ML20245B418
| ML20245B418 | |
| Person / Time | |
|---|---|
| Site: | Browns Ferry, Sequoyah, 05000000 |
| Issue date: | 08/17/1988 |
| From: | Blair C TENNESSEE VALLEY AUTHORITY |
| To: | |
| Shared Package | |
| ML20245B391 | List: |
| References | |
| NUDOCS 8906230169 | |
| Download: ML20245B418 (24) | |
Text
_ _ _ _ _ _ - - - _ -. _ - - - - _ - - _ _ - - - - _ - - _.. _ -
I.~.
i e
L INVESTIGATION REPORT DESIGN AND OPERATION OF SAMPLING, SYSTEMS FOR 4
p, ANALYSIS OF HIGH PURITY WATER CW 0m[ i 7/f/A D h a i1-s os
{PreQ"erT Date Preparer Date WW2&l -
l 1-//- W
~ReVieief ~
Date I
8906230169 890615 PDR ADOCK 05000259 P
PNV 02871/DNS01
r 4
ABSTRACT The technical bases for the design'and operation of chemical sampling systems for high purity water systems are examined. Temperature control and regulation; behavior of suspended meterials in sample lines; flow regulation and control; pressure regulation and control; materials of construction; sample line deposit - sample interactions; sample fittings; sample probes; sample conditioning filters are reviewed in detail.
It is concluded that:
1.
Sample coolers should be placed as close to the root valve as practicable.
2.
Pumped refrigerant trim coolers are preferred for most tria cooling applications.
3.
Critical sampling for trace level impurities requires a constant, continual sample flow, at a nominal velocity of 6 f t/s.
4.
Bypass streams may be used to keep flow conditions within acceptable rangec.
5.
Rotameters ma-be needed to verify proper sample flowrote. They should be installed in a downstream position.
6.
Sample systems should be constructed of 1/4 in. 0.D., 0.049 or 0.065 in.
316 stainless steel sean?.ess tubing.
7.
Compression fittings are preferred over welded fittings.
8.
Neither a positive or negative recommendation can be mW for installation of sample probes or taps.
9.
In-line filters are not recommended for normal sampling applications.
10.
The sample will be affected by the sample system, even when titanium is used. These. effects may be larger than the parameter being measured.
Ef fects cacnot be eliminated, only minimized.
- 11. Sample lines should be as short as is practicable with a minimum number of fittings and valves. 02871/DNS01
F ',,
L k
1.0 INTRODUCTION
1.1 Purpose-N This report presents the technical basis for the design and operation of sampling systems for high purity water systems (condensate, feedwater, etc.).
1.2 Scope h is report is applicable to the monitoring of condensate, feedwater, and similar high purity water process streams for chemical species-both soluble and insoluble.
It is not intended to provide guidance for the sampling of steam, tanks, nonhigh purity systems, or nonaqueous systems. Considerations discussed in this report may be applicable to sampling for radiochemical species.
1.3 Background
Sampling system design and operation is the subject of a previous report (" Investigation Report - Sampling Condensate and Feedwater for Trace Level Impurities", September 17, 1985, L29 800917 833).
Since the initial report was issued, there have been new reports of instances where " sample line phenomena" biased the measurement of contaminants in nuclear plant process streams. This report incorporates recent data and observations to define the " state of the art".
2.0 DISCUSSION 2.1 General Considerations Recent technological advances (such as the online ion chromatograph) are allowing contaminants to be measured at increasingly lower concentrations. With today's technology, the ability to get accurate, meaningful analyses is often not limited
' by analytical equipment or expertise of personnel, but by the physical and chemical processes occurring in the sampling system.
What are these physical and chemical processes that affect the ability of a sampling system to deliver a sample? How can the sampling system be designed and operated to minimise unwanted interactions between the sample atream and the sampling system?
These are the questions that will be addressed in this report.
Traditionally, it has been assumed that soluble contaminants will pass through a sampling system unchanged. While this may be a good assumption for contaminants at the parts-per-million level, it is generally a poor assumption when attempting to sample and analyre for contaminants at the parts-per-billion or parts-per-trillion level.
Species of interest will interact with sampling materials or the unavoidable deposits that form on the surfaces of the sampling j
system. These interactions can introduce " sample line effects" at l
1 I 02871/DNS01
a magnitude equal to or greater than the variable being measured.
Species of interest may chemically react, thermally decompose, precipitate, or otherwise change in nature. What goes into a sample line is not necessarily what coses out of a sample line.
2.2-Temperature 2.2.1 Sample Coolers Devices for cooling sample streams may be classified into two categories roughing or tria.. The function of the roughing cooler is to lower (the temperature of the sample stream to (1200F, which is adequate for many purposes.
Some applications, however, require that the sample stream temperature be controlled within a fairly narrow range.
c Here, an additional tria cooler may be used.
Roughing coolers are commonly manufacturered in a shell and-tube or tube-in-tube configuration. Both types can be obtained with high heat transfer coefficients in a variety of materials. Each will work well if properly sised and selected.
Roughing coolers ordinar!!y use water as the cooling fluid.
Occasionally, air is used as the beat transfer medium. This is not recommended as a general practice. The power plant environment is quite warm, with ambient air temperatures often in excess of 1200F.
