Calming Baffles and Perforated Plates: Flow Distribution Design Rules
11/09/2026Separator Problems That Start Outside the Vessel: Drain Seals, Supports, Transients and Properties
11/15/2026Most selection problems on separator internals are not sizing problems. They are technology problems: the wrong device was chosen for the duty, and then sized correctly for the wrong job. This page is the map we use internally when a customer sends a process datasheet and asks which technology we would put in the vessel.
The four technologies, and what each actually removes
| Technology | Droplet cut point | Typical pressure drop | Fouling tolerance | Where it fails |
|---|---|---|---|---|
| Wire mesh pad | about 10 microns, lower with finer mesh and deeper bed | 0.1 to 0.25 kPa | Low to moderate | Blinding by solids, blow-out on a wide support grid, flooding at high liquid load |
| Vane pack / chevron | 20 to 30 microns | 0.05 to 0.15 kPa | High | Fine droplets pass straight through, poor drainage re-entrains what it caught |
| Fiberbed element | sub-micron aerosol | 0.5 to 2 kPa | Low | Media collapse, condensation inside the bed, blower cannot supply the pressure drop |
| Coalescer / plate pack | separates two liquid phases, not droplets from gas | low, set by the settler not the media | Moderate | Emulsion stabilised by chemistry, interface held at the wrong level, solids at the interface |
The first decision is therefore not “how big”, it is “what am I removing”. Droplets from a gas stream and a dispersed phase in a liquid are different problems that happen to share a vessel.
Decision sequence: five questions in order
We ask these in this order, because each answer eliminates technologies rather than narrowing sizes.
- Is the continuous phase gas or liquid? Gas duties go to mesh, vane or fiberbed. Liquid-liquid duties go to coalescer media or a plate pack, and the settler around them matters more than the media.
- What droplet size has to be removed? Above about 30 microns a vane pack is the cheapest answer and tolerates fouling. Down to about 10 microns, mesh. Below a few microns, or where the carryover is aerosol rather than droplets, only a fiberbed element will do it.
- How much solids or scale will the stream deposit? Vane packs tolerate deposits because the passages are large and drain downwards. Mesh blinds progressively. A fiberbed element is the least tolerant of all, because the deposits sit inside the fibre where they cannot be washed out.
- How much pressure drop can the system pay for? A vane pack costs almost nothing in fan duty, mesh costs a little, a fiberbed element can cost more than the fan can supply. Check the blower curve before choosing the finest technology, not after.
- Is the vessel new or existing? In an existing vessel the available height, the manway and the shell roundness decide more than the nominal duty. A technology that needs depth you do not have is not an option, however well it fits the process.
Cross-technology comparison at one duty
The same duty, sized three ways: 2.62 m3/s of gas at 1.2 kg/m3 with liquid droplets at 150 kg per hour, requiring essentially dry gas at outlet.
| Technology | Velocity used | Required area | Vessel diameter | Pressure drop estimate |
|---|---|---|---|---|
| Mesh pad, K = 0.107 m/s | 3.26 m/s | 0.80 m2 | about 1.21 m | about 0.15 kPa |
| Vane pack, K = 0.152 to 0.183 m/s | 4.62 to 5.56 m/s | 0.67 to 0.81 m2 | about 0.93 to 1.02 m | about 0.08 kPa |
| Fiberbed, face velocity 0.3 m/s | 0.30 m/s | 8.73 m2 | about 3.33 m | 0.8 to 1.2 kPa |
The fiberbed needs roughly seven times the cross section of a vane pack for the same gas flow. That is the price of sub-micron capture, and it is why fiberbed elements are used where the outlet specification genuinely demands it rather than as a default upgrade.
Where each technology is the wrong answer
Mesh pad, when the liquid load is high or the gas is dirty
A mesh pad holds liquid inside the bed by design. At high liquid load it floods, and beyond flooding it re-entrains more than it removes. On a dirty gas it blinds, and the pressure drop rises until the fan is the limit. Both failure modes look like “the pad stopped working” but neither is solved by more mesh.
Vane pack, when the specification is written in microns
A vane pack is a momentum device. Droplets below about 20 microns follow the gas around the turn and are never captured, whatever the face area. If the outlet requirement is expressed as a droplet size below that, a vane pack cannot meet it no matter how it is sized.
Fiberbed, when the blower has no margin
The element is the only technology that removes sub-micron aerosol, and it does so at a pressure drop the rest of the system has to supply. Where the blower is already at its curve, adding a fiberbed element moves the operating point and the flow falls. That is a system problem presented as an element problem.
Coalescer, when the emulsion is chemically stabilised
Coalescer media gives dispersed droplets a surface to attach to and grow on. It cannot defeat a surfactant or a solid that holds the interface stable. In that case more coalescer area produces a thicker emulsion band, not a cleaner outlet, and the fix is upstream in mixing or chemistry.
What to send us to get a specific answer
Flow, pressure, temperature, density of both phases, the droplet or solids loading, the required outlet, the vessel dimensions including the manway, and the material. With those we can state which of the four technologies applies and what the element would look like. Without the vessel dimensions we can still size the technology but not guarantee the fit.
The families in detail
- How a wire mesh demister works – capture mechanisms, mesh density and bed depth, and where mesh stops being the right answer.
- Vane packs and chevron mist eliminators – the momentum family, passage geometry and drainage.
- Fiberbed versus wire mesh – the comparison that decides the fine-droplet duties.
- Entrainment control in LNG kettle reboilers – where the vessel diameter comes from when a mist eliminator, not the heat duty, sets the limit.
- Entrainment prediction and why a reboiler element cannot be sized freely – the correlation chain, the shell-chord limit, and the role of CFD.
- Western versus Chinese sizing conventions – coefficient, pad height and margin, and the 47% diameter gap.
- Why the same duty gives different diameters – the five conventions that move a quote by 40%.
- Reboiler eliminators beyond LNG – ethylene fouling, heat-pump approach, refrigeration suction conditions.
- Why a shorter bundle can need a larger shell – which area binds per stage, and the transport limit on diameter.
- Separator problems that start outside the vessel – drain seals, supports, transients and datasheet properties, the checks to run before replacing an element.
- Coalescer cartridges and cross-reference – the liquid-liquid family, replacement and equivalence.
If the duty is a replacement inside an existing vessel, the measured dimensions matter more than the nominal specification: see the process data we need and how a cross-reference is built.
Frequently Asked Questions
Which mist eliminator technology should be used for droplets below 10 microns?
When is a vane pack better than a mesh pad?
Why does a fiberbed element need such a large vessel?
How do I know whether my duty is a gas-liquid or a liquid-liquid problem?
What is the first thing to check before choosing a technology?
- Gas-Liquid Separation Fundamentals — terminology, separation mechanisms and internal selection
- Technical Data Sheet — material and performance reference data
- All Technical Articles — complete engineering index
If you are working on a replacement
Coalescer media and plate packs are usually replaced as a set with their support frame, so the interface dimensions are fixed before the media grade is chosen. See the process data needed to size and select it, how fabrication and inspection run and what can and cannot be compressed on an urgent replacement.
