Flue Gas Absorption Tower

What is a flue gas absorption tower?

A flue gas absorption tower is a vertical vessel in which an upward-flowing gas meets a downward-flowing liquid so that a pollutant crosses from the gas into the liquid. The vessel creates interface between the two phases, holds it for a residence time through a contact zone, and then takes the liquid back out of the gas before the gas leaves. In practice it consists of a shell, a gas inlet transition, a contact zone filled with spray headers, trays or packing, a staged mist eliminator above the last spray, a sump at the bottom, and a recirculation circuit that returns the liquid to the top. It is bought for three numbers that cannot be separated: the removal it will achieve, the liquid-to-gas ratio at which it will achieve it, and the pressure drop it will consume doing so.

Why is the liquid-to-gas ratio the number a tower is really bought for?

Because the ratio determines how much wetted surface the liquid presents, how much pump energy the circuit consumes, and how much the liquid load raises the pressure drop across the packing, and because removal rises with the ratio toward an asymptote while the cost rises with it without limit. A tower quoted at a removal figure with no ratio attached has had its most important design variable left to the supplier’s convenience. The ratio is also the number the owner can check on site, because it is the sum of the active spray header flows divided by the measured gas flow, and both of those are measurable at any time during the life of the plant.

Why do two bids quote liquid-to-gas ratios that differ by five times?

Because four mutually incompatible conventions circulate under the same abbreviation. Published sources in this class use litres of liquid per cubic metre of gas, kilograms per kilogram on a mass basis, gallons per minute per thousand actual cubic feet per minute, and at least one manufacturer’s own internal ratio. A published hundred and twenty gallons per minute per thousand actual cubic feet per minute converts to about sixteen litres per cubic metre, and a published mass ratio of ten converts to about ten litres per cubic metre of a saturated gas at fifty degrees where the density is about one kilogram per cubic metre, or about thirteen if the reference is dry air at zero. Those figures are the same order of magnitude once converted. A published single-manufacturer example that quotes fifty-six, eighty-two and a hundred and eight for three designs of the same duty cannot be placed on either basis from its own page, and a ratio of that kind compared against a ratio of ten is very often a comparison of conventions rather than of engineering.

Why is a gas velocity of three to four metres per second wrong for a packed tower?

Because that figure is a design value for an empty spray section rather than for a bed. Published design bands put a packed bed in counter-current flow at one and a half to three metres per second with one and a half to three centimetres of water gauge per metre of packing, where the same figure of three to four metres per second appears in the published patent literature specifically as the void-tower gas velocity with three to four seconds of contact time and no packing present. Velocity has no absolute meaning in a tower: it has meaning only as a fraction of the flooding velocity of the packing that is actually installed, because that fraction decides whether the bed runs wet, becomes loaded, or floods. Applying a void-tower velocity to a packed bed would put the bed above its loading point, which raises the pressure drop steeply and can hold liquid in the bed.

Spray tower, packed tower or tray tower: how is the type decided?

By what the liquor does to the internals first, and by the removal target second. An open spray tower has nothing in the contact zone to plug or to scale, so it is the default where the liquor carries solids or forms a deposit, and it pays for that immunity by needing the highest liquid ratio of the three to reach the same removal. A packed bed gives the most transfer per unit of height and the lowest liquid ratio, and it is the wrong choice the moment the liquor can coat a wetted surface, because the void volume that gives the packing its low pressure drop is the void volume a scale closes. A tray tower sits between them and buys the transfer at a lower liquid ratio than an open tower while charging additional pressure drop. Published design bands differ between the three by a factor of three in velocity and a factor of five in pressure drop, which is why the arbitration is a written record with the rejected options named.

What does a tray buy and what does it cost?