In addition, the limited ability of air to transfer best may require mechanical air circulation, thus greatly increasing the chances of mechanical failure and maintenance requirements.
There are two basic types of trim coolerst bath and pumped refrigerant (figure 1). The bath type usually employs a coil submerged in a cool fluid. Heat is exchanged through the walls of the tubing to the fluid in the bath which is refrigerated. Rath heat exchangers work best when the process stream is maintained at a constant temperature and flow rate is held constant. There is a temptation with the bath type cooler to regulate sample temperature by regulating sample flowrate; unfortunately, this often compromises the representative:*ss of the sample.
(Notes flow regulation is the only means of controlling sample temperatures when the bath cools more than one stream.)
Some bath trim coolers employ a sample splitting device to increase the ability of the cooler to cope with changes in sample temperatures. While the sample splitting feature increases the positive control over sample stream temperature, it introduces some problems. When flows are split the flow rates in each section will be reduced (assuming constant line diameter).
This may allow suspended matter (crud) to settle in the coil portion. When the temperature of the incoming sample stream increases, flow i
through the coil is increased, and a portion of the 02871/DNS01 l'
____w
_..__-_--y______
e
- f, M
Mll6 mfqR 3R amamIns
~IWF MIDS TEN' mast EnATWE -
m 3duptg g?
._ [ ;
I
?
CELim uten Im y,,g, gy r -
m,7 - --
\\2 A^MNW armistantto U
4 lastT g
costs h pq y
=
p:
4 BATN 4--
CELI 4 W47tR art ALMUMTICA(LY ASJJgfAgLg TH4E mY v4 Lug BERNM195
- lwr M
N!Get Ttrthanat '
g N.
SAMPLE sinust a
CELINE WAfta 1st =
r
%,7 - -
\\2
^
W MPRimanTISt O
M 45st?
$8'8 - _ D4 W
BATH VITH SMPLE MITTING W
AmfWIM8
~laga r MISM TDegnAtyg' CELIE i
h
- I N RCrt!GreAtta EAT "I'
- rLu
=. er i
!.._.. _ _ I=romaru
=c PUNPED REFRIGERANT FIGURE 1 - BATH AND PUMPED REFRIGERANT TRIM CDOLERS l
l
- ox,..g yy N'
inventtry cf trapped crud may be released, temporarily-4g' compromising the representativeness of the delivered sample stream.
Pumped ref rigerent, trim coolers -.(PRTC), on the other hand.
employ tube-in-tube or shell-and-tube haat exchangers. Beat is. transferred to'a' fluid which is pumped from a separate refrigeration unit. The PRTC controls the temperature of the sample.by controlling the flowrate of refrigerant and allows timely response to changes in incomins' sample stream-temperatures, while maintaining a constant sample flowrate.
.PRTCs can be purchased to independently cool sore than one-sample stream.
Of the above types of trim coolers, the PRTC allows the' maximum flexibility in using space.. The cooling baths-are 4
' typically larger than.the heat exchangers used with the PRTC, and usually must.be located near the floor level.
p Heat exchangers used with the PRTC can be placed almost anywhere; the refrigeration unit can be'placed-in a remote spot (as can be done with refrigeration units for bath type
. trim coolers).
This flexibility is important where space is limited or radiation exposure from sample line volumes are a concern.
2.2.2 Chemical Reactions'and Temperature Temperature 'ean. have a major effect on the processes occurring in the sample system.- Chemical species can-undergo irreversible changesLin-the sample systen which will introduce a measure of nonrepresentativeness into the sampling process. Reactive chemical species can thermally y
decompose or otherwise react in the time it takes to travel through the sampling system. Cooling the process stream as quickly as practicable can,.in most cases, help preserve the
~
chemical nature of the sample stream.
On the other hand, it has been observed that some species will undergo irreversible transformations on cooling.. Thus, if a sample stream is cooled, nonrepresentativeness can be introduced.
Hydrazine, an oxygen scaveng.9r, is very reactive at elevated temperatures and readily decomposes at temperatures above 4000F. The rate of reaction of hydrazine with dissolved oxygen is a strong function of temperature, with the rate being quite fast above 2500F (1).
In addition, at elevated temperatures, dissolved oxygen can rapidly react with a large number of species other than hydrazine.
It is necessary. therefore, to cool high temperature sample streams as quickly as possible if hydrazine, dissolved i
oxygen, or other similar reactive species are of interest.
The measurement of electrochemical potential (believed to be an important measure of corrosiveness) is sensitive to 02871/DNS01
, ory c
pfy
+
m.
.3 16.g y
f, temper:ture. In order to get s'truw measure of
. electrochemical potential, the measurement aust be taken at '
m :6
'very near process. temperatures.. In this case, the sample-
~
stream must be insulated. to prevent best loss, rather than cooled (2).
u
.It has been observed that many metallic species existing.in r
process' streams at high temperatures undergo irreversible
. transformations in a~ sample:line when that line is cooled.
E N;;ctative' sampling ior'these species may require high
-temperature flitration (3,4). Because_of'the difficulties l involved in performing high temperature filtration, it is a
- common-practiceito cool the sample-and then filter it.
The
. material collected'can then be.. analysed.for the. amount'of
- elemental metais' present' (speciation-determination of the.
type of metal oxide--may not be practical)..
