FRP Scrubber and Absorber Internals: Packing Supports, Spray Headers and Mist Eliminators

When an FRP scrubber or absorber underperforms, the shell is almost never the reason. The vessel is still intact, the liner is still resin-rich, and the structure is still sound. The problem is behind the manway: a packing bed that has collapsed into one side of the column because the support grid deflected, a spray header whose outer nozzles were replaced with a different spray angle during a shutdown, a vane bank running 60% above its design velocity after a production increase, or a distributor with blanked drip points that has been quietly producing zonal maldistribution for three years.

Internals are the part of an FRP scrubber that decides removal efficiency, pressure drop and the cost of the next shutdown, and they are the part most often specified by catalogue page rather than by calculation. This guide sets out how each internal is sized, what it must be made of, how it interacts with the FRP shell and its corrosion liner, and what to check during commissioning and shutdown. The numbers are drawn from the vessel standards that govern the shell (ASTM D3299, ASTM D4097, ASME RTP-1) and from mass-transfer design practice for packed and spray columns.

1. What "Internals" Covers, and What Each Part Has to Do

A wet scrubber's internals split into four functional groups, and each imposes a different kind of load on the vessel. Getting the load path right matters more in FRP than in steel, because FRP is not tolerant of point loads applied through an unprotected wall, and because a corroded-through internal support is invisible until the bed drops.

• Liquid distribution: the inlet distributor, spray headers, weir boxes, trough distributors or spray nozzles that wet the packing or fill the tower cross-section. These load the shell primarily through their own weight and through the reaction of the spray, and they must remain level to work at all.
• Packing and its support: the support grid or plate that carries the dead weight of the packing plus the liquid holdup, usually on a support ring laminated or bolted into the shell.
• Droplet capture: the mist eliminator — vane bank, mesh pad or fibre bed — plus its support grid and, critically, its wash system.
• Ancillaries: redistribution collectors and redistributors, liquid collection trays, internal piping, level and pH instrumentation, and the gas inlet geometry at the bottom of the column.

The performance consequence of getting internals wrong is not marginal. Liquid maldistribution is the single largest cause of lost packing efficiency: a badly distributed column can show a height equivalent to a theoretical plate 50 to 200% higher than the design value, which shows up as an outlet concentration above permit while every stamped drawing looks correct.

2. The Shell and Liner Set the Boundary Conditions

Everything inside an FRP scrubber lives inside a corrosion barrier, and the first rule of internals work is that the barrier is not a structural element and must not be used as one. The three-zone laminate model applies: a resin-rich inner surface typically 0.25 to 0.5 mm thick (0.010–0.020 in) reinforced with a chemical-resistant C-glass or synthetic veil at around 95% resin content; a corrosion barrier of chopped strand mat bringing the combined liner to not less than 100 mil (2.5 mm) with a glass content around 27% by weight; and behind that, the structural laminate carrying the design loads.

Two fabrication standards cover most FRP scrubber shells. ASTM D3299 covers filament-wound cylindrical tanks in polyester or vinyl ester for above-ground vertical installation at atmospheric pressure, with a pressure and vacuum scope of about ±14 inches of water column. ASTM D4097 is the contact-moulded equivalent. Where the vessel is designed as pressure equipment, ASME RTP-1 governs up to 15 psig, with ASME Section X beyond that.

Requirement	Typical specification	Source / standard
Corrosion barrier thickness	Not less than 100 mil (2.5 mm) combined; heavy-duty chemical service often 2.5–6.3 mm (100–250 mil)	ASTM D3299, ASTM D4097, ASME RTP-1
Resin-rich inner surface	0.25–0.5 mm veil layer at approximately 95% resin content	Industry practice
Reinforcement in the barrier	Chopped strand mat and surface veil only; continuous woven fabrics excluded from the barrier because they wick along fibre bundles	ASTM D3299 practice
Glass type for acid service	ECR-glass (boron-free); powder-bound mat preferred over emulsion-bound	Materials practice
Cure verification	Barcol hardness at least 90% of the resin manufacturer's specified minimum; ASTM D2583 method	ASTM D3299
Static / sustained allowable (RTP-1 regime)	Safety factor about 10:1, allowable stress of the order of 1,500 psi	ASME RTP-1
Transient allowable	Safety factor about 5:1, of the order of 3,000 psi	ASME RTP-1
Deflection limit under hydrostatic load	L/300 commonly applied to FRP panels and flat surfaces	ASME RTP-1 practice
Hydrostatic test	Typically 1.5 × MAWP for 30 minutes; 24 hours for atmospheric tanks	ASME RTP-1

