ML20205A661
| ML20205A661 | |
| Person / Time | |
|---|---|
| Issue date: | 03/31/1999 |
| From: | Bertodano M, Han J, Ishii M, Kelly S, Mi Y, Ransom V, Revankar S, Xu Y NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES), PURDUE UNIV., WEST LAFAYETTE, IN |
| To: | NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| References | |
| CON-FIN-L-2202 NUREG-CR-5578, NUDOCS 9903310040 | |
| Download: ML20205A661 (108) | |
Text
. _. _. _ _ _ _ _ _ _ _ _ _. _ _ _.. _ _ _ _ _. _ _ _ _ _ _ _ _ _ _ _ _.. _ _ _ _ _.. _ _ _ _ _. _ _ _ _. _ _ _. _ _ _ _ _ _ _ _ _. _.. _ _ -
i l
PU/NE-98-3 1
i i
i, Instrumentation for the PUMA Integral Test Facility g
l 4
l 1
Purdue University O
/
Y' O j
..-~
U.S. Nuclear Regulatory Commission i
B Office of Nuclear Regulatory Research l
Washington, DC 20555-0001
/
=
1
\\...../
h g,'0331 C;
R PDR
i f
NUREG/CR-5578 PU/NE-98-3
}
l l
Instrumentation for the l
PUMA Integral Test Facility i
i Purdue University 0
a/
,/ p U.S. Nuclear Regulatory Commission Office of Nuclecr Regulatory Research Washington, DC 20555-0001
/' %,
.Y 9903310040. -0331
~
I AVAILABILITY NOTICE L
Availability of Reference Materials Cited in NRC Publications NRC publications in the NUREG series, NRC regu-NRC Public Document Room lations, and Title 10, Enerpy, of the Code offederal 2120 L Street, N.W., Lower Level Regulations, may be purchased from one of the fol-Washington, DC 20555-0001 lowing sources:
< http://www.nrc. gov /N RC/PD R/pdr1. htm >
1 -800-397-4209 or lody 202-634-3273 1.
The Superintendent of Documents U.S. Government Printing Office Microfiche of mcst NRC documents made publicly P.O. Box 37082 available since January 1981 may be found in the Washington, DC 20402-9328 Local Public Document Rooms (LPDRs) located in
< http://www. access.gpo. gov /su_ docs >
the vicinity of nuclear power plants. The locations 202-512-1800 of the LPDRs may be obtained from the PDR (see previous paragraph) or through:
j 2.
The National Technical Informat'on Service Springfield, VA 22161-0002
<http://www.nrc. gov /NRC/NUREGS/
j
<http://www.ntis. gov /ordernow>
SR1350/V9/lpdr/html>
l 703 -487-4650 t
Publicly released documents include, to name a l
The NUREG series comprises (1) brochures few, NUREG-series reports; Federal Register no-t (NUREG/BR-XXXX), (2) proceedings of confer-tices; applicant, licencee, and vendor documents ences (NUREG/CP-XXXX), (3) reports resulting and correspondence; NRC correspondence and
{
from intemational agreements (NUREG/lA-XXXX),
internal memoranda; bulle; ins and information no-t (4) technical and administrative reports and books tices; inspection and investigation reports; licens-
[(NUREG-XXXX) or (NUREG/CR-XXXX)), and (5) ee event reports; and Commission papers and compilations of legal decisions and orders of the their attachments.
Commission and Atomic and Safety Licensing Boards and of Office Directors' decisions under Document 6 available from public and special tech-Section 2.206 of NRC's regulations (NUREG.
nicallibraries include all open literature items, such XXXX) as books, journal articles, and tansactions, Feder-al Register notices, Federal and State legislation, A single copy of each NRC draft report is available and congressional reports. Such documents as free, to the extent of supply, upon written request theses, dissertations, foreign reports and transla-as follows:
tions, and non-NRC conference proceedings may be purchased from their sponsori>1g organization.
Address: Office of the Chief Information Officer Reproduction and Distribution Copies of industry codes and standards used in a Services Section substantive manner in the NRC regulatory process U.S. Nuclear Regulatory Commiss are maintained at the NRC Library, Two White Flint e
Washington, DC 20555-0001 North, 11545 Rockville Pike, Rockville, MD E-mail:
< DISTRIBUTION @nrc. gov >
P0852-2738. These standards are available in the Facsimile: 301 - 415 -2289 library for reference use by the public. Codes and standards are usually copyrighted and may be A portion of NRC regulatory and technicalinforma-Purchased from the originating organization or, if tion is available at NRC's World Wide Web site:
they are American National Standards, from-
<http://www.ntc. gov >
- aerican National Standards institute All NRC documents released to the public are avail-11 West 42nd Street able for inspection or copying for a fee, in paper, New York, NY 10036-8002 microfiche, or, in some cases, diskette, from the
Public Document Room (PDR):
212 - 642 -4900 DISCt. AIMER This report was preparad as an account of work sponsored by any legal liability or responsibility for any third party's use, or the an agency of the United States Government. Neither the United results of such use, of any inforrnation, apparatus, product, or States Government nor any agency thereof, nor any of their em-proosss disclosed in this report, or represents that its use by ployees, makes any warranty. expressed or implied, or assumes such third party would not infringe privately owned rights.
NUREG/CR-5578 PU/NE-98-3 Instrumentation for the PUMA Integral Test Facility Manuscript Completed: March 1999 Date Published: March 1999 Prepared by S. T. Revank.r, M. Ishii, Y. Mi/Purdue Univ.
M. L. Bertodano, Y. Xu, S. Kelly, V. H. Ransom /Purdue Univ.
J. T. Han/NRC Schoolof NuclearL ineering Purdue University West Lafayette, IN 47907-1290 J. T. Han, NRC Project Manager s
Prepared for Division of Systems Technology Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Washington, DC 20555-0001 NRC Job Code L2202
,<>5**"*%
9
L l
I ABSTRACT The PUMA (Purdue University Multi-Dimensional Integral Test Assembly) facility is a i
scaled integral model of the Simplified Boiling Water Reactor (SBWR) designed by General l
Electric (GE) Nuclear Energy. The PUMA testing program includes integral tests such as the l
main steam line break tests and bottom drain line break tests as well as separate-effects tests of interest to reactor safety, and code model evaluation. The flow conditions in the PUMA tests require specialinstrumentation. In this report, discussions on those special instruments, including impedance metcr, modified magnetic flow meter, oxygen analyzer, conductivity l
probe, and vortex flow meter are presented. The impedance void-meter was developed to 1
l measure the average void fraction in pipes. The commercially available magnetic flow meter l
_ electrodes were modified to measure the liquid flow rate in two-phase flow. A multi-port gas sampling system using oxygen analymrs was developed for on line measurements of air concentration in steam at various components of PUMA. A commercial vortex flow meter was used to measure low steam flow rate in pipes. A conductivity probe was developed to measure the local void fraction in the reactor pressure vessel for water-steam at 1 MPa (150 psig) and 180 *C (356 *F) or below. These meters were tested and calibrated in the laboratory.
l r
1 1
i s
iii NUREG/CR-5578
TABLE OF CONTENTS ABSTRACT.......................................................................................
.... iii LI ST O F FIG U RES.....................................................................
... vii L I ST O F TA B L E S............................................................................................ i x A C KN O WL E DG M ENT............................................................................................. x N O M EN C L ATU RE.............................................................................................. x i l.. INTRODUCTI ON................................................................................. 1 - 1 1.1 Re fe re nc es.............................................................................
1-3
- 2. I M PEDANC E VOI D-M ETE R.....................................................................
2-1 2.1 Introd ucti on................................................................................................ 2-1 2.2 Impedance Void-Meter Design.................................................................. 2-2 2.3 Theoretical B asis........................................................................................ 2 -4 2.4 C ali b rat io n.................................................................................. 2-7 2.5 Flow Regime Identification...............................................
....... 2-9 2.6 Co nc l usio n s...........................................................................
2-11 i
2.7 Re fere n ces.............................................................................................. 2 -.12 l
l 1
3. MAGNETI C F LOW ICTER............................................................................. 3 - 1 3.1 I ntrod ucti on......................................................................................... 3-1 3.2 Hardware Design and Development........................................
3-2 3.3 Theoretic Aspects...........................................................................
. 3-2 3.4 C al i bratio n...............................................................
..... 3-5 3.4.1 Calibratt.n in Vertical Flow........................................................
3-5 3.4.2 Calibratit n in Horizontal Flow..................................................... 3-6 3.5 Co nc l usi o ns.............................................................................................. 3 -7 3.6 Re fere n ce s...................................................................
. 3-8
- 4. OXYG EN ANAL YZE R....................................................................