2.2 Temperature Summary l '.. Roughing coolers.using water as the heat exchange medium are recomunended to reduce high temperature sample streams to <1200F.
2.- Pumped refrigerant tria' coolers are preferred to bath type ecolers for most applicaticas.
3.
For most applications, and especially for hydrazine and dissolved oxygen monitoring, the roughing' cooler should
'be as close to the sample root valve'as practicable.-
Tria coolers may be located'for convenience.~
2.3 Flow
.2.3.1 Suspended Material
,i Sample streams consist of two phasest liquid and solid.
.The solid phase (principally metal oxides) is more dense than the liquid phase and tends to deposit-in the sampling system by settling and by adhering to sampling system surfaces.
Soluble species can interact with sampling system surfaces, suspended matter, and material deposited in the sampling system,'potentially leading to irreversible physical or chemical transformations.
These interactions are quite often a function of surface area. Therefore, minimizing surface area can help minimize potential sampling biases.
This may be done by reducing the amount of material deposited in the sampling line and by reducing the total length of the sampling line.
Iv Some authorities state that the deposition of suspended I '
matter will be minimized by maintaining turbulent flow with a high Reynolds number (3,5,6).
Others suggest that the total accumulation of suspended matter is also influenced by impaction-deposition and erosion of suspended material (/). 02871/DNS01 L_
=_
JQ, a
l The amount of suspended' material depositing in a sampling;
}
5 6
system is a function of the' fluid velocity (8). The L
variation in weightfof deposited material as a function of K
velocity is shown.in figure 2.
It appears that for a h p~
horizontal pipe, crud deposition is minimised at fluid l
velocities in the range.of 5 to 7 ft/s.
(Although deposition'is less at fluid velocities above 20 ft/s, these.
high velocities are impractical because of excessive L
pressure drop and. sample throughput).
In the absence of L
other data, it will.be assumed that figure 2 represents p
typical sampling system behavion-y l
. Initially.there is a net deposition on surfaces'of the sampling system. Later, some of.the deposited material is I
eroded. ' Eroded material may be redeposited and then eroded several times before emerging from the sampling' system.. De eventual outlet concentrations of suspended matter are at mixture of noneroded (" fresh") and eroded (" stale")
materials (7).' Consideration of this phenomena indicates that it is quite likely that even the most carefully designed sampling system will show a " memory effect." This
" memory effect" can make representative sampling for short-lived species quite a difficult proposition.
The deposition of suspended material will eventually' reach an equilibritan where the net rate of deposition equals the net rate of erosion. Fluid velocity has a great influence on the time it-takes for a sampling system to reach equilibrium.(figure 3). -Equilibrium is reached in about one month when the velocity is a constant 6 ft/s.
If the velocity is 1.0 f t/s, equilibritse is not achieved until after one year.
(As with figure 2, the general applicability of figure 3 may be questioned; but it is clear that fluid velocity as well as Reynolds number influences deposition of suspended material.)
ne total accumulation of suspended material is greatly affected by the presence of " crud traps." Crud traps may be expected in areas of low fluid velocity such as dead legs, valves, fittings, and sharp bends in tubing. Of course, minimizing these potential crud traps will reduce the accumulation of suspended material and attendant problems.
Suspended matter can settle when a sampling system is shut down. When flow is reestablished, a "crudburst" may be expected. Crud bursts can release not only a portion of the previously deposited materials, they can also release trapped soluble species and bias downstream measurements for hours (9).
Crud bursts can also result from varying the velocity of the sample stream. Therefore, it is important that a constant, continual flow be maintained.
The need to maintain constant flow is demonstrated in the results of a sampling campaign undertaken at the LaSalle BWR.
It was found that it was
! 02871/DNS01
,., o d
^
I g
1 I
l l
IO l
1.ii1111l l
l-1 I Illy r
Horizontal E
f:.
M
- g, g
h W'
,/
=_
{
/-
/
/ ertical V
r g
<f
=
g a
2 30-8 _.
2
~
?
m
)-
Spm Porticles 0.65 in. Ip. Tubes 250 F Averoge Temp 10
I I I IIIII I
I IfIII c'
c' 0:
Fluid Velocity, f t/see FIGURE 2.
Fluid Velocity influence on Deposition (8)
. ;)
16' 1
I I IIIIll 1
1 I i I ii1l.
I I I I Iiit
_~
Fluid Velocity Ft/Sec 6
C #
==
E 20 E
4 10 8
kKi"
)
~
=
c
~
R
=
3 C
3 g'
i
- ' S
~:
2 E
.?
'i 5
3
= m' i
{
5pm Porticles g
0.65 in. I.a Tubes E
250*F Average Temp
-3 o*
g
~
l Month-
' I Year-g-4 I
I I f li t fl l
I I I I "f l}
l l I ill11 o'
o os r
o*
Time, hours FIGURE 3.
Deposit Behavior With Time at Selected Fluid Velocities (8)
~
l I'
L' mora important ta maictain a constant flow than it was to operate at a varying, but optimal. flowrate of 6 ft/s (10).
l l-2.3.2 Bypass Streams l-Of ten the flow required to give optimum conditions (Reynolds number, fluid velocity, etc.) exceeds that which can be handled by analyser. For example, some analyzers require a flow of 100 to 200 al/ min; the optimum sample line flew, however, may be ne'arly 1200 ml/ min. A practical way of handling this dilemma is to use a bypass stream to maintain proper flow conditions in the bulk of the sampling system.
