What Is Isokinetic Sampling?
What Is Isokinetic Sampling?
What Is Isokinetic Sampling?
Flue gas is removed from a pipe by traversing at least two
axes across the pipe. A probe with a nozzle is inserted
perpendicular to the flow. The nozzle is pointed into the gas
stream. The nozzle is sized according to the velocity of the
flue gas.
Usually 8 to 24 different locations (points) within the stack
are sampled on a plane perpendicular to stack flow. The
location and position of the sampling points used during the
traverse is determined based on: 1) the size of the pipe,
2) straight run upstream of the sampling point, and
3) straight run downstream of the sampling location.
Generally, the greater the upstream undisturbed straight run
the fewer traverse points required. Sampling is most often
conducted by keeping the nozzle at each traverse point an
equal amount of time equally distributed over the time of
the sampling event. For instance, twelve (12) five (5) minute
points sampled over 1 hour is a frequent scenario.
Why might there be bias?
At each sampling point, the local velocity of the
flue gas (Vs) is measured and the flow rate of
the sampling train is adjusted so that the flow
within the nozzle (Vn) is the same velocity as
the gas in the duct you are sampling. This
condition is called isokinetic sampling. If the
velocity of the gas in the nozzle is too slow or
too fast, you cannot collect a representative
sample. If the velocity is too slow, this is
called under‐iso conditions. Some gas in‐line
with the nozzle has to divert and go around
the nozzle because the flow in the nozzle is
too slow. This represents a high bias for
particles in the stack because the particles
have inertia. Therefore, they continue on
their previous course and enter the nozzle.
So, the mix of particles to gas within the
nozzle is higher than in the stack gas – a high
bias. Conversely, if the flow in the nozzle is
too high, then gas that would have gone past
the nozzle is sucked into the nozzle, but the particles associated with that gas hit the outside of the
nozzle and don’t end up in the sample. Gas that didn’t belong in the sample gets pulled in, but the
corresponding particles did not. This creates a low bias because the extra gas dilutes the sample.
Isokinetic = No bias
Under‐iso = High bias
Over‐iso = Low bias
How is a test run conducted?
The sampling points are arranged within the stack such that each point is within an area created by
theoretically slicing the cross‐section both across the stack and in circles. Each section has the same
area as every other section. Because the circumference of the outer sections is greater than the inner
sections, the outer sections are thinner, and the outer points are closer together. Notice that none of
the points are at the center of the duct.
Because the velocity profile of the gas in the
duct is not uniform, you must measure the duct
velocity (vs) at each sampling point and then
adjust the sample flow rate up or down until the
velocity in the nozzle (Vn) matches Vs. The
calculations required to determine the nozzle
velocity on‐the‐fly are pretty complicated.
Therefore a laptop loaded with a pre‐
programmed XL sheet is used to do this. An
orifice based flow meter is installed in the
sampling train at the exit of the pump. This acts
kind of like a speedometer. You’d be hard
pressed to maintain the speed limit if all you had
to look at was your odometer. In an isokinetic
sampling train the volume meter is like the
odometer and the orifice meter is the
speedometer. A traverse usually follows a script
similar to this:
1. Leak check sampling train
2. Place probe in stack such that the nozzle is at point A1 – allow things to equilibrate for a minute.
3. Measure impact pressure of the gas using a pitot tube and measure the gas temperature.
4. Input this data to the spreadsheet. The spreadsheet gives you the setting for the orifice meter.
5. Record initial volume meter reading. Turn on the pump. Adjust the sample train flow rate to
the setting.
6. Maintain this flow rate for 5 minutes.
7. Move the train until the nozzle is at point A2. Keep the pump running while you do this.
8. Repeat steps 3‐6 until you have completed all of the points for axes A.
9. Remove the probe from the stack. Leak check it.
10. Move to Port B. Repeat steps 2‐9.
How does this apply to VOCs?
As illustrated in the above discussion inertia creates the potential for bias in the sampling. Therefore,
any chemical or physical target for the testing that might exhibit inertia needs to be sampled
isokinetically. Particulate (suspended solids) must be sampled isokinetically. However, water soluble
compounds must also be sampled isokinetically if there are water droplets in the stack. Also, aerosols
must be sampled isokinetically. Aerosols often form when a VOC is carried in a gas stream that cools as
it travels down the duct. At the hot end of the duct, the VOC saturation concentration is high, but
saturation concentration decreases with temperature. If the saturation concentration drops below the
actual concentration in the duct VOCs condense and form aerosol droplets. Droplets in the stack, unless
they are very small, behave much like particulate. VOC methods usually specify a constant rate sample
and they don’t use a nozzle. Most sample perpendicular to the flow instead of pointed into the flow.
This is equivalent to sampling in an over‐iso condition. So, any aerosol droplets or water soluble VOC
held within a water droplet are under‐represented in the sample. Thus, if a VOC method is applied to a
condition with water or aerosol droplets there is a significantly low bias to the results. There are
necessary modifications to this sampling arrangement that must be made. They will be situation
dependent. Please call to get guidance for your particular situation.