4-1 4.1 Introduction........................
4-1 4.2 Oxyge n Detec tors..........................................................................
4-1 4.2.1 Paramagnetic Detectors.....................................................
41 4.2.2 Catalytic Combustion Detectors.......................................................... 1-2 4.2.3 Electrochemical Oxygen Detectors................................
4-2 4.3 Extractive Zirconium Oxide Oxygen Analysis................................
4-3 4.3.1 Operation Principle................................
4-3 4.3.2 Analyzer Specifications............................................
. 4-5 4.3.3 Cali brati on.........................................................
..... 4-5 4.4 Measurement in PUMA Facility...........................................
4-7 4.4.1 Installatio n...................................................................
4-7 4.4.2 Analyzer Response........................................................
4-8 1
4.4.3 Sample Measurement in PUMA Test..................................
4-9 4
I i
v NUREG/CR-5578 i
TABLE OF CONTENTS (CONTINUED) 4.5 Re ferences....................................................
4-10
- 5. CONDUCTIVITY P ROB E...........................................................
. 5-!
5.1 I ntrod uc ti on................................................................................. 5 - 1 5.2 Probe Design........................................................
5-1 5.3 Electric Circuit Design..............................................................
5-1 5.4 Data Acquisition System Hardware................
... 5-2 5.5 Data Acquisition System Software.
5-2 5.6 Probe Calibration.........
5-4 t
5.7 Sample. Measurement...............................................
5-5 5.8 References.....................................................................................
... 5-6
- 6. VORTEX F LO W M ETER.............................................................................
6-1 6.1 I ntrod uc ti o n........................................................................
6-1 6.2 Principle o f Operation..........................................................
6-1 6.3 Foxboro Vortex Flow Meter.................................................
6-3 6.3.1 Specification................................................................
. 6-3 3
6.3.2 Calibration.........................................................................
6-3 6.4 Steam Flow Measurement in PUMA Test.....................................
6-3 6.5 Re fe re n c e s...................................................................
.....................6-5 7.
C O N C L U S I ON S.....................................................................
7-1 1
P i
1 NUREG/CR-5578 vi i
1 i
1 l
I LIST OF FIGURES Chapter 2 i
Figure 2.1 Impedance probe with ring-type electrodes........
.2-14 1
Figure 2.2 Functional block drawing of the circuit for impedance measurements.
.2-15 Figure 2.31mpedance circuit calibration..
.2-16 Figure 2.4 Layout of the test loop...
.2-17 Figure 2.5 Experimental system for cross-calibration in a 50.8 mm vertical pipe
.2-18 l
Figure 2.6 Cross-calibration between impedance void-meter and differential I
pressure transducer in a 30.8 mm vertical pipe Oi=0.04~2.00m/s,j,=0~20m/s)..
.2-19 Figure 2.7 Cross-calibration between impedance void-meter Lnd gamma densitometer in a 50.8 mm vertical pipe di.02-1.00m/s,j,=0-20m/s).............
.2-20
=0 Figure 2.8 Impedance signals of vertical two-phase flows in a 50.8 mm te' t section s
0,=0.04m/s).
.2-21 Figure 2.9 Probability distribution functions ofimpedance signals of vertical two-phase i
=0 I
ficws in a 50.8 mm test section di.04m/s) 2-22 Chapter 3 1
)
Figure 3.1 Magnetic flowmeter.
.3-9 i
Figure 3.2 Modified electrodes of the magnetic flowmeter........
.3-10 Figure 3.3 Calibration between the magnetic flowmeter and rotameters for
]
single-phase flows..
.3 11
]
Figure 3.4 Calibration between the magnetic flowmeter and orifice flowmeter i
for single-phase flows
.3-12
)
Figure 3.5 Measured superficial velocity versus void fraction at lower flow rates....
.3-13 i
Figure 3.6 Measured superficial velocity versus void fraction at higher flow rates....
.3-14 j
Figure 3.7 Cross-calibration results of the modified magnetic flow meter for single-phase flow..
.3-15 Figure 3.8 Cross-calibration results of the modified magnetic flow meter for two-phase flow
.3-16 i
Chapter 4 Figure 4.1 Flow of oxygen ions through a hot 7irconium oxide electrolyte producing a voltage difference across the eier
.4 11 Figure 4.2 Typical setup for OXA sw oxygen analyzer wiCi portable oxygen calibration tanks
.. 4-12 Figure 4.3 OXA 1000 oxygen ananzer installation on PUMA facility...
.4-13 Figure 4.4 Analyzer response to 3.5% oxygen concentration for line lengths of 4.3 m and 10.1 m
.4 14 Figure 4.5 Analyzer response to 10% oxygen concentration for line lengths of 4.3 m and 10.1 m.
.4-15 vii NUREG/CR-5578
LIST OF FIGURES (CONTINUED)
Figure 4.6 Analyzer response to 15% oxygen concentration for line lengths of 4.3 m and 10.1 m..
.4-16 Figure 4.7 Response of oxygen analyzer as a function ofline length and oxygen concentration
.4-17 Figure 4.8 Air concentration measured in PCCS supply lines during a bottom drain line break integral test for 16 hours1.851852e-4 days <br />0.00444 hours <br />2.645503e-5 weeks <br />6.088e-6 months <br />
.4-18 Figure 4.9 Air concentration measured in PCCS supply lines duringa bottom drain line break integral test for first 4500 seconds
.4 19 Chapter 5 Figure 5.1 Location ofinstrumentation ports in the reactor pressure vessel
.5-7 Figure 5.2 Schematic of a conductivity probe design....
.5-8 Figure 5.3 Schematic ofinstrument port.
.5-9 Figure 5.4 Schematic ofinstrument tube
.5-10 Figure s.5 Circuit diagram for a conductivity probe
.5-11 Figure 5.6 Air-water calibration loop
.5-12 Figure 5.7 Calibration curves for a conductivity probe..
.5-13 Figure 5.8 conductivity probe data for the bottom dain line break test (Probes located in the RPV chimney).
.5-14 Chapter 6 Figure 6.1 Karman vortices behind a bluff body.
. 6-8 Figure 6.2 Vortex shedding frequency as a function of Reynolds number
.6-9 Figure 6.3 Typical accuracy and K-factor linearity of the vortex meter for the calibrated range.
.6-10 Figure 6.4 Factory supplied calibration sheet for 19.1 mm vortex flow meter..
.6-11 Figure 6.5 Typical dimensions and installation of vortex flow meter on a pipe line.
.6-12 Figure 6.6 Comparison of steam flow rate measured by a vorte x flow meter (38.1 mm size) and condensed water flow rate measured by a magnetic flow meter..
6-13 i
NUREG/CR-5578 viii
-- _~-.
d LIST OF TABLES Tablel.1 Improvements on instrumentation and their application in PUMA facility......1-4 Table 4.1 Specifications of OXA 1000 Oxygen Analyzer................................
... 4-6 Table 6.1 Operating conditions and performance specifications of Foxboro vortex fl o w mete r................................................................
.6-6 i
Table 6.2 Steam flow rate limits for different size vortex flow meter......................... 6-7 5
l i
i l
l ix NUREG/CR-5578
ACKNOWLEDGMENT The authors wish to acknowledge the comments and support provided by Dr. Farouk Eltawila and other NRC Staff.
i NUREG/CR-5578 x
TT:
l NOMENCLATURE Latin Symbols A
pipe area B
magnetic flux density d
distance D
pipe diameter E
electric field intensity g
gravity G
impedance h
height i
electric current I-summation of the weighted input in a neuron j
electric current density j
total superficial velocity
/
- wall curve on pipe cross-section L.
lengtli n
direction normal to the wall V
counting number P
perimeter p;
pressure r
radial coordinate R.
reading of radiation y
electric current source U
electrical potential v
fluid velocity V
electrical potential difference W
width-X input of a neural network 2
axial coordinate alopg flow direction Greek Symbols a
void fraction A
differential e
pennitivity p
- permeability p
mass density a
electric conductivity Subscripts andSuperscripts 0
quantity for vacuum or gas phase 1
quantity for liquid phase f
fluid xi NUREG/CR-5578
l' i
i g
gas NOMENCLATURE (CONTINUED) i,j number mixture's quantity l'
m l
Mag reading of magnetic flow meter
~
number of counting
.N out output dimensionless quantity-1 r
t I
i I
i I
l
)
1 i
l EUREG/CR-5578 xii i
l
?