At a point very near the analyser, the flow may be split and a portion is routed to the analyser. The excess flow may be routed to waste or returned to the process stream, as appropriate.
The use of bypass streams has found wide acceptance (3, 6, i
11, 12). Their use is not without penalty, though. The l
bypass stream serves no function other than to maintain proper flow conditions, and the. portion of the total flow that is attributed to the bypass is typically larger than that of the rest of the. sample stream. Yet, tho bypass flow must be removed from the process stream, cooled (if necessary), and returned to the process stream or routed to-waste. Each of the preceding carries an economic penalty.
Continual flow of sample streams also caries economic penalties. Prudent design of the sampling system can minimize these penalties. For example, return of a sample or bypass stream to the process, stream will reduce the need for make-up water and reduce waste.
2.3.3 Line Diameter and Flow The economic penalties associated with maintaining proper flow conditions can be lessened by the use of small bore sample tubing. Smaller diameter sample tubing will allow proper flow conditions to be achieved with lesa total throughput than large diameter tubing. The relationship between tubing size, flowrate, and velocity is illustrated in figure 4.
Lines smaller than 1/4-inch OD are prone to plugging. Their use, therefore, should be restricted to specialized applications.
2.3.4 Flow Regulation Classically, the globe valve and analogues such as the needle valve are used to throttle, or regulate, flow. Flow l
regulation valves usually cause large pressure drops; therefore, these valves reduce pressure as they regulate flow.
Flow regulation valves are available from a number of vendors in a wide variety of types, sizes, and materials of 02871/DNS01 i
m G
/
l S
m E
8 B
00 U
/
yi 0
T 3
D E
53 s.
T C
E 5, x 0
f, z 0
n L
/
5 E
2 1
S 6,
2
/
R 1
O t
F t
0 3
0 n
E 0
/
2 T
+
R R
0 H
O 0
L 4,
/
F 0R s
f 5W 3
1
+
s is v
/
Y
. 0 T
~
0 I
i 0
C O
- N' 1
Xn L
2,
+
E f
V D
0 I
0 i
5 U
L
~
F 4
E E
O R
9 8
7 6
5 4
3 2
I U
G
^inNHL uv yFHUOJ.J> asD1_Lu a
I
.L F
construction.
Globe valves.(along with their analogues) may be used as isolation valves for low pressure operation.
Back pressure flow regulasing valves (combining both globe
'~
and diaphragm features) have been used with success in maintaining constant flows through sampling systems.
Some types of valves (gate, plus, ball, etc.)'are designed for on/off operation. They make excellent isolation valves; their use as flow regulators, however, is not recommended as a general practice.
A sampling' system will often contain a combination of two or more of the above types of valves, with each type performing the function for which it is most suited. Figure 5 illustrates the potential use of valves in a sampling situation.
2.3.5 Flow Measurement Flow measurement is a requisite if the sampling system is to be operated at the proper flow conditions. Brscause of the
(
relatively low flows (<1500 ml/ min) encountered in sampling systems, the devices to measure flow are limited.. Typical process flow measurement devices (orifice and venturi
'i meters,. vortex shedding meters, etc.) are impractical.
Rotameters are suitable and are, by far, the most common flow measuring instrument used in sampling applications.
Rotameters must be used with some caution. The types of rotameters normally used for sampling cannot withstand high temperatures and pressures.
In particular, rotameters cannot be used in instances where isolation of a downstream valve will cause the rotameter to be subject to high pressure. Rotameters are potential crud traps and have been known to allow air ingress, especially when equipped with an integral flow control valve. For many applications, it is wiser to locate the rotameter downstream of the instrument or analyser to avoid potential contamination problems.
4 Downstream location of a rotometer may not eliminate all of the problems associated with crud traps or oxygen ingress.
A case in point is a dissolved oxygen analyzer which must be verified by a grab sample.
Pickup of oxygen through a downstream rotameter might cause an apparent discrepancy between the instrument and at the analyzer when, in fact, none may exist.
Integrating flowmeters, which measure the total flow over a period of time, are very useful for some applications, such as corrosion product sampling.
The types of integrating flowmeters that are available for the typical sampling application are limited.
The " drip meter" is a relatively 1
inexpensive possibility.
The drip meter has the 8-02871/DNS01
]
FC Z
.. c L
T i
Y PROCESS STREAN ROOT. VALVE (GATE, BALL,OR PLUG TYPEI ROUGHING COOLER, r.
--)
u
~1000 nt/nin p
es PRESSURE REDUCING VALVE e
j BYPASS LINE ISOLATION VALVE' ISOLATION VALVE NEEDLE. VALVE TRIM CDOLER
.)