It buys liquid circulation. A published comparison on one six-hundred-and-seventeen-megawatt duty runs a two-tray design at a liquid ratio of fifty-six with thirty-four thousand five hundred gallons per minute of circulation, a one-tray design at eighty-two with fifty thousand, and an open tower at a hundred and eight with sixty-six thousand. The open tower therefore needs about ninety-three per cent more liquid to do the same job. What the trays cost is pressure drop: ten and two-tenths inches of water gauge for two trays and eight and two-tenths for one, against seven inches for the open tower, a forty-six per cent increase. Adding the two together, the total fan and pump power in the published example is six thousand four hundred and seventy kilowatts for the two-tray design against seven thousand four hundred and seventy for the open tower, so the trays come out ahead on the total energy account. That is the arithmetic a buyer has to see before choosing a type, because the type is an energy decision rather than a quality decision.

Why is a packed bed risky in a scaling slurry?

Because the properties that make packing attractive are the same properties that make it vulnerable. Packing gives a large wetted area and a low pressure drop per metre because it has a large surface and a large void fraction, and a scale deposits on surfaces and grows into voids. The consequence is that the pressure drop rises and the wetted area falls at the same time, so the tower loses both its hydraulics and its transfer, and it does so gradually enough that the loss is often noticed at a stack reading rather than at the bed. Published selection guidance in this field describes packing types and rates them on pressure drop and capacity, but does not connect the choice to the scaling and solids behaviour of a particular liquor, so the connection has to be made by the specifying engineer. Where the liquor carries solids or deposits, the void volume, the wash-down arrangement and the zonal pressure-drop monitoring that detects blinding before it becomes a failure are design items rather than maintenance items.

How is the contact height of a tower set?

The height is the output of a transfer calculation, not a figure copied from a similar installation. The calculation takes the inlet pollutant load, the removal required and the driving force available at the liquid-to-gas ratio chosen, and produces the contact volume the duty needs. That volume is then split between cross-sectional area and height by the velocity decision, which is why a plant that doubles its flow gets a wider tower rather than a taller one and a plant that tightens its limit gets a taller bed rather than a blindly higher liquid ratio. The height is then made up of the packing depth plus the spray-to-bed distances, the space the redistribution device occupies, and the eliminator space above the last spray, and the assumptions of the calculation are issued with it so that they can be revisited when the gas changes.

What is the flooding fraction and why design below it?

The flooding velocity is the gas velocity at which the liquid can no longer descend through the bed and begins to be held up in it, so the pressure drop climbs steeply and the tower’s hydraulics become unstable. The flooding fraction is the operating velocity expressed as a percentage of that velocity, and it is the only meaningful way to state a tower’s gas velocity because the flooding velocity depends on the packing, on the liquid-to-gas ratio and on the liquor’s properties. Operating well below the flooding point keeps the bed wet, the pressure drop predictable and the turndown available; operating close to it buys throughput per unit of steel at the cost of hydraulic stability. Because the fraction is stated against the packing actually proposed, a later change of packing does not silently move the bed off its design point.

How is the pressure drop budget built, and what does it cost to run?

The budget is the sum of five line items: the inlet transition, the spray zone, the packing bed, any redistribution device, and the mist eliminator. The velocity head of the gas at three metres per second and a density of about one kilogram per cubic metre is only about five pascal, so with a loss coefficient of order one the inlet transition costs a few pascal and under one per cent of a two-thousand-pascal tower. The bill is in the packing and the eliminator. Published figures for packing are one and a half to three centimetres of water gauge per metre in the counter-current design band, which is about a hundred and forty-seven to two hundred and ninety-four pascal per metre, against about four and a fifth centimetres per metre, about four hundred and ten pascal per metre, for the published general packing figure. The cost to run is arithmetic: at a hundred thousand cubic metres per hour, which is twenty-seven point eight cubic metres per second, two thousand pascal at seventy per cent fan efficiency is about seventy-nine kilowatts and fifteen hundred pascal is about sixty kilowatts, so the nineteen-kilowatt difference over eight thousand hours is about a hundred and fifty-two thousand kilowatt hours.

Why does the mist eliminator limit how far a tower can be turned down?