- 1. INTRODUCTION l
The PUMA (Purdue. University Multi-Dimensional Integral Test Assembly) test facility is a scaled integral model of the General Electric (GE) Nuclear Energy, Simplified Boiling j
Water Reactor (SBWR) [1.1-1.2]. This integral test facility was built under sponsorship of U.S. Nuclear Regulatory Commission (NRC) to reproduce the major thermal-hydraulic phenomena at low-pressure (<1033 kPa) [1.3]. PUMA has all the engineered safety systems and safety-grade systems of the SBWR. These systems include the following: (1) Automatic Depressurization System (ADS), (2) Gravity-Driven Cooling System (GDCS), (3) Passive j
Containment Cooling System (PCCS), and (4) Isolation Condenser System (ICS). The GDCS and PCCS are two new passive safety systems unique to the SBWR and do not exist in currently operating BWRs. The ICS is functionally similar to those in some current BWRs. Both the GDCS and PCCS are designed for low-pressure operation (less than 1033 kPa or 150 psia), whereas the ICS in the SBWR is capable of high-pressure operation as well (up to 7580 kPa or 1100 psia). The ADS is actuated at a prescribed water level in the reactor pressure vessel (RPV)~ to depressurize the RPV so that the GDCS can be activated to inject water to the RPV using the gravity head.
The PUMA testing program includes performing integral tests such as the main steam line break tests and bottom drain line break tests as well as condue:ing separate effects tests for understanding the key two-phase flow phenomena and for evaluating and modifying code models. It is very important to measure the mass and energy inventory and thermodynamic state in each component, and the boundary flow between various components.
To accomplish this the facility has over 400 instruments. The PUMA instruments include thermocouples, absolute and differential pressure transducers, nozzles, venturis, impedance void-meters, magnetic flow meters, vortex flow meters, oxygen concentration meters, local conductivity probes and video cameras. The data acquisition and control systems are based on a network of eight Intel 40486 personal computers.
The flow conditions in PUMA include steam-water two-phase flow, low steam flow (typically 10-100 kg/h in 38 mm pipe), and non-condensable gas and steam mixture flow.
Some of the pipelines have two-phase flow during blowdown from the RPV to the drywell.
1-1 NUREG/CR-5578 L
To measure the two-phase flow rate, the void fraction or quality is required. However, no standard commercial two-phase flow meters or void meters are available. Hence some instruments were developed at Purdue University specifically for the void fraction and liquid mass flow measurement. The impedance void-meter and the local conductivity probe are two examples. Certain im? '.ments such as the magnetic flow meter and the oxygen j
concentration meter had to be modified so that they could be used for ?UMA test conditions.
The following are the special instruments that required either in-house development or modi 5 cations.
The impedance void-meter was developed to measure the average void fraction in pipes.
The commercially available magnetic flow meters can only measure liquid flow rate in single-phase flow. They cannot measure two-phase flow rate. The magnetic flow meter electrodes were modified at Purdue University to measure the liquid flow rate in two-phase flow. The oxygen gas analyzer was used to measure the concentration of oxygen (or air) in the steam. A multi-port gas sampling system was developed for on-line measurement of air concentration in the steam in various PUMA components during a test. A commercial vortex flow meter was used to measure low mam flow rates in pipelines. A conductivity probe was developed to measure the local void fraction for conditions in the RPV at 1 MPa and 180 C.
These meters were tested in the Purdue laborato y and were calibrated. In the following chapters each of these instruments is presented in terms of the measurement principle, design and typical data. In Table 1.1 the summary of the instruments discussed in this report is given.
NUREG/CR-5578 1-2
1,1 References 1.1 GE Nuclear Energy, SBWR Standard Safety Analysis Repon, Repon No. 25A5113 Rev. A 1992.
1.2 Ishii, M., Revankar, S. T., Dowiati, R., Benodano, M. L., Babelli, I., Wang, W.,
Pokharna, H., Ransom, V. H., Viskanta, R., Wilmarth, T., and Han, J. T., Scientific Design of Purdue University Muhi-Dimensional Integral Test Assembly (PUMA) for GE SBWR, Purdue University Report PU-ME-94/1, U.S. Nuclear Regulatory Commission Repon NUREG/CR-63091996.
1.3 Han, J.T., Bessette, D.E., Shotkin, L.M., NRC Confirmatory Testing Program for SBWR, Proceedings of the Twenty-First Water Reactor Safety Information Meeting, Bethesda, Maryland October 25-27,1993.
l 4
1-3 NUREG/CR-5578 l'
l
_. _... _ __... _.-.___. _..___ _ _. _._ _ _ _._.m Table 1.1 Improvements on instrumentation and their application in PUMA facility Instrument Application Modification / improvement Impedance Void-Void fraction in steam / water Ring type electrode; new design; Meter flow at 1 MPa and 180"C in Purdue built and tested; cross-vertical and horizontal pipes calibration Magnetic Flow _
Two-phase steam-water The magnetic flow tube electrode Meter mixture flow rate in is modified into strip type horizontal and vertical pipes; clectrode for application ia flow regime identification stratified flow, cross-calibration Oxygen Air concentration'in steam-air Design, construction and testing Analyzer mixture in PUMA vessels of multi-port sampling system; in-situ calibration Conductivity Local void fraction in Single point needle type Probe steam / water system at 1 MPa electrode; new design; Purdue and 180 C built and tested; cross-calibration Vortex Flow Low range steam flow rate In-situ cross-calibration Meter NUREG/ Cit-5578 1-4
- 2. IMPEDANCE VOID-METER Knowledge of the void fraction in two-phase flow is important for both theoretical and experimental aspects of the research conducted in the PUMA facility. It is required for the physical description in two-phase flow analysis, especially in the development of closure
' laws for two-phase flow models such as the two-fluid model developed by Ishii [2.1]. To obtain the liquid superficial velocity during a two-phase experiment, the void fraction information, as well as the liquid velocity, is also desirable.
2.1 Introduction Many methods have been employed for void fraction measurements [2.2]. For example, Revankar and Ishii [2.3] and Miller and Mitchie [2.4] used conductivity and optical probes for local void fraction measurements. For the area-averaged void fraction, attenuation techniques were used by Chan and Bannerjee [2.5] and Eberle et al. [2.6], whereas for the volume-averaged void fraction the traditional " quick closing valves" method was used by Oliver and Hoon [2.7].
Among these techniques, the one based on the impedance measurement is the most suitable for the pressure and temperature conditions of PUMA experiments (1 MPa and 180 C).
An impedance void-meter has been. The information about void fraction is determined by measuring the electricalimpedance of two-phase flow, and applying the relationship between the void fraction and the impedance. The fast response of impedance void-meters makes it possible to use them for measurement during transient situation as well as in steady state.
Impedance void-meters have attractive economic features as well, since it is much easier to construct them than other void measurement meters.
The literature on impedance void fraction measurements before 1978 was reviewed l
extensively by Hewitt [2.2]. Cimorelli and Evengelisti [2.8], first presented a simple analysis of impedance measurements, and applied Maxwell's relation for bubbly flows. In their experiments, for vertical steam-water two-phase flows, total capacitance values on a 18mm ID and 1485mm long test section were measured. According to their results, Maxwell's 2--I NUREG/CR-5578
i I
relation appears to be reliable when the void fraction is less than 40%, almost covering the entire bubbly flow regime.
. It has been realized that the relationship between void fraction and impedance depends on flow regimes. To overcome the difficulty of the dependence on flow regimes, a number of i
alternative probe designs have been investigated. Some promise seemed to be offered by the development of Mcrilo et al's six-clectrode probe (2.9], in which a rotating electrical field was generated and distributed throughout the probe volume. However, this six-clectrode i
probe did not show any advantage over a simple two-strip-clectrode probe for venical flow measurements [2.10].
i Among designs of impedance void-fraction meters, the most successful is the non-l intrusive probe with ring-type electrodes mounted Hush to the tube wall [2.11,2.12]. This design concept was employed in the present investigation and was applied for both vertical and horizontal Dows.
For application in steam 4 vater two-phase flow in vertical and horizontal pipe flows, a prototype impedance void-meter with two ring-clectrodes in a stainless-steel shcIl has been developed. This impedance void-meter can function well at a pressure up to 1.5MPa and a 1
temperature up to 200 "C. provided the temperature effect on void measurements is compensated properly. However, void measurements with this design still depend on flow regimes so the meter has to be cross-calibrated in order to establish the relationship between l
impedance and void fraction in each flow regime. The output signals from the impedance void-meter carry some information on flow regimes.
By analyzing these signals and evaluating their probability distribution functions, flow regime identification was performed for vertical flows.
For horizontal flows, flow regimes can be determined based on experimental conditions, and justified by futput signals and tlcir probability distribution functions. In the following sections, thu design of this impedance void-meter, experimental study, cross-calibration and flow regime identification from impedance signals are presented.