BACK PRESSURE v
FLOW REGULATOR GRAB SAMPLE 200 nl/ntn
/
ANALYZER ROTAMETER
[NEEDLEVALVE DRAIN-ROTAMETER
)
4 DRAIN I
Figure 5 FLDV REGULATION NETVORK
- 1. GRAB SAMPLE ISOLATION VALVE NORMALLY CLOSED-0THER VALVES NORMALLY OPEN i
- 2. BACK FLOV REGULATOR HELPS MAINTAIN CONSTANT FLOV AND PRESSURE THROUGH i
ANALYZER VHEN GRAB SAMPLE VALVE IS OPENED
- 3. COMBINATION ISOLATION VALVE-NEEDLE VALVE ON GRAB SAMPLE ALLOVS FLOV RAT TO BE SET (NEEDLC VALVE) AT SELECTED FLOV RATE, ISOLATON VALVE PERFORMS
.ON/DFF FUNCTION, GRAB SAMPLES CAN BE TAKEN AT REPRODUCIBLE FLOV RATE.
iy os'?
,w~
p (h < ~
j
+
Qt,e
' disadvantages of temperature and' pressure limitations and y
ip '
- due to.its large-internal volume, can be a pot etial L ',, ' ' "
. radiation exposure source.
J2.3.6 Flow Summary 1.
The deposition-of suspended matter in sampling: lines must be'ainimized'in order.to maximize the
,f representativeness of the agple.
2.
'The sample stream should have a fluid velocity.of about e
t6 ft/s and be turbulent in order to minimize deposition and allow the sample line to. equilibrate in.a reasonable
' amount-of time.
e 3.'
Flows should be constant to minimize crud bursts and allow equilibrium between deposition and erosion.
r 4.
When'it is not_ feasible to allow a sample line to' flow i -
' continuously; an extended flush at a. constant'flowrate.-
r
'_may be used. ' Twenty-four hours at 6 f t/s is recommended m
.as a goal.
5.
Bypass streams may be necessary to maintain flows within proper bounds. The bypass stream should allow these flow conditions.to be maintained for as much'of the
-length of'the sample line above the analyzer or grab s'emple.. point as possible.
- 6. : In order to minimize costs associated with maintaining-
~
preferred flows, small. internal diameter tubing should be used.
7.
Valves must be carefully matched to their intended'
- function, 8.
Rotameters, or other flow measurement devices, are i
a recommended for most continuously flowing zample streams.
Flowmeters should be located downstream of g
analyzers.
' I 2.4 pressure 2.4.1 pressure Reduction and Regulation In many sampling situations, the pressure of the sample stream must.be reduced from that of the process stream
{
(minus head losses resulting from passage through the sampling system). pressure reduction may be necessary as analyzer requirement or for personnel safety.
I 4,
02871/DNS01
l, 4-
.i-L,',
In lower. pressure applications. globe type valves (includiAg i
needle: valves) may be used to regulate.both flow and pressure. This is a consequence of the characteristic high pressure drop across these valves.-
gigh pressure. applications require special pressure reducing devices..Many' devices.(e.g... drag valves or rod-in-tube valves)~are prone.to plugging and may be considered high L.
. maintenance items-(13).
A self-contained pressure, regulator valve may be expected to be more maintenance free and has;the capability of both-L controlling and regulating pressure. Self-contained pressure regulating valves for sampling use are normally
. spring loaded / diaphragm mechanisms. These valves have the advantage of automatically responding to changes in process.
1 or;in downstream pressure using only the energy of-the flowing fluid (16).
l It is good practice to consider the need'for pressure relief 4
valves in the design of sampling systems. Automatic pressure regulators can fail, manual controllers can be Ll improperly manipulated, or the sample stream discharge may be unexpectedly isolated or plugged.. Pressure relief valves 4
.can prevent personnel injury or damage.to expensive process 1
. analyzers.
2.4.2' Sampling low-Pressure Streams When sampling low pressure process streams, artificial means-(pumps) may be needed to force the sample through the-sampling system. The classical example is condenser hotwell sampling, where it is necessary to increase the pressure of the sample' stream from approximately 27-inch Ng vacuum to 20 psig or more. Both centrifugal and positive displacement pumps have'been used for low pressure sampling.
An important consideration in low pressure sampling is prevention of air ingress from the atmosphere into the sample stream. Air can enter through fittings, valves, or through pump seals.. Other potential problems are leakage of hydraulic or lubricating fluids from the pump into the sample stream and damage.to positive displacement pumps by entrained solids.
In-line strainers may be necessary to prevent damage to susceptible pumps. Contamination of the sample stream by 1
the pump can be eliminated by using a pump with no seals that can leak. A magnetically coupled gear pump is an 4
example.
{
I Air ingress through fittings and valves can be eliminated by careful selection, proper assembly, and good maintenance 3
practices. This ingress can be a serious problem that is I
- 02871/DNS01 j
difficult to d:tect, and even harder to diagnose;.therefore, 1
it is important that careful attention be paid to preventing the problem.-
,2.4.3 Pressure' Summary f
}
1.
Many throttling valves can serve a dual function as
{
pressure reducers.
i
- 2. - Self-contained pressure reducing valves are recommended
.1 for reducing-high pressure (<2500 psi) sample streams.
j i
'3.
Magnetically coupled gear pumps are ~ reconsnended for low j
pressure sampling applications, such as hotwell sampling.
2.5 Materials 2.5.1 General Considerations
,1 i
The. materials used in the sampling system must be capable of wi?hstanding high temperature, high pressure, and corrosive en.
raments. The materials should be inert-the sampling system should not add anything to or remove anything from the sample streas.
