically caused by downcomer filled with foam or too-high tray liquid. dP profile shows a uniform rise.

Symptom 5: Poor separation with adequate reflux and reboil

Damaged internals: common after water slug, compressor surge, relief event. Run a gamma scan.
Feed entering wrong tray: Feed distributor blocked or stripped. Tower behaves as if feed enters 5-10 trays off target.
Column operating in wrong regime: Packed tower with vapor velocity below loading point — poor HETP performance.
Vapor bypass: Chimney trays or total trapout trays with missing caps. Usually post-turnaround finding.

The Pressure-Drop Signature Guide

dP pattern	Likely cause	Confirming test
Uniform dP rise across all sections	Foaming or tower overloading	Reduce feed rate 20% — dP drops immediately if flooding, lingers if foaming
Spike at feed tray	Two-phase feed, slugging, or feed nozzle damage	Reduce feed preheater outlet temp 10 F; watch dP settle
Zero dP in stripping section	Tray weeping or damage	Increase reboil; if dP doesn't rise, internals are damaged
dP near zero everywhere	Vapor bypass (e.g., manhole gasket, trapout tray)	Gamma scan is definitive
Steady slow rise over weeks	Fouling	Lab on bottoms for iron sulfide / polymer / amine degradation products
Asymmetric side-to-side	Tray tilt or partial collapse	Differential thermocouples at same elevation; gamma scan

Gamma Scanning: Worth the Cost

A gamma scan is a non-intrusive density profile of the column. It can identify flooding, weeping, fouling, tray damage, foaming, and vapor bypass in a single 4-6 hour test. It costs $15-30k but typically saves days of incorrect troubleshooting. When you have unexplained performance loss, scan first, tune later.

Flooding signature: High-density plateau in the downcomer area above design.
Weeping signature: Low density at tray decks plus high density at the tray below.
Foaming signature: Smooth gradient instead of discrete tray steps.
Missing tray signature: Null zone (no density change) at the tray elevation.
Packed tower maldistribution: Asymmetric side-to-side scans — one-sided wetting.

Common Industry-Specific Issues

Crude atmospheric tower

Desalter upset → chloride in overhead → corrosion + salt plating on top trays. Neutralizer and wash water fix.
Product draw tray swap → operator pulls kero from diesel draw; returns look fine but yield is off.
Stripping steam tube leaks → increased column base temp but no separation benefit.

Amine regenerator

Heat-stable salt buildup → caustic neutralization once HSS > 2 wt%.
Oxygen ingress → amine oxidation products foul trays; fix the storage tank blanket.
Foaming is the #1 call — always run antifoam continuously at 5-10 ppm; do not bolus-dose.

Sour water stripper

Ammonium bisulfide (NH4HS) deposits on trays above the feed at cooler zones. Rx: hot water wash during shutdown.
Feed nitrogen too high → NH3/H2S ratio off → cannot strip both simultaneously without side draw.

FCC main fractionator

Coke fines fouling the slurry pumparound → loses pumparound duty → temperature profile pulls up → overhead gets heavier.
Slurry settler upset → catalyst in HCO draw → erodes pumps.

Deethanizer / debutanizer

C3 slipping to overhead → almost always condensing temp too cold (subcooled reflux).
C5 in LPG → reboiler fouled or trays damaged. Rarely an actual separation issue.

The Fix Order: Cheapest First

Check lab: are you measuring the right thing at the right frequency? A bad sample invalidates everything.
Check trends: when did it change? What else changed then?
Check mass balance: does vapor/liquid leaving match feed entering? If not, you have a leak or a meter error.
Adjust reflux and reboiler first: they are free to adjust.
Adjust feed preheat: small duty changes shift the feed enthalpy and column profile.
Cut feed rate 10-20%: diagnostic for flooding vs damaged internals.
Run gamma scan: definitive on internals, foaming, flooding.
Column entry: last resort — costs a turnaround window.

When to Stop and Ask for Help

Pressure spikes exceeding design — risk of relief or damage. Cut rates and investigate before continuing.
Product off-spec across both overhead and bottoms simultaneously — strong indication of vapor bypass, gross foaming, or feed enthalpy issue.
dP is uniformly zero — vapor is finding another path. The column is essentially not doing its job.
Tower shell or feed line temperature anomalies — internal damage, localized flashing, or leak.

Instrument Cross-Checks Before Everything

Half of "column troubleshooting" calls end at a bad instrument. Always check:

Differential pressure transmitter: condensed liquid in taps, isolation valves closed, tubing plugged.
Overhead temperature: thermowell damage or process condensate lining the TI.
Reboiler flow: fouled orifice plate, cavitation in the meter run, meter factor out of date.
Level transmitters: interface measurement invalid when foam or wax is present.

References

Kister, H. Z., "Distillation Troubleshooting", Wiley, 2006 (the definitive practical text).
Kister, H. Z., "Distillation Design", McGraw-Hill, 1992.
Tray Pressure Drop: Fair's correlation, Perry's Chemical Engineers' Handbook, Section 14.
Gamma scan methods: Tracerco, Quest TruTec technical papers.
API Standard 670 — column instrumentation reliability.
GPSA Engineering Data Book, Section 19 — Fractionation.

Rev 1 — practical reference. Every column has a personality; the structured diagnostic order is more valuable than any individual fix.