2.2 Impedance Void-Meter Design The impedance void-meter consists of two major components: a probe and an electronic circuit. The design of the impedance probe was based on the requirement of pressure up to IMPa and temperature up to 180 "C.
For good mechanical and non-corrosive properties,
'NUREG/CR-5578 2-2
stainless steel was chosen as the material for the electrodes and shcIl of the probe. Tellon was used as an electrical insulator between c!cetrodes, and between the electrodes and the shell. The structure of the probe is shown in figure 2.1. The two ring-clectrodes were 11ush mounted on the inside wall and insulated from the stainless steel shell. These ring-type electrodes allow us to measure the void fraction from 0 to 1 for horizontal Hows as well as vertical flows. A series of o-rings were installed in gaps between electrodes and insulators.
The probe was tested with a steam-water system at 1.2 MPa pressure and 180 C temperature condition. A total of five probe si7cs were designed and constructed. Their internal diamelets are: 12.7 mm,19.1 mm, 25.4 mm, 38.1 mm, 50.8 mm and 76.2 mm. For all of them, the height of the dectrodes is 12.7 mm and the distance between two electrodes is 50.8 mm.
As shown in figure 2.2, the impedance circuit consists of several parts including a buffer, a cunent-voltage amplifier, a demodulator, a low-pass filter and a voltage amplifier.
Normally a linear relation is desired between input and output signals. The output of the circuit is designed to be directly proportional to the measured impedance. A calibration of this circuit was conducted with a series of resistors, and it was found that the integral nonlinearity (INL) was less than 1.5%, as shown in figure 2.3. An alternating current is supplied at 100kHz to the electrodes on the impedance probe. This avoids the double layer effect [2.13]. The void-meter is versatile such that it could be adapted for void fraction measurements of multiphase flow with liquids of different electrical conductivities and different probe sizes by adjusting the gain of amplifiers of the circuit.
In impedance-based void measurements, some errors might be introduced. These errors can be avoided or reduced as discussed below.
(1) Effects on the Liouid Electrical Conductivity The conductivity of water varies with the temperature. This effect increases the difficulty of void fraction measurements. In order to account for the temperature effect on liquid conductivity, the relationship between temperature and water conductivity was measured from the lowest experimental temperature to the highest experimental temperature (180 C) before each integral experiment.
Based on this relationship, the measured impedance readings were converted into dimensionless readings at a reference temperature.
2-3 NUREG/CR-5578
The reference temperature may be any temperature between the lowest experimental temperature to the highest experimental temperature. Because the cross-calibration of the impedance void meter was conducted at room temperature, it had to be assumed that the relationship between the void fraction and impedance at the reference temperature is the same as that at the room temperature.
Conductive chemicals are added to the water to increase the conductivity. The chemicals added are Morpholine and Ammonia Hydroxide. The water conductivity was kept around 360 Si before each experiment. Small portion of these chemicals might vaporize during experiments and would change water conductivity. However, it was difficult to trace the conductivity change during the experiments. It was assumed that the water conductivity was constant during the experiments. This assumption should be justified based on the experimental results.
(2) Electronics Drift As discussed previously, the output of the impedance measurement circuit is proportional to the amplitude of the carrier signal from a function generator, values of feedback resistors of voltage amplifier, and conductance of a two-phase mixture. Therefore, the drift of the i
amplitude of the carrier signal and values of feedback resistors should be minimized. For j
this, feedback resistors with 0.1% temperature coefficient were used. Typically tests were conducted after keeping the instruments on for more than one hour to stabilize the circuit including the function generator.
(3) Mechanical Installation Error It should be noted that mechanical errors could be introduced if the impedance probes were not installed with proper alignment.
In order to reduce mechanical errors, the impedance probes were installed with extreme care and properly aligned with the test section.
2.3 Theoretical Basis In practical casts, the impedance of a two-phase mixture is dominated by either conductance or capacitance depending on properties of the two-phase fluid and carrier frequencies of impedance circuits. In the present tests, since the carrier frequency was not NUREG/CR-5578 2-4
4 larger than 100kHz, the water acts as conductive liquid. Hence, conductance measurements were performed.
Normally conductance of a two-phase mixtun: depends on conductivitics of the two-i phase fluid and the distribution. If conductivity can be assumed constant, the relationship between impedance and void fraction can be predicted for some ideal distributions such as uniform dispersed flow and concentric separated flow.
2 The dimensionless impedance of a two-phase mixture, G*, is defined as G"'- G, G.
(2.1)
=
G - G,,
o where G,,, is the measured impedance value between two electrodes, Go is the impedance value when the pipe is full of conductive liquid, and G is the impedance value when the pipe i
is full of non-conductive gas.
' For a dispersed flow, especially bubbly flow, the electrical field between electrodes is distributed in the two-phase mixture. Hence the impedance of the two-phase mixture i
depends only on the conductivity of the mixture.
In such a case, the dimensionless impedance of the two-phase mixture, G', can be predicted by Maxwell's relation [2.14],
G. = 1 3a (2.2) 2+a Maxwell's relation is based on the assumption that the non-conductive dispersed phase is composed of non-interacting and equal-size spheres, and distributed uniformly in the continuous phase fluid, which is consistent with the characteristic of bubbly flows when the void fraction is low.
i For an annular flow, a void fraction measurement is no more than a film thickness measurement, if the droplets entrained into the gas phase are ignored. In this case, ring-tv-electrodes of the impedance probe exhibit more advantages than other type of electrodes.
Theoretical analysis of the behavior of a probe with parallel rectangular electrodes separated by a short distance was developed by Coney [2.13]. Coney computed the impedance between two rectangular electrodes by the conformal transformation and obtained 2-5 NUREG/CR-5578
K(m)
G., = cr l, K(1 - m),
(2.3) o where K(m) = [, (1-msin,)-a5dr is the complete elliptic integral of the first kind, La is 2
the effective length of electrodes, and
'nW' 2
sinh 26,,
r (2.4) m=
r n(W + D)3,
2
.smh 26,
(
where W is the width of electrodes, D is the distance between electrodes and 6, is the (equivalent)ligtiid film thickness.
Andreussi et al. [2.11] extended the above analysis to the impedance probes with ring-type electrodes by introducing an equivalent thickness, 6, =
p,.
(2.5) where Pi is the wetted perimeter, A is the flow area. For annular flows, the equivalent i
thickness is 6, = D*(1 - a)/4.
(2.6)
If it can be assumed that the liquid film attached on an electrode has the same potential as the electrode's, the dimensionless impedance of the two-phase mixture can be approximately expressed as G*=l-a.
(2.7)
This implies that, for very thin liquid films, the dimensionless impedance measured by the impedance void-meter can be related linearly to the void fraction.
NUREG/CR-5578 2-6
l
_2.4 Calibration The cross-calibration of the' impedance void meter was performed in both 50.8 mm and 76.2 mm vertical and horizontal test loops, which were constructed and instrumented especially for this investigation. The layout of the verticalloop is shown in figure 2.4. In the
- cross-calibration experiments, de-ionized water and air were used as liquid and gas, respectively.
A pump drives the liquid flow, while gas is discharged either from a pressurized tank or from a compressor through an accumulator. The two-phase fluids are injected into a mixing chamber and flow upward through a reducer into a test section.
Figure 2.5 shows major parts of the vertical test section as installed. Probes or detectors l
ofIhe impedance void-meter, gamma densitometer and differential pressure transmitter.wcre installed on the test section. The gamma source and Nal(TI) detector of the gamma densitometer were set up downstream of the impedance void-meter. Two testing taps were loca"<1 on the llanges for differential pressure transmitters, to convert pressure drops to global void fraction. Both gas and liquid superficial velocities can reach speeds over 20 m/s.
The data acquisition system included.a 486 PC, DAS-20 A/D board and its data acquisition software. All programs of data acquisition were written in C language under the DOS environment.
The purpose of cross-canbration was to establish the relationship between impedance and void fraction for each type of probe geometry with different reliable void fraction measurement instruments. For each condition of cross-calibration, impedance values were sampled at 200 Hz in 60-second duration. In this investigation, the cross-calibration of the impedance void-meter was conducted with a differential pressure transmitter and a gamma l
densitometer. Both of these methods have been established as reliable methods for void fraction measurements [2.5].
]
The total pressure difference in a vertical test section is the summation of the friction loss, acceleration, and gravitational pressure drop. In the situation of very low flow rate and constant area of the test section, both friction and acce!cration terms are neglqible compared L
with the gravitational term. Also, since the density of air is much sn. aller than that of water i
l_
in the calibration condition, the total pressure drop measured along the test section is 2-7 NUREG/CR-5378
AP,,, = AP,,,, = p.gh = pf(1 - a )gh,
(2.8) where h is the distare between the two pressure taps. Therefore the dimensionless pressure,
- i.e., the total measured pressure divided by hydraulic pressure generated by single phase liquid, is equal to liquid holdup; i.e.,
(2.9)
Thus, there is a linear relationship between the void fraction and dimensionless differential pressure. The differential pressure transducers were used for the cross-calibration at relatively low flow rate conditions.
