| General Considerations
Many gas compressors in use today employ direct lubrication where the
oil is admitted to the compressor and thence to the compressed gas. However, in may
instances, the presence of this oil is detrimental to subsequent operations and the oil
must be removed.
The oil is present in two states. Some of it appears as a fine mist or
as small droplets. Oil in this form can best be removed by mechanical means. Any number of
types of filters are employed from cotton to specially shaped porous ceramic or sintered
metal filters. These filters, however, are unable to remove the oil in the vapor state,
and this is where activated carbon can be used to an advantage.
There is very little data concerning the most important aspects of this
application. It is very difficult to measure how much oil is in the compressed gas. Most
oil removal measurements are based upon the amount a filter picks up in an extended test.
Data obtained by this method indicate oil concentrations from 5 to 20 vppm. A more
accurate figure may be obtained if the amount of oil added to the compressor is known. A
conservative estimate for design purposes is 15 ppm by weight. This is on the order of 1.3
lbs. of oil per million ft.; of gas at STP.
The amount of oil removed by the mechanical filter must be known to
size the carbon section. In tests conducted by the Mellon Institute, the ratio of
mechanical to adsorptive removal ranged from 260:1 for a highly effective ceramic filter
to 1.3:1 for a cotton pad.
The capacity of activated carbon for oil vapor adsorption would
ordinarily be very high. However, the concentration of oil in most cases is extremely low
especially if an efficient mechanical filter is used upstream of the carbon bed, resulting
in a low concentration of the oil entering the carbon bed. Tests have shown that a
capacity of approximately 3% by wt. is obtained in an extended test with values ranging
from 2% at the outlet end of a deep bed to about 25% at the inlet end where oil leakage
into the carbon bed from the mechanical filler (cotton pad) was experienced.
The difficulty in detecting oil in the compressed gas before a
filtering operation indicates a magnification of the problem in measuring the effluent oil
concentration from a carbon bed. The only method of determining the efficiency of the
filter is by careful observation of the behavior of the cleaned gas in subsequent
operations. Detailed calculation and common sense sometimes indicate that a filter just
cannot be doing an efficient job and yet the small amount of oil that is removed seems to
make the difference between acceptable and non-acceptable gas.
Commercially Available Adsorbers
There are any number of commercially available filters employing
activated carbon designed mainly for CO2 and compressed air. They typically
employ three filters in series, the second one containing activated carbon. The oil
adsorbers are also used in soft drink bottling plants to purify CO2. The
recommended NUCON grade is NUSORB GC60-3.
Design Considerations
The same design considerations generally apply to the removal of oil
from compressed gas as in any other adsorption system although there will be some slight
differences.
Concentration of oil in compressed gas:
Assume 1.3 lbs. per million ft.3 at STP unless there is other data.
Percentage of oil removed by the mechanical filter:
Assume 60% unless data indicates more.
Adsorptive capacity of carbon for oil vapor:
Assume 3 wt.% capacity at temperatures below 100EF and down to .5 wt.%
at temperature as high as 300EF.
Cycle Time:
Most units are designed to operated for more than 6 months. If the
quantity of carbon involved is small, it is discarded. If it is large, it can be returned
to facilities licensed for reactivation, which will cost approximately $.75/lb. If
conditions are such that the carbon capacity is quite low, regeneration in place can be
considered. Either steam or hot gas can be used. Temperatures should approach 600EF for
complete removal. Care must be taken not to admit gasses containing oxygen to the bed at
these high temperatures, i.e. the carbon is to be cooled before returning to air service.
Dimensions of the carbon bed:
Capacity considerations generally fix the volume of the carbon bed.
Maximum flow rates are pressures dependent. If these rates are approached, the bed depth
should not be less than 4 ft. Eighteen inches should be the minimum depth for any flow
rate. Allowable pressure drop will sometimes determine the length of the bed and,
therefore, the diameter. In some cases, vessels are already available which also may fix
the carbon volume used.
Grades to recommend:
If the velocity is high but the pressure drop is not important,
NUSORB G60 6x12 is recommended. When pressure drop is a problem, use NUSORB GC60-3.
Sample Calculation: |