In practice, it is possible to select materials that can withstand system conditions. Truly inert
]
materials do not exist.
j The overwhelming choice for sample lines throughout the 1
industry is 316 stainless steel, although titanium is
)
sometimes recommended (15,16). Plastics should be avoided as they are usually permeable to atmospheric oxygen or 3
carbon dioxide to a significant degree and can be a source
]
of organic bias. There is evidence that even polypropylene and teflon can absorb / release measurable quantitles of organics (17).
i Consideration must be given to the other materials in the 1
sampling system. Sample coolers are available in a variety 3
of materials including stainless steel and titanitse. Valves and fittings often contain materials which can introduce bias into the sampling process.
Seamless tubing is recommended for sampling applications.
Tubing is available in a variety of small diameters and wall j
z thicknesses while piping is not.
Tubing is smoother than piping and will therefore transmit the sample with less pressure drop, incur less deposition of suspended material, and be less subject to surface phenomena (heterogeneous solid-liquid interactions).
2.5.2 Surface Phenomena The sample stream will be in contact with the sampling system materials and any deposits in the sample system.
There will be interactions, both chemical and physical, as a 02871/DNS01
-2i s
3 s
g
[
rxult cf. this contact. These interactions may be considered to be a form of. surface.. phenomena, that is, the interactions between the sample stream'and.the sampling system may be treated as if they. occur on the surface of the f'
g( u-sampling system (and deposits)..
Many chemical reactions are surface catalysed. Physical
,y
' adsorption occurs on the surfaces of chemically-inert i
material. ' Ion exchange can take pir.ce between a variety of materials, especially if ion exchange resin fragments are present in deposits.- Effects.such as these often lead to subtle, but measurable changes' occurring in the sampling systes.-
[.
Surface phenomena cannot be totally avoided. They can be diminished;:however. Since-surface phenomena are proportional to surface area, reducing surface area will reduce surface related problems.. Surface. area is decreased S'
. hen deposits are minimized and when sample lines are as!
w short as practical.
o Some of the examples.of surface phenomena are sufficiently interesting to warrant their mention in this report.
.1..Hydrazine is'used to deoxygenate makeup water. The rate of the reaction of hydrazine and dissolved oxygen ~is
.normally very slow at low temperatures.
It has been found, however, that the hydrazine/ oxygen reaction will proceed at reasonable rates if the process stream is flowed through a vessel of activated carbon (chemically inert) whose surface contains sufficient adsorbed' hydrazine.- This has direct significance for sampling where there may be much higher temperatures and mort reactive surfaces (than activated carbon) to catalyse the hydrazine/ oxygen reaction.
2.
Oxygen consumption has been noted in sample lines, even in low temperature, well conditioned sample lines. This dissolved oxygen consumption has been related to line length, material, and age. The magnitude of this effect
.can be'nearly as large as the amount of dissolved oxygen
'being measured (PWR applications) (18,19).
3.
Chemisorption of cobalt in sample lines has been demonstrated at the Vermont Yankee BWR.
In a test, Co-59 was injected in the condensate demineralized o
effluent.
It took about six hours (instead of the expected few minutes) for the injected Co to appear in feedwater samples, a phenomena the authors attributed to chemisorption.
The authors also state that I
chemisorption may be expected even on titanium (16).
! 02871/DNS01 l
i 2.5.3 Fittiras l
Either welded or compression fittinga may be used for
[j sampling systems. Compression fittingo might leak where l'
welded fittings might not. Welding of small diameter lines, on-the other hand. can cause line plugging as the result of l
bore shrinkage (9). Experience at TVA's nuclear plants has shown that compression fittings can be used reliably.
~ If compression fittings are selected, care should be taken during the assembly of the sampling system. Specifically, since sample tubing diameter can,be greatly reduced during-
-cutting operations, it may be necesscry to ream cut tubing to its original diameter before tubing pieces are joined.
)
Each fitting. whether welded or compression, will cause a small perturbation in flow.
this will lead to a pressure drop and increase the potential for. crud trap formation.
Therefore, it is prudent to minimize the naber of fittings used.
2.5.4 Materials Summary 1.
Seamless 316 stainless steel tubing is recommended for normal sampling applications.
2.
The sample stream alll be affected by the sampling system materials, even when stainless steel or titanium is used.
3.
The number of fittings, valves, etc., should be minimized.
2.6 Additional Considerations 2.6.1 Sample Probes Problems involved in sampling aerosols leo to the development of sample probes. The concept or using a sample probe has since been extended to the sampling of liquid streams such as condensate and feedwater. Sample probes are designed to minimize the preferential routing of particles i
(or droplets in an aerosol) by virtue of their momentum.
It should be noted that there is tremendous dissimilarity in momenta of aerosol droplets when compared to suspended corrosion products in water (compare relative densities and fluid velocities).
Sampling probes are often recommended as needed for representative sampling of suspended raaterial in liquid streams. Sample probes usually extend one-third pipe diameters or more into the process piping to avoid sampling the region of laminar flow which exists very near the pipe wall.
! 02871/DNS01
pT There are potential problems.in.using sample probes. : Probes Lcan break off and travel'through process piping, possibly' causing equipmentL damagef and creating a safety concem.