Never drill, tap or through-bolt the corrosion barrier to fix an internal. A hole in the liner is a permanent, wetted path for the process liquor into the structural laminate, and it will open up by capillary action long before anyone notices it from the outside.

Instead, internals are supported either on a laminated support ring built up on the inside wall with the same resin system, or on a bolted ring using blind or through-shell fixings that are subsequently protected by a local overlay of the same corrosion barrier construction. ASME RTP-1 also defines visual acceptance levels for the inner layers — a "critically corrosion resistant" level capping allowable repairs at about 3% of the critical inner surface — which is the standard to quote when a fabricator proposes to patch a liner rather than rebuild it.

3. Packing Support Grids: Loads, Free Area and Deflection

The packing support is the internal with the least glamour and the highest consequence of failure. It carries the dry weight of the bed, the weight of liquid held up in the packing during operation, the weight of liquid that accumulates during irrigation and shutdown, and any load imposed by differential pressure across a fouled bed.

Sizing it correctly starts with an honest estimate of the wet bed weight. Random packing bulk densities for polypropylene packings are modest — a 50 mm nominal packing is typically in the range of 60 to 90 kg/m³ as dumped — but liquid holdup during operation can add a comparable mass, and a bed that is flooded at startup or during a wash cycle briefly weighs far more than the design case. Design practice is to take the packing weight plus holdup plus an allowance for fouling and liquid accumulation, typically in the range of 1.5 to 2 times the calculated operating wet weight, and to verify the support deflection at that load.

Design input	Practical figure or method	Why it matters
Support free area	60–70% minimum; 70–90% achievable with well-designed grid or bar supports	Low free area raises gas velocity through the support and adds pressure drop at the highest liquid loading point in the column
Design load	Wet packing weight plus holdup, times a factor of roughly 1.5–2.0	Covers fouling, liquid accumulation and upset
Deflection limit	L/300 as a working target for FRP support bars under sustained load	Deflection redistributes packing, creates void channels and drops bed efficiency
Support type	Gas-injection or grid-type support; bar grid with a coarse wire or FRP top layer under small packings	Small packings fall through large openings; a coarse layer prevents it
Attachment to shell	Laminated support ring, or bolted ring with a local corrosion barrier overlay	Point loads on an uninforced FRP wall are a crack initiation site
Pressure drop budget	Allow 0.1–0.25 kPa across a clean support, more once fouling begins	Fan power is a lifetime cost; a support that adds 1 kPa in a large column is a real operating expense

Free area deserves a specific comment because it is where material choice interacts with process behaviour. In an acid mist scrubber the gas velocity through a packed bed is typically kept between 0.5 and 1.5 m/s: below 0.5 m/s the inertial impaction that captures droplets becomes too weak, and above 1.5 m/s pressure drop rises sharply and captured liquid starts to be re-entrained from the packing surface. A support plate with 60% free area passing the same gas flow sees a local velocity some 1.7 times the bed velocity, which can push the support zone into the re-entrainment regime even though the bed itself is fine. That is one reason high-free-area supports are worth paying for.

4. Liquid Distributors: Drip Point Density and Uniformity

A liquid distributor has to do five things: deliver a uniform flow across the cross-section, provide enough drip points for the packing to redistribute properly, irrigate the wall region without over-wetting it, stay open enough for the vapour to pass, and avoid generating entrainment. Each of these has a number attached.