Because it normally occupies the tower’s own cross-section and therefore runs at the tower’s own superficial velocity, so when the gas flow falls the eliminator’s velocity falls with it. Its separation is inertial, which means its cut point improves as the velocity rises and worsens as the velocity falls, and its upper limit is re-entrainment, the velocity at which the collected film is torn back off the vanes. The result is an operating window in the same variable as the tower’s own velocity, and the tower’s minimum load is the point at which the eliminator stops separating rather than the point at which the packing stops transferring. A tower quoted for a wide turndown with a full-cross-section eliminator is a tower whose eliminator was never checked. Where the required turndown exceeds what a shared cross-section allows, the eliminator can be given its own smaller flow area so that its velocity is held while the tower’s falls, and the price of that is additional pressure drop.

What happens to removal and carryover at half load?

At half the design gas flow with the circulation unchanged, the tower’s superficial velocity halves, the liquid-to-gas ratio doubles, and the two move in opposite directions in cost. The packing’s pressure drop falls to roughly a quarter because it follows about the square of the velocity, and the fan’s energy for the tower’s own drop falls to about an eighth because it follows the flow times the drop. The pump’s energy per unit of gas, by contrast, doubles, because the liquid flow is unchanged while the gas flow is not. The removal falls, because both the contact time and the driving force fall, and the carryover margin is lost, because the eliminator is now at the lower limit of its window. The part-load penalty of a tower is therefore a pumping penalty rather than a fan penalty, and the lever that removes it is the spray-level schedule, which takes headers out of service as the load falls and returns the liquid-to-gas ratio toward its design value.

How much wash water does a mist eliminator need?

The published catalogues in this class state that effective irrigation of the eliminator is essential to prevent fouling and that the wash has to reach the whole vane face, and they give no number for the flow that achieves it. The design answer is therefore not a flow rate but a coverage: the arrangement is specified as a coverage per unit plan area, a cycle interval and a spray pattern that overlaps the vane geometry, sized against the solids the liquor carries and the rate at which they deposit. What matters is whether every square metre of vane face is reached between two cycles, not how many litres per hour the wash consumes, because a generous flow that misses a corner of the vane leaves that corner blinding while the total looks adequate. The wash drainage is specified with it, returning the collected liquid to the sump or to the blowdown according to its load.

Why does the gas inlet need a transition and wall rings?

Because a tower whose gas arrives asymmetrically spends its cross-section unevenly and returns less removal than a well-distributed bed of the same height would give. The transition brings a duct velocity of a few metres per second up to a tower velocity of one to three over a length that lets the gas spread before it reaches the first spray plane, and its design criterion is not its shape but the velocity deviation across the section at that plane. Wall rings or an internal awning are placed above the inlet because without them a film of gas runs up the inside of the shell and bypasses the sprays, which is the failure that published equipment vendors in this field describe in exactly those terms and remedy with exactly those devices. The inlet is also placed above the maximum liquor level, so that the entering gas cannot impinge on the liquor surface and lift droplets back into the duct.

What is the wet-dry interface and why does it fail first?

It is the junction where the hot dry inlet gas meets the saturated zone of the tower, and it is the point at which this class of vessel most often fails. At that junction acid condenses on a surface that may also be dust-laden, salt deposits nucleate and grow in the condensed film, and chlorides concentrate against the lining as the film evaporates and re-condenses. Because it is a small area with a large temperature and moisture gradient across it, it is also the least tolerable place for a fabrication shortcut. The detail is therefore part of the tower drawing rather than site work: the inlet is placed above the maximum liquor level, the junction is drained so that condensate has somewhere to go other than down the wall or back down the duct, that zone is detailed in a material rated above the rest of the vessel, and the joint is arranged so that it can be inspected and repaired without cutting the shell.

Which material for the tower shell, and why is stainless steel refused?