For all flow rates, the area-averaged void fraction measurements were performed by gamma densitometer. The range of void fraction covered by this method is from 0 to 1. The relationship between readings of gamma densitometer and void fraction depends on flow regimes [2.5]. Typically for bubbly flows, we have 4
In(R,,, / R,)
a=
(2.10) in(R / R ),
)
i where Ro, R, and Rm are gamma densiaeter's readings of liquid phase and gas phase and i
two-phase mixture, respectively. For slug, churn and annular flows, we have R,,, - Ri a = R - R,,
(2.11) e The relationship between void fraction and impedance of two-phase flow was cross-calibrated in vertical and horizontal test sections. Cross-calibration of impedance meters in several flow regimes was carried out against a differential pressure transducer and a gamma densitometer.
For bubbly. flows, agreement between experimental results and their theoretical prediction was satisfactory as shown in figures 2.6 and 2.7. This indicates that the Maxwell's relation is valid for the bubbly flow regime. Dimensionless impedance of t'vo-phase mixture was approximately correlated to the liquid holdup (1-a), for slug and churn flows. As shown in figure 2.7, Coney's theoretical prediction agreed very well with experimental data of annular flows.
NUREG/CR-5578 2-8
_. _.~ _ _
t 1
1 In a horizontal test section, cross-calibration between impedance void-meter and a i
gamma densitometer was carried out for stratified and intermittent flows. It has been shown that the experimental results have the same trend as the theoretical prediction by Andreussi et al[2.11]. Fifth order polynomial functions were fitted to the experimental data so that void fraction can be determined from impedance readings. For the 50.8 mm probe, the Ibnction was a = -5.5378G + 14.8310G" - 11.9025G 2 + 2.390G 2 -0.7398G* +1.
(2.12)
Similarly to vertical flows, for horizontal flows, Maxwell's relation and Coney's prediction were applicable to bubbly and annular flows, respectively.
l 2.5 Flow Regime Identification i
l Vertical two-phase flows are often categorized into groups sharing similar characteristics.
Most flow conditions could be classified in the categories of single i quid phaseflow, bubbly i
flow, slug flow, churn flow, annularflow and single phase gas flow. It has been shown by previous literature and present experiments that the determination of void fraction with impedance technique is strongly dependent on geometric distributions of two-phase mixture, i.e., flow regimes.
Many researchers have been working on developing non-intrusive methodologies for l
flow regime identification. Jones and Zuber [2.15] depicted flow regimes with void fraction l
fluctuations. To obtain fluctuations of void fraction in an air-water two-phase flow through a rectangular channel, they employed a linearized X-ray rneasurement system. The probability density function (PDF) of void fraction was suggested as an objective and qualitative flow pattern discriminator for the three dominant flow regimes, i.e., hubbly, slug and annular flow.
Some researchers have employed a differential pressure technique to estimate flow regimes from the configuration of probability density function of differential pressures [2.16,2.171 j
However, the technique applies to the case where the friction and acceleration losses are negligible compared with the gravity pressure losses. In addition, there may be some difficulties related to the presence of two-phase fluids; for example, the presence of gas trapped into the pressure lines [2.18].
2-9 NUREG/CR-5578
In this work, vertical Dow regimes were identified prior to void fraction measurements through impedance signals and their 11uctuations as illustrated in figure 2.8. Assuming that the geometric distribution of two phases is a random process, the probability ' density functions (PDF) from the impedance void-meter can be obtained in the time domain, as shown in Ogure 2.9.
Basically, in a segment of a two-phase pipe flow, when various small spherical interfaces between dispersed bubbles and continuous liquid are scattered randomly, a bubbly flow can i
be recognized. Normally there is a maximum limit of area-averaged void fraction. Area-averaged void fraction fluctuates around an averaged value and accordingly, the maximum impedance value might not be less than 0.4 as predicted by Maxwell's relation (see figure 2.8). Considered as a Gaussian distribution, there is only one peak appearing at the higher end of PDF of a bubbly flow.
Bubbly flow will change into slug flow when void fraction increases above 25-40%
Geometrically, slug flow can be viewed as a combination of bubbly flow and annular flow.
A typical unit of slug flow consists of a Taylor bubble, similar to downward annular flow, and a liquid slug, in which small bubbles are surrounded by continuous liquid. Impedance
' signals of slug f9w in the time domain are shown in figure 2.8. At any instant, signals are similar to either annular or bubbly flow. This is the reason that there are two peaks appearing in the PDF of an slug flow as shown in figure 2.9. Churn flow can normally be considered as a transition region between slug and annular flow. Transitions between churn flow and other flows are not as clear as transition between bubbly flow and slug flow. Stable Taylor bubbles can not exist and liquid bridges are broken due to higher gas flow rates. In figure 2.8, the signals of impedance void-meter exhibit this complex phenomenon rather clearly.
Compared with slug flow, in figure 2.9, the PDFs of churn flow show that the peak associated with liquid slug is lower and eventually disappears as gas flow rate increases.
For annular flow, signals of the impedance void-meter, are opposite to those of bubbly flow.' Geometrically, they have completely different distributions. For relatively low flow rates, annular flow is a separated flow where the gas phase flows through the pipe center and only a thin liquid film is attached on the wall of the pipe. Corresponding to a minimum void NUREG/CR-5578 2-10
-i--
9
..- +
fraction larger than 0.8, the maximum impedance is not lager than 0.2 and accordingly there is only one peak appearing between 0 and 0.2 in the PDF of annular flow.
Based on the discussion above, for vertical Dows,11ow regimes can be characterized and j
identified through impedance signals and their PDFs. This can facilitate impedance-based void measurements significantly.
l For horizontal Dows, How regimes can not be identined based on a signal form the impedance void-meter with the ring-type electrodes. For exampic, for a low void condition, the typical impedance signal of bubbly flow is similar to the impedance signal of a stratified Dow; and for a low void condition, it is difficult to distinguish stratified flow and annular flow from cach other based on the impedance measurement..
2.6 Conclusions Void fraction is one of the fundamental quantitics necessary to describe the characteristics ofliquid-gas two-phase flow. Void fraction measurements with the electrical impedance technique have been proven attractive for many applications because of their fast response, non-intrusiveness and relative simplicity of operation. A number of impedance
)
probe designs have been developed in recent years. There is noticeable sensitivity of l
impedance probes to different now regimes, flow orientations - and fluid temperature conditions.
1 In order to overcome the difficulty mentioned above, a very simple ring-type probe was developed and tested. Cross-calibration between the impedance technique and other void measurement methods have been performed so that the relationship between impedance and void fraction can be correlated in each How regime for both vertical and horizontal flows. In addition, flow regimes of vertical Hows were identified from the characteristics of impedance signals.
'1
~
2-11 NUREG/CR-5578
r.
2.7 References 2.1 M. Ishii, Thenno-Fluid Dynamic Theory of Tuv-Phase Flow. Eyrolles, Paris.
Scientific and Medical Publication of France, New York (1975).
2.2 G. F. Hewitt, Measurement of tuv-phase flow parameters. Academic Press, New York (1978).
2.3 S. T. Revankar and M. Ishii, Local interfacial area measurement in bubbly flow, Int.
J. Heat Mass Transfer 87,453-468 (1992).
2.4 N. Miller and R. E. Mitchie, Measurement of local voidage in liquid / gas two phase l
I How systems using a universal mobe, J. Br. Nucl. Energy Soc. 9,94-100 (1970).
2.5 A. M. C. Chan and S. Banerjee, Design aspects of gamma densitometers for void fraction measurernents in small scale two-phase flow, Nucl. Instr. Meth. 190, 135-148 (1981).
i 2.6 C. S. Eberic, W. H. lxung, M. Ishii and S. T. Revankar, Optimization of a one-shot ramma densitometer for measuring arca-averaged void fractions of gas-liquid flows in
- e. rows pipelines, Meas. Sci. Technol. 5, i 146-1158 (1994).
2.7 D. R. Oliver and A. Young Hoon, 'Two-phase non-newtonian flow. Part 1: Pressure drop and hold-up," Trans. Instn. Chem. Engrs. 46, T106-T115 (1968).
2.8 L. Cimorelli and R. Evangelisti, The application of the capacitance method for void fraction measurement in bulk boiling conditions, Int. J. Heat Mass Transfer 10,277-288 (1957).