Drip point density is the first. For a nominal 25 mm or larger random packing, a density of about 60 points per square metre (5.5 points per square foot) is the baseline; high-performance random and structured packings want progressively more, up to 130 to 215 points per square metre for the highest-surface-area structured packings. The reason is that structured packings have poor lateral spreading of liquid, so whatever the distributor fails to place at the top of the bed stays where it landed all the way down, while random packing self-distributes to some degree.

Distributor type	Irrigation density range	Typical turndown	Where it is used
Closed pipe / perforated pipe distributor	8–30 m³/m²·h (3–12 gpm/ft²), up to about 60 m³/m²·h in high-rate designs	about 1.7:1	Small to medium columns, clean service
Pan or orifice tray distributor	12.5–80 m³/m²·h (5–33 gpm/ft²)	about 2:1	Random packing, medium columns; limited turndown with single-level orifices
Trough distributor with downcomers	12.5–50 m³/m²·h (5–20 gpm/ft²), higher with multiple hole rows	2:1 to 2.5:1, up to 4:1 with triangular slots	Large columns, fouling service, structured packing
Spray nozzle header	3–180 m³/m²·h (1.2–74 gpm/ft²)	about 1.5:1	Spray towers, quench sections, top spray in packed absorbers

Uniformity is the second number. A useful acceptance metric is the standard deviation of flow through the drip points divided by the mean flow, sometimes written as a coefficient of variation: for mass-transfer duties it should be held to roughly ±5%, and the tolerance loosens for pure heat transfer or washing duties. Uniformity is unmeasurable after the column is closed, so it has to be verified at the factory — a distributor delivered as a set of parts and assembled on site has no uniformity guarantee at all.

Geometry follows from hydraulics. A typical distributor occupies about 33% of the cross-section, which means the vapour velocity in the riser area is around three times the velocity leaving the top of the packed bed; the higher the column operates relative to flood, the more care is needed so that liquid streams do not break up into droplets and get carried up the risers. Distributors are normally set about 150 mm (6 in) above the top of the bed, and a spray catcher or anti-entrainment section of at least 250 mm in height, with the downcomer ends buried in 38 to 50 mm saddles, significantly reduces mist generation above the distributor — a detail that is cheap at the design stage and almost impossible to add later.

Turndown is the third constraint. Distributor flow is proportional to the square root of the liquid head, so as the plant turns down, the head falls and both the levelness tolerance and any gradient across the distributor become a larger fraction of the driving head. Single-level orifices hold uniformity over roughly 60 to 120% of design flow; below that, multi-level orifices or slotted weirs are needed, at the cost of slightly worse uniformity across the whole range. Spray nozzle headers tolerate turndown less well than trough designs, at around 1.5:1.

5. Spray Headers and Nozzle Selection

In spray towers, quench sections and the top spray of a packed absorber, coverage replaces drip point density as the governing parameter. The design question is whether the cone from each nozzle reaches its neighbour's cone with enough overlap to leave no dry patch, and the geometric answer is straightforward: for a single central nozzle, the required half-angle is the arctangent of the vessel radius divided by the nozzle mounting height, and the next wider standard angle with margin is selected. For a multi-nozzle header the better geometry is usually a narrower spray angle — 60° to 80° — at closer nozzle spacing, which produces a more even distribution than wide-angle nozzles widely spaced.

• Pattern: hollow cone is the dominant pattern in wet scrubbers, because it maximises droplet surface area per unit of liquid volume and therefore the number of pollutant-droplet collisions.
• Free passage: in recirculating systems, use a spiral-body hollow cone nozzle with no internal vane. Recirculated liquor concentrates scale-forming compounds and particulates, and any vane or narrow passage will plug.
• Pressure and flow: nozzle flow scales with the square root of pressure, so a header running below design pressure loses both flow and pattern. Confirm the flow-pressure curve against the pump curve at the actual operating point rather than at rating.
• Strainers and blowdown: fit coarse strainers upstream of the header, and manage blowdown so that liquor conductivity stays within limits. Scaling load on nozzles is a chemistry problem, not a nozzle problem.
• Header velocity: keep branch velocities in the range of about 1.5 to 3 m/s to limit erosion and pressure loss, and provide a removable end cap or flush connection for each branch.

