Perspectives: Plant InSites
Effectively Remove Droplets
Various factors affect the choice and operation of liquid/vapor separators.
Plants often use separators to remove liquid droplets from vapor streams. To evaluate the gas capacity of these vessels, engineers typically rely on the empirical Souders-Brown correlation:
V = K [(ρL – ρV)/ρV]0.5
where V is the superficial vapor velocity, K is a constant, ρL is the density of the liquid phase and ρV is the density of the vapor phase. The same equation frequently is used to evaluate packing in distillation, stripping and absorption towers. As a concept this makes sense. Tower packing comes into contact with rising vapor and falling liquid. If the liquid can’t fall, the tower will flood. For distillation towers, the limiting value for K often is given as 0.45 for systems requiring low entrainment and up to 0.50 for systems that can tolerate high entrainment. These values are for high-capacity packings that can operate close to the system limit. However, a specific packing may have a much lower capacity.
The Souders-Brown correlation doesn’t account for any effect of separator height. For towers, that isn’t a major issue because most installations have spacings between internal elements of three feet or less.
Separators, however, often have much more vertical space. This allows for higher capacities in a given plot area. With V in ft/sec and density in consistent units, experience-based values of K for separators are:
• 0.167 for 5-ft-high vertical vessels with the stream entering in the middle of the vessel (values for specific services run from 0.12 to 0.24);
• 0.210 for 10-ft-high vertical vessels with entrance in the middle of the vessel (a range of 0.18 to 0.35 for specific services); and
• 0.35 for spherical vessels and vessels with wire-mesh mist eliminators (a variation of 0.22 to 0.39 for specific services).
For horizontal vessels, you can estimate K via:
K = 0.45 (L/10)0.56
where L is the horizontal vessel length in ft. For such vessels, V is calculated based on the open area above the liquid level. This effectively relates the horizontal travel required for the vapor to give enough time for a droplet to settle horizontally.
Vertical separators are best in applications where gas doesn’t contain much liquid. Horizontal separators are more effective for dealing with large liquid volumes. They also better suit three-phase systems if the two liquid phases are being separated at the same time.
The Souders-Brown correlation and K-factor selected imply a minimum droplet size. Here, the situation becomes a bit unclear. Stokes’ Law, Newton’s Law or the Intermediate Law govern droplet settling. Smaller droplets have lower settling velocities; larger droplets have higher ones. Without special internals (e.g., baffles, impingement plates, cyclones or mesh pads), the maximum size of a droplet that will settle depends upon its terminal velocity. If this is less than the velocity of the rising vapor, the droplet will fall out of the vapor. If the terminal velocity exceeds the velocity of the rising vapor, the droplet will ascend with the vapor.
The upper limit for Stokes’ Law is a droplet Reynolds number of 2. Newton’s Law pertains when Reynolds numbers lie in the 500–200,000 range. Between 2 and 500 — the situation for most industrial systems — the Intermediate Law applies. Droplet removal efficiency is a complex function of droplet size, liquid density, vapor density and vapor viscosity. The effective removal limit for vertical separators varies from a droplet size of 40 μ to 170 μ depending upon the system, vapor velocities and liquid load.
Wire-mesh mist eliminators, by providing both inertial and impaction droplet-removal mechanisms, can push limits down into the 3–10-μ range. However, quoted removal efficiencies of 99–99.5% for such droplets often reflect best-case operation. Contamination of the mesh pad with scale, rust or other solids may dramatically reduce this. Corrosion damage to the pad is another major concern.
Mesh pads have a relatively sharp upper limit on vapor velocity. Operation may alter dramatically with as little as 10% change in vapor rate as the pad approaches its upper operating limit.
Finally, mesh pads create liquid/vapor surface area. Much like with packing, this may promote liquid/vapor mass transfer. In fact, mesh pads often serve as the packing in small diameter columns. Liquid composition in the mesh pad may not be as expected. This can lead to fouled pads.
Like all equipment, separators without and with mesh pads have a specific operating envelope for best performance. Running outside that envelope will likely lower efficiencies. Respect their limits and both should work well.
ANDREW SLOLEY, Contributing Editor
Asloley@putman.net
More from this perspective...
| Title | Date |
|---|---|
|
More Exchanger Area Can Pose Pitfalls This "obvious solution" for heat-transfer problems may not work. |
02/01/2012 |
|
Avoid flow problems by considering the impact of inlet losses. |
05/20/2009 |
|
Odd Layout May Make Even More Sense Unusual tower configuration better copes with control valve failure. |
01/23/2013 |
|
Some equations have stood the test of time. |
04/04/2012 |
|
Pick the Proper Centrifugal Pump Consider the impact of startup conditions and prospective operating points. |
07/11/2011 |
|
Piping Geometry: Grasp Line Layout Draining, venting, fouling and other considerations affect piping geometry. |
09/06/2012 |
|
Plant InSites: Cast a cold eye on columns Subcooling can complicate and compromise tower performance |
05/23/2008 |
|
Plant InSites: Come Up With a Solid Estimate Use geometry to figure out the amount of material in storage or equipment. |
12/14/2009 |
|
Plant InSites: Dont Make Problems for Your Plant Local fabrication may seem attractive but can degrade performance. |
10/22/2009 |
|
Plant InSites: Look Beyond the Lore Dont rely on recollections about how a unit had performed |
10/10/2008 |
|
Preheater points out the value of cooling off Lack of attention to detail in conceptual design can hide many sins. Put the problem aside and then ponder it again. A fresh look may lead to an even better solution. |
05/12/2004 |
|
Process Engineering: Inspect with your mind, not just your eyes A thorough inspection of plant equipment following tiny installations can prevent a lot of future problems. Go beyond just checking for compliance with drawings when looking at hardware and physically look at the equipment itself. |
03/11/2005 |
|
Process Engineering: Poor compressor design puts pressure on pumps When a poor compressor design put too much pressure on the pumps, a new solution had to be constructed. Sometimes the ideal solution is just an intricate compromise. |
01/18/2005 |
|
Properly Assess Energy Recovery Projects Impact on other operations and transfer prices may alter the economics. |
03/09/2011 |
|
Properly Position Interlock Valves The obvious location may not be the most cost effective |
10/16/2012 |
|
Properly Trigger Standby Pumps A pressure signal works for many but not all situations. |
05/06/2010 |
|
Properly protect centrifugal pumps Consider various factors when selecting how to guard against low flow, Andrew Sloley says in this month's Plant Insites column. |
07/11/2007 |
|
Rethink Reactor Temperature Control Cascade strategy offers simplicity and fast response. |
07/13/2010 |
|
Right Vacuum Control Choice Takes the Pressure Off Choosing the proper recycle stream for regulation is crucial |
08/07/2008 |
|
Save a bundle solving pressure-drop problems Pressure drop in compressor-suction and interstage coolers often creates problems. In some cases, just a few pounds of extra pressure drop make a revamp unworkable. |
11/04/2004 |