2.9
% Merilo, R. L Dechene and W. M. Cichowlas, Void fraction measurement with a i
1 rotating field conductance gauge." 1. Heat Transfer, Transactions of ASME 99,330-332 (1977).
J 2.10 J. M. Delhaye, C. Favrcau, J. M. Saiz-Jabardo and A. Tournaire, Experimental investigation on the performance of impedance sensors with two and six clectrodes for area averaged void fraction measurements, ANS Proc.1987 National Heat Transfer C<mf.,234-239 (1987).
1 I
NUREG/CR-5578 2-12 1
2.!!
P. Andreussi, A. Di Donfrancesco, M. Messia, An impedance method for the measurement of liquid hold-up in two-phase flow, Int. J. Multiphase Flow 6,777-787 (1988).
2.12 N. A. Tsochatzidis, T. D. Karapantsios, M. V. Kostoglou and A. J. Karabelas, A conductance probe for measuring liquid fraction in pipes and packed beds, Int. J.
Multiphase Flow 18,653-667 (1992).
2.13 M. W. E. Coney, The theory and application of conductance probes for the measurement of liquid film thickness in two phase flow, J. Phys. El Scient. Instrum. 6, 903-910 (1973).
2.14 J. C. Maxwell, A Treatise on Electricity and Magnetism, Clarendon Press, Oxford (1881).
2.15 O. C. Jones and N. Zuber, The interrelation between void fraction fluctuations arx1 4
flow patterns in two-phase flow, Int. J. Multiphase Flow 2,273-3% (1975).
2.16 N. K. Tutu, Pressure iluctuations and flow pattern recognition in vertical two phase gas-liquid flows, Int. J. Multiphase Flow 8,443-447 (1982).
2.17 G. Matsui, Identification of flow regimes % vertical gas-liquid two-phase flow using differential pressure fluctuations, Int. J. Mun. phase Flow 10,711-720 (1984).
2.18 O. C. Jones and J. M. Delhaye, Transient and statistical measurement techniques for two-phase flows: a critical review, Int. J. Multiphase Flow 3,89-116 (1976).
2-13 NUREG/CR-55,8
I i
}..]
Stainless Steel Shell Teflon Insulator Stainless Steel Electrode Figure 2.1 Impedance probe with ring-type electrodes.
NUREG/CR-5578 2-14
F 1
i From a function i
generator To electrodes f
p BUFFER CURRENT-VOLTAGE AMPLIFIER DEMODULATOR c
To A/D converter l
y LOW-PASS FILTER
=
OUTPUT AMPLIFIER z
i i
Figure 2.2 Functional block drawing of the circuit for impedance measurements.
2-15
~_.
6.0 5.0
- l-4.0
.0 8
2.0 1.0 0.0 0.0 0.1 0.2 0.3 0.4 CONDUCTANCE ( 1/kohm )
Figure 2.3 Impedance circuit calibration.
NUREG/CR-5578 2-16
.... -....~-
. ~............ _.
h AIR N..
LJHRRIM.M
- +h-g-- REHIEGALK3E lESTSLTON
{
/
WATERR.1ER CDtMRINTOMNNEL
- ARROWAEIIR FED VAliR w
hREBIECALI3E 7
W-- +
N nuamIDARTANK 1
ll==m m
N
,/
ORRT LDWRARRINLi#'
h(-
y ARINLETFRHUECALGE m
b omm Figure 2.4 Layout of the test loop.
2-17 NUREG/CR-5578
M1 Gamma Gamma Amplifier
- ___jp, s:.,
_g___
Sointe'-
Detector
/SCA 0.254m N
s f
Impedance Computer
'Cirhuit A/D' l.778m 1.168m v
v i
Water I
DP Air
)
Figure 2.5 Experimental system for cross-calibration in a 50.8mm vertical pipe.
NUREG/CR-5578 2-I o
l l
1.0 0.8 G
= *.,
cg Eq. (2.7) 8.
0.6 5
Maxwell's relation mee li 0.4 O
c i
e E
02 g:;
0.0 0.0 0.2 0.4 0.6 0.8 1.0 Void Fration ( DP )
Figure 2.6 Cross-calibration between impedance void-meter and differential pressure transducer in a 50.8mm vertical pipe (ji=0.04~2.00m/s,jg=0~20m/s).
2-19 NUREG/CR-5578
_ _ ~... _ _ _ _.. _ _ - _ _ _ _ _ _ _ _ _..
1.0 8
,5 0.8 g
E
~
0.6 m
g Max ' ell's relation Coney's theory d
c
,o 0.4 eee E
Eq. (2.7) 5 0.2 0.0 0.0 0.2 0.4 0.6 0.8 1.0 Vold Fraction ( Gamma Densitometer )
l Figure 2.7 Cross-calibration between impedance void-meter and gamma densitometer in a 50.8mm venical pipe (ji=0.02-1.00m/s,jg=0~20m/s).
NUREG/CR-5578 2-20
l 1.0
_ _ _ =, _ - - -
Bubbly Flo
.0i 0.8 o
E Churn Flow v
.6 Slug Flow 0
E l#.O#-4 h
~'
^
\\
ae, E 0.4
- ~
1
}
k f 0.2 Annular Flow 0.0 O
1 2
3 4
5 Time (sec)
Figure 2.8 Impedance signals of vertical two-phase flows in a 50.8mm test section (ji=0.04m/s).
2-21 NUREG/CR-5578
I 0.6 7, --
I Bubbly Flow )
h--
... -..... -....h. $..
T I
Annular Flow Iil
..I'..
0.4 u.
O S
I
.... y..
'I I
I I
0.2
,.L I
f I
I A
- -- " Churn Flow t'-
4;i-
'l
, Slug Flow 0
.x l '/
't
\\
0 0.2 0.4 0.6 0.8 1
Dimensionless impedance, G' l
i
{
Figure 2.9 Probability distribution functions ofimpedance signals of vertical two-phase flows in a 50.8mm test section (ji=0.04m/s).
NUREG/CR-5578 2-22
- 3. MAGNETIC FLOWMETER Reliable and accurate information about liquid flow is very important in both industrial practices and laboratory research that involve two-phase flow. However, most flowmeters, designed for single-phase flow, are often not suitable for two-phase flow measurements because of poor response and poor accuracy. In the PUMA experiments, the application of magnetic flowmeters was extended to the measurements of the liquid superficial velocity and liquid velocity for both vertical and horizontal two-phase flows.
For vertical flow measurements, commercial flowmeters were used. For horizontal flow measurements, a conventional magnetic flowmeter was modified to cope with the nonsymmetrical phase distribution. The flow regime dependence of the magnetic flowmeter was overcome through I
cross-calibrating it with other liquid flowmeters in vertical bubbly, horizontal stratified and intermittent flows. In this section, development of the magnetic flowmeter is discussed in detail, and the result of cross-calibration is presented.
3.1 Intmduction The electromagnetic flowmeter is based on the principle that material moving through a magnetic field experiences an electromotive force which is perpendicular to both the imposed magnetic field and the direction of motion of the material and satisfies the right hand rule, as shown in figure 3.1. Investigations using magnetic flowmeters for single-phase flows were conducted during the 60's and 70's [3.1,3.2]. Now the magnetic flowmeter has become one of the most reliable commercialinstruments for single-phase flow measurements. Even so, its application in two-phase flow conditions is still under development. Hori et al. [3.3]
conducted the preliminary experimental study on the performance of a transverse field magnetic flowmeter and showed indirectly that the output of the magnetic flowmeter is proportional to the mean liquid velocity. Bernier and Brennen {3.4], Murakami et at [3.5]
and Knoll [3.6] also performed similar investigations in two-phase flow conditions. It has been shown from their results for vertical flow that the accuracy of the magnetic flowmeter depe.~!s on flow regime. For bubbly flows the error of the indicated liquid flow rate was 3-1 NUREG/CR-5578
within 5% of the actual value. While void fraction is larger than 0.3, the difference between the flowmeter indication and the actual flow rate increased drastically.
No report about the application of the magnetic flowmeter in horizontal two-phase flows has been found. For horizontal flows, commercial magnetic flowmeters with point -
type electrodes can not function properly because non-conductive phase is not uniformly distributed and liquid velocity profile is non-axisymmetric. In this research, a conventional magnetic flowmeter with point-type sensors was modified into one with strip-type sensors.
The modified magnetic flowmeter can perform liquid flow rate measurements over a wider range of two-phase flow conditions.
3.2 Hardware Design and Development Normally commercial magnetic flowmeters are equipped with two point-type sensors so that they are only applicable to developed single-phase liquid flow and vertical bubbly flow measurements. To apply this technique under two-phase flow conditions, point-type electrodes in commercial magnetic flowmeters were replaced with strip-type electrodes.