Liquid-to-gas ratio is set by the duty, and the two cases are not interchangeable. For acid gas absorption in a packed bed, an L/G of roughly 0.5 to 2.0 L/m³ is typical; for acid mist removal the requirement is lower, around 0.3 to 1.0 L/m³, because the liquid film's job is to wash captured acid off the packing rather than to drive mass transfer. Specifying a mist-removal L/G on an absorption column under-wets the packing and leaves dry, channelled sections.

6. Mist Eliminators: Vane, Mesh and Fibre Bed

The mist eliminator determines whether the mass of pollutant the scrubber captured stays in the vessel or leaves through the stack as fine droplets. Selection is driven by the droplet size distribution at the outlet, which in turn depends on the scrubber type: a spray tower produces a coarser outlet distribution because its own large droplets are incompletely captured, while a packed bed produces finer mist because the droplets that survive the packing are already small.

Type	Design gas velocity	Typical cut size / efficiency	Pressure drop	Fouling tolerance
Chevron / vane bank	2–6 m/s depending on vane geometry and pocket design	About 20–25 µm at 99%; special designs lower	0.15–0.40 kPa clean	High; wide blade spacing, typically 35 mm, resists plugging and allows flush sprays
Composite mesh pad	1–3 m/s	Above about 1 µm at 99%+	Higher than vane at equal duty	Moderate; prone to plugging with sticky or solid-laden droplets
Fibre bed	0.1–0.5 m/s	Sub-micron droplets	Highest	Low; used for polishing where visible emission limits are tight
Structured or packed bed eliminator	1–3 m/s	10 µm and above at 99%+	Very low	Good, but requires support and irrigation

The velocity limit is not arbitrary; it comes from the balance between gravity, drag and surface tension on the captured droplets. The classical sizing relationship is the Souder-Brown equation, in which the maximum allowable vapour velocity is a constant K multiplied by the square root of the density ratio:

U = K × √[(ρ_L − ρ_G) / ρ_G]

Published K values for vane eliminators are of the order of 0.12 m/s for vertical flow with horizontally placed vanes, 0.18 m/s for horizontal flow with vertically placed vanes, and up to 0.26 m/s for double-pocket designs. Above the resulting velocity, captured liquid films are torn off the vanes and re-entrained, and the mist eliminator starts exporting droplets instead of removing them — the scrubber's removal efficiency appears to fall off a cliff after a production increase, and the cause is usually a demister running above its design velocity rather than a failed packing.

Two installation details decide whether a correctly sized eliminator actually performs. First, it must be installed without gaps at the perimeter: any bypass path around the vane bank is an unobstructed route for droplets straight to the stack, and modular units must be sealed to each other and to the shell. Second, the drainage path must be real — vane pockets must drain by gravity without the collected liquid sitting in the pocket, or the pocket fills and the vane behaves like a wetted wall.

7. Wash Systems and the Wet-Dry Cycle

Vane and mesh eliminators in continuous service accumulate solids. A wash system is therefore part of the design, not an option, and it should be specified with the rest of the vessel rather than retrofitted during the first fouling outage.

The workable arrangement for a vane bank with 35 mm blade spacing is a header on the upstream or downstream face with flush sprays directed at the blade inlet, sized to wet the surface and flush the drainage pockets without driving liquid through the bank. Washes are usually intermittent: a short wash at a defined frequency maintains the eliminator, while continuous washing costs pump energy and can itself generate entrainment. Where a distributed control system is available, drive the wash from a pressure drop or fan power signal so that washing happens when the bank is actually loading up. For mesh pads, wash intensity has to be lower and more carefully distributed, because high-velocity jets can damage the pad structure. For fibre beds, washing is usually counter-current and at low velocity, and the bed is often replaced rather than washed.

