Chemical Engineering January 2014 - 43

Jonas
Causes of sampling
system errors
O
Figure 2. This weld-in style, single-port isokinetic sampling nozzle meets current
ASTM standards for sampling water and steam. Flanged connections to the process
pipe are also acceptable
sentative samples can lead to large
sampling errors [3]. Even in lowerpurity
systems, sampling errors due
to improperly designed sampling
systems can be significant.
A well-designed sampling system
(Figure 1) consists of an isokinetic
sampling nozzle (discussed below),
isolation valves, sample tubing, a
primary cooler (for steam and hightemperature
liquid samples), a
secondary sample cooler, pressurereduction
and total-flow-regulation
valves, a thermal-shutoff valve
(for process temperatures above
100ºF), back-pressure regulator and
sample drains.
Because steam impurities are easily
adsorbed by magnetite (Fe3O4),
the oxide buildup on the inner diameter
of the sampling nozzle and
tubing should be minimized. For
this reason, all wetted components
of the sampling system should be
made from at least Type 316 stainless
steel. Carbon and low-alloy
steels should be avoided.
Deposit buildup in the sample
lines can result in plugging of the
sample line or seizing of sample
isolation valves. Even when not
directly affecting sample flow, deposits
in the sampling system can
affect the sample accuracy. Deposits
can act as ion-exchange media and
adsorb or release impurities during
changes in the flow conditions. Even
the best sampling-system design is
still susceptible to deposition and
plugging if the cycle chemistry at
the plant is not maintained within
industry standards, particularly
when high concentrations of corrosion
products (such as iron oxide or
copper oxide) are present. Lengthy
sample lines (for instance >100 ft) or
low sample velocities (for instance
< 4 ft/s) increase the probability of
sample line blockage and can cause
unacceptable time lags between
sample collection and analysis. A
sample flowing at 2 ft/s through 500
ft of tubing will take over four minutes
to reach the analyzers.
Why isokinetic sampling?
Isokinetic sampling is the extraction
of a representative portion of the
process stream without altering the
physical and chemical properties of
the sample. In isokinetic sampling,
all phases (solid oxides and precipitates,
liquid droplets and vapor) of
the sampled fluid enter the sampling
nozzle with the same velocity
vector (meaning the same velocity
and direction of flow). The main
reason isokinetic sampling is necessary
is that the sampled stream
is almost always a two-phase fluid
(gas-liquid, gas-solid, liquid-solid)
and the second phase typically has
a very different chemistry composition
than the steam or water [2]. In
addition, the second phase (droplets
or particles) typically has a different
density and inertia compared to the
primary phase (gas or liquid) and
therefore would not be proportionally
represented in a sample that
was not withdrawn isokinetically.
The benefits of isokinetic sampling
have been verified during an Electric
Power Research Inst. (EPRI)
project [3] and through an independent
analysis [10].
Sampling nozzle design
The design of the isokinetic sampling
nozzle (Figures 2 and 3) is a
critical part of the sampling system,
and should be performed prior to
the selection of the other sampling
system components. As noted, if
designed incorrectly, the sampling
nozzle could provide a sample that
perators should be mindful of these
common sources of sampling errors
(in order of priority and impact):
* Sample withdrawal - Sample does
not (1) represent the stream (due to
wall effects, stratification, not isokinetic,
or mixing issues), and (2) represent
all phases (solid, liquid, gas)
* Deposition in the sample line (could
also result in plugging)
* Release of built-up deposits into the
sample stream (leading to spikes)
* High pressure drop leading to insufficient
sample flow (typically due to
long sample lines)
* Changes in sample flowrate (for instance,
sampling system can take up
to 6 hours to reach equilibrium from
the start of sample flow)
* High sample temperature (can lead to
pH and conductivity errors)
* Chemical reactions in sample lines or
coolers (reduction of oxygen concentration,
change in pH and so on)
* Corrosion of the sampling system may
lead to generation of corrosion products
(as a result of improper materials
of construction)
* Filters in the system interfere with desire
to sample suspended solids
* Sorption on sample tubing and suspended
oxides may remove a portion
of the chemical species being
monitored
is not representative of the conditions
in the pipe. Proper sampling
-nozzle design must consider the effects
of flow- and vibration-induced
forces on the nozzle, as well as the
design pressure and temperature.
Prior to 2006, the ASTM Standard
D1066 " Standard Practice for Sampling
Steam " [4] included a multiport
sampling nozzle, which - in
its most basic form - consisted of
a piece of pipe with multiple holes
in it. The sampling pipe extended
most or all of the way across the
process pipe and was supposed to
simultaneously sample from several
locations across the diameter
of the pipe. However, research has
shown that such a multiport design
operates non-isokinetically, is prone
to plugging, and is susceptible to
failure due to vibration [3].
In many piping applications, the
flow in the process pipe is fully turChemiCal
engineering www.Che.Com January 2014 43
http://www.Che.Com
Chemical Engineering January 2014

Table of Contents for the Digital Edition of Chemical Engineering January 2014