Original electrodes were made of stainless steet Since the modified electrodes were located
)
in magnetic fields, brass was used as the electrode material to avoid any distortion of the magnetic field. In order to protect these electrodes from corrosion, they are coated by gok!
film of 0.01 mm thickness. The schematic of a strip-type electrode is shown in figure 3.2.
The geometrical parameters of electrodes for several pipe sizes are listed in Table 3.1. To minimize intrusiveness, the maximum thickness of these electrodes was limited to 2 mm.
3.3 Theoretical Aspects When a magnetic flowmeter with point-type sensors is applied to a single phase flow, potential differences appearing on the wall can be related to flow pararmters such as the liquid velocity of two-phase flow according to Maxwell's electromagnetic theory [3.2]. To analyze the magnetic flowmeter, some assumptions should be made based on experimental conditions. The effect ofinduced currents on the magnetic field distribution and the effect on 4
the fluid ~ motion of the electromagnetic body forces associated with these currents are NUREG/CR-5578 3-2
negligible since de-ionized water and air have been used in our experiments. % ionized water and air can be considered as non-magnetic fluids, of which the permeabihu,
, is the 4
same as the vacuum permeability,4xx10. The conductivity of water, o, is isotropic and is not affected by the magnetic field or liquid motion. Currents due to convection of electric charges by fluid motion can be neglected because the amounts of such charges in the liquid are very small.
By ohm's law, the induced current, j, in the magnetic flowmeter can be expressed as j = a(E + v x B).
(3.1) l where v x B is the electromagnetic field induced by the fluid motion, E is the electric field j
in the stationary coordinate system due to charges distributed in and around fluids and to any variation of the magnetic field in time. Maxwell's equations relate the electric field to the magnetic field by dB VxE= 7, (3.2) and gg, VxB=a j+c (3.3) o where po is the pern, ability of vacuum and to is the permitivity of vacuum; both are constants at a low frequency of the magnetic field B. The skin-effect due to self-inductance is negligible; i.e., the right hand side of Eq. (3.2) is zero.
Also, the Maxwell term c (dF/dt)in Eq. (3.3), is much smaller than the conduction current j, and hence is negligible o
unless the frequency is very high. Based on these assumptions, Eqs. (3.2) and (3.3) can be reduced to VxE=0, (3.4) and VxB= oj.
(3.5)
According to Eq. (3.4), the electric field E is irrotational Hence, an electric potential U can be defined as E = -V U.
(3.6) 3-3 NUREG/CR-5578
From Eq. (3.5), if the divergence is taken on both side, we have:
V.J = V V xB = 0.
(3.7)
Combining the equations above, we obtain the governing equation for a magnetic flowmeter in the form of Poisson's equation, V2U = V.y xB.
(3.8)
Eq. (3.8) hokis good even for turbulent flows. Eq. (3.8) indicates that the liquid velocity can be related to the induced electrical potential distribution around the flow field where a magnetic field is imposed. For liquid single-phase flows, to obtain the averaged liquid velocity, we can simply measure an electrical potential between two diagonal electrodes perpendicular to both the magnetic field and flow direction. If we treat this as a two-dimensional problem, the governing equation of the magnetic flowmeter can be reduced to V'U = B av'.
(3.9)
Solving for the potential difference between the diagonal electrodes yields AU = DBvf.
(3.10)
It is shown from Eq. (3.10) that, for vertical single-phase flows, the averaged liquid velocity, vf, is proportional to the potential difference between the point-type diagonal electrodes, used in commercial magnetic flowmeters.
Due to the complexity of the distribution of the non-conducting gas phase and the velocity profile of the conducting liquid phase, it is very difficult to analytically predict the liquid phase velocity from the output of the magnetic flowmeter under most two phase flow conditions, except for several simple or ideal two-phase flow conditions. The sensitivity analysis of the magnetic flowmeter for ideal two-phase flow conditions such as vertical j
bubbly and annular flows was performed by. Wyatt (3.7].
Normally for a vertical bubbly flow, the void fraction is lower than 0.3. Bubbles are small in comparison with the flowmeter diameter, and they are distributed axisymmetrically.
Globally, the induced electrical field distribution within the liquid phase will not be affected i
NUREG/CR-5578 3-4
by the presence of gas phase. It is obvious that the potential difference between the diagon21 electrodes is still proportional to the mean liquid phase velocity, given by Eq. (3.iG).
j Combining information about the void fraction, the liquid superficial velecity can be obtained through the expression i
2 AU
--i.= (1 -a,1 (3.11) i f
where the upper bars indicate time-averaged quantities for bubbly flow. Eq. (3.11) implies that the void fraction should be independently measured if the superficial velocity of liquid phase is preferred instead of the mean liquid velocity. Several experiments have been conducted which support the discussion presented above [3.4,3.5 and 3.6].
For the modified magnetic flow meter with strip-type electrodes, it is formidable to predict its performance under two-phase flow condition. The relationship between the liquid velocity and the output of the magnetic flowmeter has to be relied on the cross-calibration in the flow regimes that cover experimental conditions.
i 3.4 Calibration Under various flow conditions, a point-type electrode magnetic flowmeter and a modified strip-type magnetic flowmeter were cross-calibrated with rotameters and orifice flowmeters in a vertical test section and a horizontal test section, respectively. In this subsection, the results of the cross-calibration are discussed in detail.
3.4.1 Calibratien in Vertical Flow A 50.8 mm commercial paint-type electrode magnetic flowmeter was cross-calibrated with rotameters and orifice flowmeters under single-phase liquid flows. The results are shown in ')ut 3.3 and 3.4. The rotameters have 5% error in measurement. From these figures, it can be said that the rotameters are not as accurate as orifice meter and magnetic flow meter. The calibration showed good linearity and repeatability.
3-5 NUREG/CR-5578
Several sets of two-phase flow experiments were carried out. For each experimental set, the water flow rate or the liquid superficial velocity was kept at constant value, and the air flowrate or the gas superficial velocity was gradually increased such that the void fraction increased from zero to about 0.8.
Two-phase flow regimes were identified by visual observation. A gamma densitometer or an impedance void-meter was used to determine the void fraction. The superficial liquid velocity was computed by multiplying the time-averaged void fraction values with time-averaged values of output signals of the magnetic flowmeter, and the results are plotted in figums 3.5 and 3.6. The superficial liquid velocity 1
measumd by the magnetic flow meter is approximately constant for each liquid flow rate when the void fraction is lower than 0.25. The result is consistent with previous analysis such as Eq. (3.11). But when the void fraction is higher than 0.25, the time-averaged superficial velocity decreases. The flow regimes for high void fractions are slug and churn flows. The time-averaged superficial liquid velocity in slug flow shouki be calculated as the time average of product of the void fraction and liquid-phase velocity obtained from the
)
magnetic flowmeter. Obviously, that is not equal to the product of time average of the void iraction and liquid-phase velocity. Therefore, the application of the point-type magnetic flowmeter is limited in vertical bubbly flows.
3.4.2 Calibration in Horizontal Flow For horizontal flows, a modified 50.8 mm magnetic flowmeter is used. Owing to the complex phase distribution, prediction of the relationship between readings of the magnetic flowmeter and the liquid velocity is a formidable task. In this research, extensive cross-calibration was adopted to obtain the required relationship.
The calibration results of the modified magnetic flow meter for single-phase flow showed that the output of the flow meter is linearly proportional to the liquid velocity, as plotted in figure 3.7. A linear relation is fitted as i
v = 1.6046Vu,, -0.0005, (3.12) f where the Vu,, is the reading of the magnetic flowmeter. With this relationship, the error in the single-phase liquid velocity measurement is less than 3% of the full scale.
NUREG/CR-5578 3-6
.. - - -. - - ~ - ~
l The two-phase calibration of the magnetic flowmeter was conducted in stratified, slug (intermittent) and annular flows. For annular flows, single reading of the flowmeter might conespond to different liquid velocities. 'Ihis implies that the magnetic flow meter can not be used to measure the liquid velocity in annular flow. For stratified and slug flows, the calibration results are presented in figum 3.8. A parabolic and a linear relationship were fitted for stratified and slug flows, respectively. These relationships are v = 7.9221V,,,' + 1.8287Vu., - 0.0458,
_ (3.13) f and vf = 3.5490V,,, -0.6367, (3.14)
I The error in the liquid velocity cross-calibration is introduced by the magnetic flowmeter, the impedance void meter and rotameters. The error in horizontal two-phase flow is much larger l
than in single phase flow. The measurement error of the modified magnetic flowmeter is within i15%.
