Understanding Droplet Formation
Filtration and Separation Basics by JCI
The selection of filtration and separation technology to be used in an application depends on how the entrained droplets were formed. Understanding this concept can ensure that the correct equipment is employed in any application.
Droplet distributions follow a Gaussian probability curve which is centered around a peak representing the average size. The important characteristics of the distribution, both the average or mean size and the standard deviation, are highly dependent upon the manner in which the droplets were formed, as well as the process conditions including pressure and temperature, and the characteristics of the fluids.
Consider a lube oil that is introduced into a compressor to protect its internal components. One mechanism for liquid droplet formation is that the lube oil will be vaporized by frictional heat generated by the moving parts, resulting in vaporized lube oil becoming mixed with an air or natural gas stream, and then condensed back into liquid form as the stream cools in the downstream equipment and piping. Entrained liquids arising through this condensation are typically very fine, and usually display a distribution curve with a mean droplet size of less than one micron.
Another common source of entrained liquids is through mechanical agitation, such as is commonly found when phases violently mix in mass transfer equipment. Mechanical droplet generation results in droplet distributions peaking at greater than 20 microns, and often well above 100 µm.
Also significant is that droplet distributions are not stagnant but change with time (maturing), or with variations of pressure and temperature. This is especially true of pipeline applications.
Figure 1 depicts three different liquid droplet distributions in a single vapor stream, measured from actual field sampling. This graph reflects that three different mechanisms were involved in bringing about the entrainment of liquids.
The experienced designer will be able to estimate the mean of the droplet size distribution based on the application or the means by which the droplets were formed. The device used to separate a given droplet distribution must be selected on the basis of that distribution. Generally speaking, the smaller the average droplet, the smaller or tighter the collection media required.