The wet-dry cycle is the other maintenance trap in the top of an FRP vessel. On shutdown, liquid retained in pockets, on packing surfaces and in the wash header drains slowly, and any residual acid or chloride solution concentrates as it evaporates. That concentrated liquor sits against the liner for days or weeks at a much higher concentration than the operating condition, which is why a shutdown inspection that finds liner spidering or discolouration near the top of the vessel is often looking at a wet-dry effect rather than a process chemistry failure. Draining the wash header, purging the top spray line and rinsing the demister before a long shutdown cost almost nothing and prevent a repair that involves relining the top of the shell.

8. Materials for Internals: Thermoplastics, FRP and Temperature Ceilings

Internals are rarely made from the same material as the shell. Thermoplastics — polypropylene, PVC, PVDF and PTFE — offer better chemical and temperature flexibility in small parts such as nozzles and packing elements, and can be field-repaired by thermal welding, while FRP is used for structural grids, support rings and large vane banks where stiffness and lamination thickness matter more than mouldability. The specification has to match the material to the chemistry and to the temperature, because every internal in the vessel is at the top of the temperature profile, not the middle.

Material	Practical continuous temperature ceiling	Typical chemistry fit	Notes
PVC	About 60 °C	Dilute acids	Cheapest nozzle and header material; drops off quickly with temperature
Polypropylene	80–90 °C depending on grade	Broad pH range, caustic scrubbing, most acids	Standard for scrubber nozzles and packing; not for strong oxidisers at temperature
PVDF	120–140 °C	Halogens, oxidising acids, chlorinated and fluorinated service	The default upgrade for chlorine, hypochlorite and nitric acid duties
PTFE	About 250 °C	Effectively all chemistry	Used where nothing else survives; higher cost, lower mechanical stiffness
FRP — polyester matrix	HDT around 90–100 °C	Mild chemical service	Not for strong acid or alkaline splash at temperature
FRP — vinyl ester matrix	HDT around 105–120 °C	Acids, oxidisers, chlorides	The practical default for scrubber shells and structural internals
FRP — phenolic matrix	HDT around 140–160 °C	Fire-sensitive duties	Specified where fire performance is mandatory, with a cost premium

Two cross-checks apply to any material table. First, continuous service temperature must be compared with the glass transition temperature, not with the melting point: design rules for pultruded and laminated FRP commonly limit service temperature to Tg minus 20 to 22 °C, which is why a vinyl ester vessel rated at a Tg of 105 °C is a 85 °C vessel in practice. Second, mechanical properties fall as temperature rises, so a support grid that passes its deflection check at 20 °C may not pass it at 80 °C; the right approach is to check the deflection at the maximum operating temperature with the hot modulus.

9. Support Rings, Fasteners and Differential Thermal Expansion

Bolted internal supports in an FRP vessel have to resolve a thermal expansion mismatch between the shell and the internal. Pultruded FRP has a transverse expansion of the order of 20 to 30 × 10⁻⁶/°C, while polypropylene expands by more than ten times that figure and PVDF by roughly eight times. Over a 750 mm diameter vessel with an 80 °C temperature rise, that difference becomes several millimetres of radial movement at the support ring, and a thermoplastic grid bolted through fixed holes will either crack or buckle unless the connection is designed to let it move.

• Fasteners: use 316 stainless steel for all internal fasteners, or a thermoplastic fastener where the chemistry attacks stainless steel. Never mix carbon steel into an FRP or thermoplastic assembly.
• Holes: oversize or slot the holes in the member that has the higher expansion, and use a washer or backing plate under every nut so the clamping load is spread over a large area of laminate.
• Gaskets: use PTFE or a chemistry-matched elastomer at the interface, and avoid compressing a soft gasket into a hard-stop arrangement where thermal cycling will crush it.
• Liner protection: fit a protection plate, angle or a strip of the same resin laminate at every point where an internal touches the liner, because abrasion of the corrosion barrier by a vibrating grid is a slow but certain failure.
• Lifting and installation: never lift an internal on a sling bearing against the liner, and never let a packing support or vane bank slide down the wall on its edge.

