| Overview
Of
all the factors which need to be considered in humidity measurement,
one of the most important, and that which is often given the least
attention is the sampling system. Considerations of leakage, pressure
and temperature gradients, and moisture absorption characteristics
are often overlooked.
The
problem of leakage is relative; i.e., if the dew point being measured
is close to the ambient room dew point, leakage into the system
may not bias the reading substantially. If however, the system is
pressurized above atmospheric so as to create a leakage out of,
rather than into, the system, the error introduced will be less.
The degree to which leakage can be tolerated also depends heavily
on the actual dew point being measured. As an example, when measuring
a dew point of -100°F (-73°C) with a sample flow rate of
4 SCFH, at an ambient or surrounding dew point of 50°F (10°C),
a leakage in flow of 5x10-5 SCFH will cause an error of 1°F
(-17°C). However, at a measured dew point of 100°F (38°C),
the same leakage rate would cause an error of only 0.00001°F.
The area of leakage becomes significantly more important and the
error becomes much larger in systems operating below ambient pressure.
Pre-Heating
If the dew point of the gas under measurement is greater than the
ambient temperature of the installation and the sampling lines,
both the lines and the sensor must be heated with some type of heater
tape, or the line must be steam-traced in the usual fashion. The
approach used will vary widely with the specific nature of the installation,
and the user must use their own ingenuity to assure that none of
the sampling components be at a temperature lower than the highest
dew point anticipated. If electrical heater lines are used, it is
desirable to connect these to a variable transformer or controller
to adjust the heating level. If the sample lines are too long, it
may be necessary to wrap them in insulation to minimize the amount
of heat required to accomplish the preheating. The line should be
heated well above the dew point and should not exceed the temperature
rating of the sensor. A maximum of 200°F (93°C) is usually
recommended. Heating above the dew point does not change the dew
point of the sample.
Selection
of Sampling Components
Of equal importance is the effect that material absorption/desorption
characteristics have on overall system response. Although not true
of all applications, stainless steel, glass and nickel alloy tubing
are the best possible non-hygroscopic materials and should be used
for low dew point applications (0 to -100°F [-18 to -73°C]
and below). Teflon is also satisfactory, but begins reducing system
response due to desorption at the lower dew points. Copper and aluminum
alloys, as well as stabilized polypropylene tubing are acceptable
above 20°F (-7°C) dew point. Most plastic and rubber tubing
is unacceptable in all ranges. Unless attacked by the sample, the
effect of the more hygroscopic materials is not of contaminating
nature, but actually one of introducing severe lag into the system
during the establishment of an equilibrium condition. For example,
plastics such as nylon cannot be used at low dew points simply because
the equilibrium condition may actually take days to stabilize. The
actual selection of the sample line material should be based on
the degree of permanency of the installation, with a minimum of
joints, fittings, and other plumbing prior to the hygrometer. Generally,
stainless steel is preferred for permanent installations operating
at low dew points. On stainless steel lines, either swage or flare-type
fittings can be used.
Pressure
Measurements
The dew point temperature of gas is a measure of the absolute moisture
content of the gas, regardless of the temperature and pressure of
the gas. Most conversion tables for dew point (or frost point),
to parts-per-million, grains-per-pound, etc., are normalized at
atmospheric pressure (14.7 psia); therefore, if accurate absolute
moisture content measurements are to be converted to atmospheric-pressure-referenced
values, the pressure must be known. A pressure tap after the hygrometer
sensor can be fitted with an appropriate pressure gauge. Basic Humidity
Definitions are explained in a
separate EdgeTech document.
Material
Moisture Properties
All materials will absorb moisture to some extent. The curves relate
typical desorption properties of common sampling line materials
after being exposed to a “wet” gas such as the ambient
atmosphere. The curves illustrate the difficulty of obtaining a
fast system response when switching from a high dew point sample
to a low dew point sample. Even if the instrument responds instantly,
the sampling lines dictate the overall response.
Cleaning
Sampling Systems
Most types of metal tubing contain oil deposits on the interior
walls due to the manufacturing process. This residue must be removed
before putting the lines into service in a gas sampling system.