It-may be argued'that' sample proces introduce noncepresenta-s tiveness into the sampling process by avoiding sampling of L
the laminar sublayer alcng'the pipe wall; a truly "y
representative sample must include this small fraction.
- In the absence of experimental data, the benefit from using sample probes is arguable. A representative sample is not' guaranteed when a sample probe is used. Likewise, the:
absence of a sample probe does not necessarily-indicate 4
-nonrepresentative sampling.
2.6.2 Sampling Conditioning Filters Sampling' filters are sometimes' recommended to keep process instrumentation from fouling with suspended material (20).
Although filters may be advisable in some instances, there.
are problems'concerning their use.
Filters can be thought-of as a type of fitting, that is,-
they will cause a perturbation in flow leading to a pressure drop.. Most types of f11ters collect suspended materialand the pressure drop increases es the filter stays in service.
This change in p: essure drop can reduce flows through the sample' system, with attendant probises regarding representativeness. Filters can' introduce foreign materials into the sampling system to bias downstream results. Worst
'of.all, most filters are,'by definition, crud traps.
Trapped crud can ' severely compromise the representativeness of downstream analyses and can be a potential source of
~
radiation exposure.
In. addition, in a well-designed and operated sampling system, suspended materials are present in direct proportion and kind to the process stream. Removal J
of this suspended material makes the sample stream, to some degree, nonrepresentative.
i Most modern process analysers are designed to operate with a minimum of fouling when operated at proper flow rates. This reduces the need for filtration or frequent cleaning..
Some analyzers will require filtration to function properly, The in-line ion chromatograph, for example, will plug with suspended material within a short period of time if the sample stream is not filtered.
i Filters, if required, should be either self-cleaning or easily cleaned. One type of filter that has been used with some success employs a cyclone type construction where flow i
is tangential to the body of the filter, and suspended material (usually more dense) is swept to a side stream where it can be discharged. 02871/DNS01 i
...m...
.J
.Thero are potential problems in using sample probes. Probes D
can break off.and travel through process piping, possibly
. causing equipment damage and creating a safety concern.
It may be argued that sample proces introduce nonrepresenta-tiveness into the sampling process by avoiding sampling of j
the laminar sublayer along the pipe wall;'a truly j
i representative saQ1e must include this small frriction.
J q
In the absence of experimental data, the benefit from using sample probes is' arguable. A representative' sample is not guaranteed when a sample probe is used. Likewise, the b
absence of a' sample probe does apt necessarily indicate nonrepresentative sampling.
2.6.2 Sampling Conditioning Filters Sampling filters.are sometimes recommended to keep process instrumentation from fouling with suspended material (20).
Although filters may be advisable in some instances.there are problems concerning their use.
y Filters can be thought of as a type.of fitting, that is, they will cause a perturbation in flow leading to a pressure drop. Most. types of filters collect suspended material and the pressure drop increases as the filter stays in service.
This change in pressure drop can reduce flows through the sample system, with attendant problems regarding representativeness. Filters can introduce foreign materials into the sampling system to bias downstream results. Worst of all, most filters are, by definition, crud traps.
Trapped crud can severely compromise the representativeness of downstream analyses and can t,e a potential source of radiation exposure.
In addition, in a well-designed and i
operated sampling system, suspended materials'are present in direct proportion and kind to the process stream. Removal i
of this suspended material makes the sample stream, to some degree, nonrepresentative.
Most modern process analyzers are designed to operate with a minimum of fouling when operated at proper flow rates. This reduces the need for filtration or frequent cleaning.
Some analyzers will require filtration to function properly, The in-line ion chromatograph, for example, will plug with suspended material within a short period of time if the sample stream is not filtered.
Filters, if required, should be either self-cleaning or easily cleaned. One type of filter that has been used with some success employs a cyclone type construction where flow is tangential to the body of the filter, and suspended material (usually more dense) is swept to a side stream where it can be discharged. 02871/DNS01
, 3.c l'
.2.6.3 ' Additional Considerations Summary L Y a
1.
Sample probes may beluseful. The requirement for' sample probes for condensate /feedwater-like sampling is
'3 f
debatable.
e 2.
In-line filters canl introduce nonrepresentativeness into the sample stress. Their use is not recommended except.
in'spwelalised applications.
3.0 CONCLUSION
S AND REC 0fftDIDATIONS 3.1-Representative sampling, especially at ' race levels (ppb and ppt) t 1s.a difficult proposition. The sampling system must be carefully designed, constructed, and operated. Shortcuts in any of these-areas can compromise the ability to get meaningful results from r
Process analysers or from grab samples.
UI
- 3.2 Roughing coolers should be lecated as close to the sampling root valve as practical. Forced air roughing coolers are not recommended, except in very special cases.
3.3 Pumped refrigerant tria coolers are preferred to bath type tria coolers.
It is not necessary to locate the trim cooler in close proximity to the sampling root valve or roughing cooler.
,3.4
.When sampling for trace level. impurities or suspended material, a D
nominal fluid velocity of 6 ft/s should be used. This fluid velocity should remain constant so that equilibrium conditions can-be reached and maintained. The Reynolds number should reflect turbulent, conditions.
