‹ Home
KMUTT Department of Chemical Engineering

Pilot ethanol distillation column

Unit operations laboratory · Department of Chemical Engineering, KMUTT

Designed by Dr. Jatupon Chaiwasu

|
Top of the column28.0 °C target 78
Bottom of the column28.0 °C target 100
Sight glass0 % hold at 50
Distillate0.0 L/h bottoms 0.0 L/h

Click any part

Drag to turn, scroll to zoom. Arrow keys, the middle mouse button or the pad below left move the model about. Press Open the column to see the trays, the packing, the chimney trays and the reboiler heater.

1 / 7 · Ready
1 / 7
Ready

Nothing is running. Press play to start the feed.

everything off

ball valve globe valve hot side cooling water Vessel diameters drawn 3 times true size so the internals are readable. Heights, lengths and positions are true.

01How the column is stacked

02Where the valves are

Nine ball valves and two globe valves. Ball valves are open or shut. Globe valves throttle, which is why both pump bypasses use them. Click a row to find it in the model.

Bottom loop

Top loop

V9 decides everything. Shut it and all the condensate returns to the column, which is total reflux. Open it and the column makes product at partial reflux.

01Watch the start-up

The same run as the first tab, drawn flat so you can see every line at once. Live pipes carry a moving stripe. The dashed ring marks the equipment the current step is about.

1 / 7
Ready

Nothing is running. Press play to start the feed.

02Start-up, in order

  1. Cooling water on Open V6 before any heat goes anywhere near the column.
  2. Connect the return hose to the feed tank From the bottom of the column back to the tank. The column hangs off the frame, so the hose runs underneath it.
  3. Start the feed pump Feed enters between tray 1 and tray 2. It drains out of the bottom of the column and back to the tank, so you are filling and draining at the same time.
  4. Bring the sight glass to half Trim the hose against the pump. The bottom of the column is tied to the bottom of the reboiler, so this glass is reading the reboiler as well. Let it climb well above half and the float lifts and the heater cuts out. The reboiler is a heater, not a heat exchanger. Open the hose further to bring it back.
  5. Open the reboiler vent V10 Straight to atmosphere. The air in the column has to be able to leave or vapour cannot climb.
  6. Switch the heater on and set the kW The bottom climbs to boiling. More kW gets there faster and drives more vapour, but the float trip does not care how fast you were going.
  7. Watch the vapour front climb The top temperature follows it up. This is the column filling with vapour, not separation.
  8. Start the reflux pump, total reflux Everything that condenses goes back to the top. Now the separation starts and the top temperature turns round and falls.
  9. Wait for 78 at the top and 100 at the bottom Both have to stop moving. Until they do, any sample belongs to no particular moment.
  10. Open the distillate and bottoms valves Read both rotameters. The top drifts up as soon as you draw, which is the price of making product.

03Why total reflux first

Total reflux  at start-up only

  • Concentration the highest this column can ever give
  • Production zero
  • Use it to build the profile fast and see the ceiling

Partial reflux  normal operation

  • Concentration lower, and it falls the more product you draw
  • Production whatever rate you set
  • Use it to make 92 %v/v ethanol and test the column against the design
Every drop of product costs purity. Open V9 and the top temperature rises straight away.

04What you trade when you open V9

total reflux ceiling partial reflux
Total reflux ceiling, top, %v/v At this reflux ratio, top Bottoms at partial reflux Ethanol recovered overhead Minimum reflux for this duty Product rate

05Purity has a wall

Read this before setting a target

90 to 92 %v/v costs a jump in reflux ratio from 2.1 to 9.1. 92 to 93 costs a jump to 51.7. Past that the column cannot get there at any reflux ratio.

The red line is the ethanol and water azeotrope at 97.2 %v/v. No ordinary column passes it.

The honest target for this column is about 92 %v/v.

01You already know the equations. Here you run them backwards

In the sizing problems from your course, a separation is given and you work out how big the column must be. This column already exists. You measure how it behaved, and you work out how well it did. Same diagram, same equations, opposite direction.

Tray section

DirectionYou knowYou find
Sizingideal stages needed, an assumed efficiencyhow many real trays are needed
Ratingideal stages achieved, 5 real trays countedthe efficiency it really has
Nreal = Nideal / EO   ⇄   EO = Nideal / Nreal One equation. Sizing solves it for Nreal, rating solves it for EO.

Packed section

DirectionYou knowYou find
Sizingideal stages needed, an HETP from a charthow tall the bed must be
Ratingideal stages achieved, 2.591 m of bed measuredthe HETP it really has
Z = Nideal × HETP   ⇄   HETP = Z / Nideal One equation. Sizing solves it for Z, rating solves it for HETP.
The one new idea. In a sizing problem the ideal stage count comes from the separation the column is asked for. In rating it comes from the separation the column produced. Everything else you have done before.

02What is fixed and what you measure

Fixed by the steelwork, never changes

Real trays below the feed5
Packed height above the feed2.591 m
Reboiler1 ideal stage
Total condenser0 stages

Read off the column each run

Feed composition zFsample the feed line
Distillate xDsample the product
Bottoms xBsample the bottoms
Reflux ratio Rreflux rate / product rate
Feed temperaturethermocouple on the feed line

03Do it in six steps

above the feed, packed below the feed, trays q line

Your measurements

04The tray answer

How to read it

Count the steps on the diagram below the feed. One of them is the reboiler, so take it off. What is left is the number of ideal stages the five real trays managed between them.