The lines should be cleaned and purged dry with air or nitrogen
before being placed into service. In addition to the initial installation,
the process itself may constitute a source of contamination and
in many applications these are volatile hydrocarbons. Care should
be taken not to introduce such hydrocarbons. Dew point hygrometers
are provided with a cleaning solution for use in cleaning and conditioning
the sensor mirror. The cleaning solution is isopropyl alcohol (IPA)
and is also locally available.
Contamination
Effects
System contamination and its effect on dew point measurement can
be subdivided into two categories: condensables and noncondensables.
Before proceeding, it is important that one understands that the
optical dew point analyzer measures the dew point, hence the vapor
pressure, of any substance that condenses on its mirror surface.
Conversely, regardless of concentration, contamination constituents
in a sample will not condense on the mirror unless its dew point
temperature is reached.
Condensables
Condensable can be further subdivided into soluble and insoluble
condensables. If insoluble, and its dew point is at or above that
of the constituent being measured, the relative concentration level
will mainly determine the effect on the measured dew point. If the
concentration level of the contaminant is low, i.e., has a low partial
pressure compared to the water vapor, then the effect of its presence
can be standardized periodically before it degrades the primary
measurement. This is done by heating the mirror surface to evaporate
the condensate and rebalancing the optical detection system. At
high concentration levels, the dew point analyzer may measure the
dew point of the contaminant rather than the water vapor compared
to many of the common contaminant. For example: if a water vapor
dew point of 32°F (0°C) was being measured at atmospheric
pressure (760 mm Hg) and the ethylene oxide were present as a contaminant
at a concentration level of 10% (76 mm Hg), its dew point would
be -31°F ( -35°C ). Since this is below the water vapor
dew point, it will not condense on the sensor’s mirror. However,
this means that there would be interference if the water vapor dew
point was below -31°F ( -35°C ). If the contaminant is,
in addition, soluble in the constituent being measured, it will
modify the vapor pressure and, hence, the dew point of the sample.
The overall effect will depend on the degree of solubility.
Noncondensable
The secondary category of contaminant is the noncondensables, which
can again be subdivided into solubles, primarily salts, and insolubles,
consisting of particulate matter. The soluble contaminant similarly
will modify the partial pressure, or dew point, being measured.
This type of contaminant affects all types of humidity instruments
and necessitates frequent cleaning of the dew point mirror, since
heating the mirror will not remove the salts. Insoluble matter is
most easily avoided through sample line filtration.
Sampling
Configurations
A suggested sampling system for use with dew point hygrometers would
be one where a portion of the gas line to be sampled is brought
to the hygrometer location from a pressure tap either by using a
suitably designed vacuum pump, or by expanding the sample to a lower
pressure. The flow rate through the main sampling line should be
sufficient to ensure continuous flushing of the lines, in order
to provide a fast response time for the sampling system. Usually,
the flow rate of 0.5 - 5.0 SCFH in a 1/4” line is adequate;
however, this number must be adjusted with the length of the line,
the level of absolute moisture content of the sample, and the desired
response time of the sampling system. A bypass line may be used
to increase the main sampling line velocity and improve the overall
response time. It is necessary that the sampling line be equipped
with a valve for adjusting the sample flow rate. The sample for
the hygrometer is obtained from the pressure drop across the by
pass as shown. It is desirable to provide the hygrometer input with
a filter, especially if the gas under study contains particulate
contaminants. Several sintered stainless steel filters are available
which are suitable. It must be remembered that the filter element
is considered a hygroscopic item, which will contribute some lag
to the sampling system. A rule-of-thumb in the design of hygrometer
sample systems is to minimize the number of components, such as
valves, tees, and filters prior to the hygrometer input. The hygrometer
output is connected to a flow meter and valve for adjusting the
flow rate to the recommended range of 0.5-5.0 SCFH.
NOTE:
Considerable cost savings can sometimes be made by recognizing that
the sample exhaust lines and related components need not be as high
a quality and as non-hygroscopic as those prior to the hygrometer.
Download
PDF version |