The material follows the size, the temperature and the chemistry at each location rather than one grade for the whole vessel. Published construction options in this class run from corrosion-free thermoplastics through glass-reinforced plastic, steel with a rubber lining and concrete with a polypropylene lining, and published moulded towers reach above ten metres in diameter and thirty metres in height with a corrosion liner over a hundred mils thick. Stabilised polypropylene is the default for a moulded shell in saturated acid service; glass-reinforced plastic extends the size and the temperature; steel with a lining and concrete with a lining take over at the scale where a moulding is no longer practical. Stainless steel is refused in the wet acid phase of this service because the passive film is broken by chlorides and by the reducing conditions that occur in pockets and under deposits, so the grades that look like the obvious answer are the ones that pit. The real constraint that decides the material is temperature: a thermoplastic has a softening point, and the tower inlet limit is that softening point with a margin, which is why a quench or a dilution stage is frequently part of the tower package rather than an accessory.

Why does the tower need a sump with oxidation air?

Because the sump is the reactor of the circuit rather than a storage volume. The liquor returning from the spray headers has to be given the residence time the neutralisation needs, and where the process route requires the conversion of a sulphite to a sulphate that conversion has to take place somewhere with enough time and enough oxygen. Introducing the oxidation air at the bottom of the sump against the liquor head makes it dissolve into the liquor; introducing it across the surface lets most of it escape. The sump volume is therefore set from the residence time multiplied by the circulation rate, the air rate is set from the conversion the chemistry demands rather than from the sump volume, and the agitation that keeps solids in suspension is designed together with the air because the same air also moves the liquor. The blowdown is then taken off at the point of the highest dissolved or suspended load and controlled on that measurement, which is what gives the circuit a defined steady state instead of an accumulating one.

Why divide the packing into sections with a redistributor?

Because liquid descending a deep bed migrates to the wall and starves the centre of the bed, so the gas passing through the middle meets a surface that is no longer fully wetted and the transfer efficiency falls below what the height was bought for. Published descriptions of this equipment name the mechanism precisely, in terms of wall flow causing an uneven distribution of the two phases in the bed and a loss of transfer efficiency, and prescribe the remedy: divide the bed into two or more sections and install a redistribution device between them, so that the liquid is re-presented to the top of each section instead of being allowed to complete the whole descent in one pass. The division is standard practice in this class rather than an option, and each section is given the access that lets its distribution be inspected, because a bed that cannot be inspected becomes an unknown after the first campaign.

What is measured to accept an absorption tower?

Eight things, all of them before the vessel is ordered rather than after it is built. The gas flow is measured at the inlet including the dilution air the capture admits; the liquid-to-gas ratio is established by summing the flows of the active spray headers and dividing by the measured gas flow, because a pump curve proves only that the pump was running; the pressure drop is recorded by zone across the inlet transition, the spray zone, the bed, the redistributor and the eliminator, so that the as-built can be compared with the budget line by line; the outlet droplet carryover is measured with an instrument of stated accuracy at both the design flow and the minimum flow; and the removal is demonstrated at the design ratio and the design load. The turndown curve is then walked down with spray levels taken out of service until the eliminator’s lower limit is reached, and that flow is marked as the minimum at which the guarantee still applies. The water and solids balance is measured over a steady period, and the liner and the wet-dry interface detail are examined before commissioning and again after the first campaign. Nothing in that list depends on a design code, because no published page in this class cites one for the absorber vessel: the governing instrument is the owner’s specification and the test is what defines the guarantee.

When is an absorption tower the wrong answer?

When the gas has never been measured, because a tower sized on an assumed flow and an assumed inlet load is a tower that will be re-specified. When the liquor’s solids or scaling behaviour rules out the packing that the removal target would otherwise need and there is neither the height nor the footprint for a spray or tray alternative. When the gas arrives hotter than the softening point of every available polymer and there is no quench and no upstream cooling available. When the site has no legal route for the blowdown, because the tower moves a pollutant into a liquid rather than destroying it and a liquor with nowhere to go is not a solution. When the requested turndown is outside the velocity window the eliminator can hold, because the guarantee would apply only at the design point. And where the honest answer is a change upstream — a different reagent, a lower-solids route, a capture change that reduces the dilution air, or a dust removal stage ahead of the vessel — because a smaller tower handling a better gas is usually cheaper over its life than a larger tower handling the gas the plant happens to produce.