EO = 3.0 ideal / 5 real = 60 %

A laboratory column of this size usually lands between 40 and 80 %. Below that, look for weeping through the tray holes, liquid bypassing the froth, or a sample taken before steady state.

Three things this number is not. It is not a whole number of steps. The last step lands past xB, so only part of it is used. Round it up and five perfect trays read more than 100 %. It is not the efficiency of one tray. That is EMV, measured in section 08. EO is the whole section at once. It is not a property of the trays alone. The same trays read a different EO at a different reflux ratio. Always quote the run conditions beside it.

05The packing answer

How to read it

Count the steps on the diagram above the feed. Those are the ideal stages the whole 2.591 m of packing produced. Divide the height by that number and you have the height needed for one ideal stage.

HETP = 2.591 m / 4 ideal = 0.65 m

Small random packing in a narrow column usually gives 0.3 to 0.7 m. A larger number means the bed is doing less than the design assumed, often because liquid is running down the wall instead of through the packing.

06Put the two together and judge it

SectionAchievedDesignResult

Design values for comparison: HETP 0.43 m, tray efficiency 100 %.

stages the column gives = 2.591⁄HETP + 5 EO + 1 Height and efficiency cannot be added. Turn both into ideal stages first, then they can.

What the answer means

07When the column does not reach the design purity

This is the normal result on an old rig. The top comes out at 89 %v/v instead of 92, and the question is not whether the run failed but which number moved and by how much.

 Design basisThis runGap

The design basis is the duty the equipment was built for: 10 %v/v feed at 52 L/h, R 9.06, 92 %v/v overhead, 6 ideal stages of packing and 5 trays counted as 5 ideal stages.

Four checks, in this order

Can more reflux fix it?

this bed, at the stage count you measured the same bed giving its design 6 stages

Can the reboiler even boil that hard?

15 kW is the installed rating. An element that has scaled, or lagging that has fallen off, delivers less. Measure it from the electrical input, or from the condenser cooling water duty.

What each cause does to the numbers

Physical causeWhich number moves
Packing fouled, or liquid running down the wallHETP rises, tray EO unchanged
A redistributor blocked, so one bed runs dryHETP of that bed alone rises
Tray holes fouled or the column weeping at low vapour rateEO falls, HETP unchanged
Reboiler element scaled, so the boil-up is shortR falls below the value you set, both sections read low
Vapour leaving through the vent instead of condensingreal R is lower than the pump setting, mass balance does not close
Cooling water too warm, reflux returning hotinternal reflux differs from the measured external reflux
Sample taken before the profile settledevery number moves, and none of them mean anything

08Measure the trays one at a time

EO divides one number by five. It only works when the tray section is doing enough separation to show up in the bottoms reading. Sampling each tray gives a separate number for every tray and works however lightly the section is loaded.

Liquid sample from each tray, %v/v

The sample above tray 1 is the liquid arriving from the feed point. The sample below tray 5 is the sump, which you already entered as xB.

What comes out

EMV = (yn − yn+1) / (y*n − yn+1) You measure liquid, so both vapour compositions come from the stripping operating line: yn = mS xn-1 + bS and yn+1 = mS xn + bS. Only y*n comes from the equilibrium curve, at the liquid on the tray itself.
One tray is not the whole section. EMV describes a single tray, EO describes all five together, and they are not the same number. They are linked by EO = ln[1 + EMV(λ − 1)] / lnλ with λ = m V / L. Quote both and say which is which.

09Which bed lost the stages

The packing is three separate beds with a redistributor between them. One HETP for the whole 2.591 m hides a bed that has stopped working. Sample at each redistributor and you get one HETP per bed.

Liquid sample down the packing, %v/v

BedHeightIdeal stagesHETP

Design HETP is 0.43 m in every bed. A bed well above that is where to look first.

10Before you calculate anything, convert

%w/w ethanol mole fraction x bubble point at 1 atm relative volatility here
An alcoholmeter reads volume percent, and ethanol and water shrink when mixed. Every equation on this page uses mole fraction.

11If you want to go further

Describing the packing with transfer units instead of stages

Stages treat the bed as if it were a stack of perfect trays. Transfer units follow the real driving force up the bed instead. Both describe the same packing and they are linked by one relation.

NOG = ∫ dy / (y* − y)HOG = Z / NOG
HETP = HOG · lnλ / (λ − 1)λ = mV/L, with m the slope of the equilibrium curve in that part of the bed.
NOG for your run HOG λ averaged over the section HETP from transfer units HETP from counting stages

The two will not match exactly, because ethanol and water have a strongly curved equilibrium line so λ changes from the bottom of the bed to the top. Quoting both, and explaining why they differ, is a better answer than quoting one.

Three ways these numbers can mislead you

Near the azeotrope. Above about 93 %v/v the equilibrium curve and the diagonal almost touch, so the step count swings wildly with a small sample error.

A section that is barely loaded. At the design reflux ratio the stripping duty is so small that the five trays and the reboiler together are asked for under two ideal stages. Any tray efficiency then gives nearly the same bottoms reading, so EO is not measurable from the end compositions. Measure the trays one at a time instead.

Rounding the step count. The staircase almost never lands exactly on xB. Count the last step as the fraction actually used. Rounding up adds a whole stage and can push a perfect tray above 100 %.

Steady state. A 1 m sump holds a lot of liquid. If the thermocouples are still drifting, the composition profile is still moving and the stage count belongs to no particular moment.

Ethanol and water at 1 atm

Relative volatility is large at the water end and small near the azeotrope. That is why a few trays are enough at the bottom and a tall packed bed is needed at the top.

01Check yourself