When it is not possible to maintain continuous flow, a minimum 24-hour flush at a constant 6 f t/s is recommended as a goal.
3.5 Bypass streams may be used to keep flow conditions within accepted ranges. The bypass stream should be split as close to the sample outlet (analyzer, filter holder, etc.) as practien1.
3.6 Rotameters should be installed, when needed, to demonstrate that proper flowrates are being maintained. Rotameters should be installed downstream of process analyzers.
3.7 Sample systems should be constructed of 1/4-inch 0.D., 0.049 or 0.065-inch wall thickness seamless tubing. Type 316 stainless steel is preferred for most applications. Tubing runs should be as short as possible with a minimum number of bends, valves, and fittings.
3.8 Compression fittings are preferred over welded fittings. Cut tubing should be restored to its original inside diameter before joining. 02871/DNS01
_x_-___-__.
gmy m
- 3. x
?
'3,9' 1/
Is th) cb; ento of date tn d:cument the benefit from using a sample probe, a recommendation for the' installation of sampling' probes-cannot be made; 3.10 In-line filters are not recoassended for nomal sampling applications. ' Continuously self-cleaning sample filters are the preferred type, for tho' e rare ' instances when the sample must be g
[
filtered.-
s E
4.0 REFERENCES
1.
Dickinson, N.
L., Felgar, D. N., and Firsh, E. A., AN EXPERIMENTAL INVESTIGATION OF HYDRAZINE-0XYGEN REACTION RATES IN BOILER FEEDWATER, Proceedings of the American Power Confercnce, Volume XIX, 1957.
2.
EPRI NP-3589-SR-LD, BWR WATER CHEMISTRY GUIDELINES, April 1985.
7 3.1 EPRI NP-3402-SR, WORKSHOP: CORROSION PRODUCT SAMPLING FROM HOT WATER SYSTEMS, March 1984.
4.
EPRI NP-3704f TRANSFORMATIONS OF COPPER SPECIES DURING SAMPLING FROM PWR SYSTEMS, September 1984.
5.
Coulter, E.' E., ' SAMPLING STEAM AND WATER IN THERMAL POWER PIANTS, Presented at the Operating Symposium for Thermal Utilities Boiler Reliability, McMaster University, Hamilton, Ontario, May 1983.
6.
-Sundberg,,J.
J., SAMPLING PRACTICES,-1982 BWR Water Chemistry-Workshop, Bethesda, Maryland, May 1982.
I 7.
Beal, S.
K., THE EFFECT OF EROSION AND DEPOSITION ON THE SAMPLING OF ENTRAINED PARTICLES, AEC Research and Development Report WAPD-7M-1014, Bettis Atomic Power Laboratory, Pittsburgh, Pennsylvania, April 1972.
8.
EPRI NP-522, SURVEY OF CORROSION PRODUCT GENERATION, TRANSPORT, AND DEPOSITION IN LIGHT WATER REACTORS, March 1979.
9.
Private Communication, L. E. Eater, Radiological and Chemica'l Technology Inc. to P. Derenphal, NUS Corp., February 25, 1985.
10.
EPRI NP-4823, WATER CHDIISTRY AND RADIATION BUILDUP AT THE COMMONWEALTH EDISON COMPANY LASALLE-1 BWR, September 1986.
1 11.
EPRI NP-2149, CORROSILN PRODUCT TRANSPORT IN PWR SECONDARY SYSTEMS, I
December 1981.
12.
Eater, L. E., and Holloway, J. H., REVIEW OF CHEMISTRY SYSTEMS AT BROWNS FERRY NUCLEAR PLANT - FINAL REPORT, Radiological and Chemical Technology, Inc., April 1984.
)
13.
EPRI RP 2712, Draf t Report, UTILITY GUIDELINE MANUAL ON WATER I
CHEMISTRY INSTRUMENTATION AND CONTROL FOR FOSSIL ELECTRIC PLANTS, September 1986. 02871/DNS01 1
yr.. _..
.s
' ll^
'f. *
,,js
- 4..au:u a. M =d-~ ~ b h[$%hkYU
< g.
14.
INSTRIBEINT ENGINEER Comp r, Rad r, re.S' EAND300E, Liptak, B. G., Ed., Chilta. B
' ' ' [/.
..yly i., 1985.
'15.
ook EPRI NP-3789, CORROSION PRODUCT SUILDUP CN L
/
1985.
WR FUEL RODS, April 16.
f EPRI NP-4560, CORALT TRANSPORT AT TIE VERN 17.
Barcelona, N.
J., selfrich, J. 4.
ONT YANKEE BWR, May 1986 Volume 57, p. 460,1985.TUBINC EFFECTS ON GRO E. E.,
SANFLING Analytical Chemistry, 18.
APPLICATIONS FOR PWR SECONDARY
!~
S MONITORING SYSTEM WITH 19.
RY, May 1986 SECONDARY CYCLE AT FARLEY 1 AND 2M T., Memo Report, DISSOLVED GETCEN RED
%s. : dAuw Nfs
, Jelby, E. A., Eater, L. E., and R 20.
Nestel, W. A.3 CHENISTRY CONTROL FOR F0WER FIANTS l.
, Power Engineering April 1980.
, WATER e
0 x
' r.
W
^^ ~^~ '~
I" n