10. Bed Sizing, Redistributors and Turndown

The internals layout is constrained by how tall a packed bed can grow before liquid maldistribution accumulates. Working limits are a bed height of not more than twice the column diameter, or 6 m (20 ft), whichever is less, for random packing, and one and a half times the diameter, or 9 m (30 ft), for structured packing. Beyond those limits a liquid collector and redistributor is required, and each one adds roughly 0.6 to 0.9 m of column height.

The practical sizing sequence for a scrubber or absorber column runs as follows.

1. Fix the gas flow and the target removal efficiency, then select the packing type and the design pressure drop — typically 0.02 to 0.05 kPa per metre for random packing and about half that for structured packing in clean service.
2. Size the diameter at 70 to 80% of flood, which leaves margin for turndown, fouling and future capacity increase. Columns designed much above 85% of flood tend to flood during upsets rather than degrade gracefully.
3. Calculate the required bed height from the mass-transfer requirement, then split it into beds that respect the height limits and insert redistributors.
4. Convert the liquid flow into an irrigation density in m³/m²·h and select a distributor type from the table above that covers the range with the required turndown.
5. Verify the drip point density against the packing selected, not against the packing originally assumed at quotation stage — a packing substitution is exactly where distributor mismatch is introduced.
6. Size the mist eliminator on the outlet droplet distribution and the Souder-Brown velocity limit, then size the wash system for the expected solids loading.
7. Check the pressure drop budget across support, beds, distributor and eliminator against the fan curve, with a fouling allowance.

11. Commissioning, Shutdown Inspection and a Specification Checklist

Most internals problems are detectable before the vessel is closed, and the checks are simple. Distributor troughs and pans must be level to within a few millimetres across the whole cross-section, because the flow through each orifice depends on the square root of the liquid head above it. Spray patterns must be visually verified with water at operating flow before the manway is bolted, with the overlap photographed and the header pressure recorded. Support grids must be checked for level, for the correct free area, and for the correct coarse layer under small packings. The corrosion barrier must be spark-tested where local repairs were made, and Barcol hardness readings taken on the repaired areas. The eliminator must be verified for perimeter sealing and for free drainage.

For procurement, the following items belong in the specification rather than in a follow-up email.

• Duty definition: gas flow and turndown range, inlet and outlet concentration, temperature profile, droplet size distribution at the eliminator, and the emission limit being targeted.
• Materials of construction for each internal: not "PP or equivalent", but the specific material and grade for nozzles, headers, packing, support grid, eliminator vanes, fasteners and gaskets, with the continuous temperature ceiling stated for each.
• Hydraulic data: irrigation density, drip point density, spray angle and overlap, nozzle flow-pressure curves, and the header velocity at design flow.
• Structural data: support grid design load and deflection limit, support ring attachment method, and the corrosion barrier overlay detail at every penetration.
• Verification requirements: factory levelness and uniformity check for distributors, spark test and Barcol results for liner repairs, hydrostatic test per ASME RTP-1 (typically 1.5 × MAWP for 30 minutes, or 24 hours for atmospheric vessels), and a documented packing installation record.
• Spares and access: manway size adequate to remove the largest internal, plus spare nozzles of the correct pattern, angle and orifice size in the same material.

The recurring theme is that the internals are what the vessel is for, and the FRP shell is the containment that makes them possible. Specifying the shell to ASTM D3299 or ASME RTP-1 and then choosing the internals from a catalogue is how plants end up with clean liners, correct wall thicknesses and an outlet that has been out of permit for two years. Our filament wound FRP chemical storage and mixing tanks, antistatic FRP ventilation duct and scrubber ductwork and filament wound FRP process pipe are built with vinyl ester corrosion barriers for this class of service, and we supply liner construction details, resin and cure documentation with the quotation. For the surrounding design decisions, the dual-laminate and thermoplastic liner guide covers PP, PVC and PVDF lining selection, the gel coat and corrosion barrier design guide explains the liner layer stack, the nozzle and manway design guide covers penetration reinforcement and bolt torque, and the duct and hood fabrication guide deals with the stiffening and support of large FRP surfaces.