3.5 Conclusions In this section, liquid velocity measurements with magnetic flowmeters are presented.
Commercial magnetic flowmeters were modified and equipped with strip-type electrodes.
Theoretical aspects of this technique were investigated extensively for single-and two-phase conditions.
Cross-calibration between the magnetic flowmeter and other flowrate measurement instruments were conducted. Based on the cross-calibration results, it can be concluded that original commercial magnetic flowmeters and modified magnetic flowmeters are applicable to vertical bubbly flow and horizontal stratified and intermittent flow, respectively.
3-7 NUREG/CR-5578
~
3.9 References 3.1 M. Hori, T. Kobori and V. Ouchi, Method for Measuring Void Fraction by Electromagnetic Flowmeters, JAERI-1111 (1966).
3.2
- 1. A. Shercliff, The Theory of Electromagnetic Flow-Measurement, Cambridge University Press, Cambridge (1%2),
3.3 M. K. Bevir, The Theory of Induced Voltage Electromagnetic Flowmeters, J. Fluid Mech. 43, pan 3, pp. 577-590 (1970).
3.4 R. N. Bernier and C. E. Brennen, Use of the Electromagnetic Flowmeter, in a Two-Phase Flow Int. J. Multiphase Flow 9,251-257 (1983).
3.5 M. Murakami, E. Maruo and T. Yoshiki, T. Development of an Electromagnetic for Studying Gas-Liquid, Two-Phase Flow Int. Chemical Engineering 30, 669-702 (1990).
3.6 K. E. Knoll. Investigation of an Electromagnetic Flowmeter for Gas Liquid Two-Phase Flow Measurement,1991 ANS Winter Meeting, RDTPA 91-65,1-6 (1991).
3.7 D. G. Wyatt, Electromagnetic Flowmeter Sensitivity with Two-Phase Flow Int. J.
Multiphase Flow 12,1009-1017 (1986).
NUREG/CR-5578 3-8
i Magnetic Flux (Voltage) y Electromotive
' (j e 10 h
i i
., y 1
,..,.,o i-gV 3(
, q;;
e i
g 3
N
~
/g Magnet Coil Electrodes Figure 3.1 The magnetic flowmeter.
3-9 NUREG/CR-5578
^
.N N
T O
^
w l
v (a). Geometry of C-type electrode i
n D
v (b). Flow channel with C-type electrodes Figure 3.2 The modified electrodes of the mag'.;.ic flowmeter.
NUREG/CR-5578 3-10
1.2 o
E.1.0 -
O O
bo 0.8 -
ts:::
O O$
O E 0.6 -
o e
E O
i m
i 0.4 -
"O O
O
\\
s 0.2 -
Q)
O E*
O 0.0 :
i i
i i
i 0.0 0.2 0.4 0.6 0.8 1.0 1.2 jf measured by rotamters (m/s)
.}
v Figure 3.3 she calibration between the magnetic flowmeter and rotameters l
for single phase flows.
i 3-11 NUREG/CR-5578
10 I
g 8 '-
e b
a 6-
,g E
lir E
4-x t
-o 2a 2-e E
C O
i i
i 0
2 4
6 8
10 v, measured by orifice flow-meters (m/s)
Figure 3.4 The calibration between the magnetic flowmeter and orifice flowmeter for single phase flows.
l NUREG/CR-5578 3-12
1.5 J,=0.31m/s J,=0.62m/s 1.2 gi' j,=1.12m/s i
i a
i
~
A A k
~
A 5
0.9 a
b, J 0.6,
=
.s
".Y 0.3 '
0.0 0.0 0.2 0.4 0.6 0.8 1.0 a
l l
l i
i Figure 3.5 Measured superficial velocity versus void fraction at lower flow rates.
]
3-13 NUREG/CR-5578 i
l l
5 dii i
i A
i i
4 ime a
e S_.
3 p
k
~
g-2 a...
.n 4
j,=1.87m/s J,=3.11 m/s 1
=
j,=4.67m/s i
~
0 O.0 0.1 0.2 0.3 0.4 0.5 0.6 a
Figure 3.6 Measured superficial velocity versus void fraction at higher flow rates.
l l
t
. NUREG/CR-5578 3-14
't 1.5 1.0 -
7 M
C 0.5 -
0.0 L i
i i
i i
i i
i i.
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Mag Figure 3.7 The cross-calibration results of the modified magnet flow meter for single phase flow.
3-15 NUREG/CR-5578
i l
10 Eq. (2.14)
Data of Statified Flow Eq. (2.15) g_
Data of Slug Flow e
U 6-a i
f 4-o
$x 2-
/
09 cP 0
i 0
1 2
3 Mag i
Figure 3.8 The cross-calibration results of the modified magnet flow meter for two-phase flow.
NUREG/CR-5578 3-16
1 i
- 4. OXYGEN ANALYZER 4.1 Introduction j
It is imponant to know the amount of non-condensable gas in the condensing vapor if it contains some non-condensable gas. The efficiency of the condenser decreases with the i
increase in the amount of non-condensable gas. In a closed system, the presence of the non-condensable gas degrades the condensation heat transfer and may eventually stop the condensation process if the non-condensable concentration is sufficient to block the diffusion of the vapor onto the condensing film or wall. In most industrial system, air is always present in the condensers. To measure the concentration of air in a vapor mixture, first the oxygen concentration is measured. Then the air concentration is calculated assuming the oxygen concentration in air at 20.95%. Direct measurement of nitrogen concentration is expensive; techniques like gas chromatography and mass spectrometry are needed to measure the nitrogen concentrations.
Oxygen is vital to a large variety of industrial and life processes that involve oxidation and combustion. Many industries use pure oxygen or inert gases containing only a few particle per billion (ppb) of oxygen as a contaminant and both of these applications usually require analysis for concentrations of oxygen.
4.2 Oxygen Detectors Oxygen analyzers used today utilize the paramagnetic and electrochemical propenies of oxygen or apply catalytic combustion techniques.
4.2.1 Paramagnetic Detectors The paramagnetic materials are those within which an applied magnetic field is slightly l
increased by the alignment of electron orbits. The degree of paramagnetism is expressed in
(
terms of magnetic susceptibility. The magnetic susceptibility is measured by the ratio of the
(
intensity of magnetization produced in a substance to the ma aetizing force or intensity of
(
4-1 NUREG/CR-5578 l
field to which it is subjected. Oxygen has a positive magnetic susceptibility. The magnetic susceptibility decreases with increase in the temperature. Three basic types of instruments exploit the paramagnetic property of gaseous oxygen and these are, denection type, thermal type and the dual-gas type [4.1,4.2].
In the deflection type analyzer, the magne:ic force acts on a body that is free to rotate about a single axis [4.1]. The force is in proponion to the difference in the magnetic susceptibilities of the test gas and of the gas in the body. In the thermal paramagnetic oxygen analyzer, the oxygen sample is heated near a magnetic pole face. Due to heating the oxygen 3
looses its paramagnetism and is replace <l by fresh cooler oxygen from the incoming sample.
This action produces convection current commonly called " magnetic wind". This " wind" How rate is a function of oxygen concentration and is detected by the thermistor. In the dual-gas paramagnetic oxygen analyzer two gases with different oxygen concentrations are brought together producing a differential pressure. The reference gas passes through two ducts, one of which meets the sample gas in the magnetic field. Since both ducts are connected, the pressure produces a flow that can be measured. The pressure is proportional to the oxygen concentration of the gas sampic.
4.2.2 Catalytic Combustion Oxygen Detectors In the catalytic combustion oxygen detector, c feel is oxidized and the amount of heat generated is measured. The sensor consists of a measuring cell and a reference cell, with a filament in each. The measuring filament is provided with a catalytic surface to oxidize the fuel, while the reference filament serves only to compensate for vatiations in sample temperature and thermal conductivity. In the measuring cell, the fuel is oxidized in the presence of noble metal catalytic filaments. The restJting filament temperature is sensed by detceting its resistance, which is a measure of the oxygen content of the gas sample.
4.2.3 Electrochemical Oxygen Detectors
. Electrochemical oxygen detectors fall into three main categories: 1) high-temperature fuel cell-type detectors involving the conduction of oxygen ions (0 2) from one electrode to NUREG/CR-5578 4-2
another through a solid oxide electrolyte (e.g., zirconium oxide); 2) ambient-temperature galvanic-type detectors involving oxygen reduction at the cathode and dissolution of an active anode such as cadmium or lead in an electrolyte; and 3) polargraphic-type detectors consisting of three electrodes (cathode, anode, and a reference) and an electrolyte. The operation of a polargraphic detcetor is similar to that of the galvanic detectors execpt that in a polargraphic detector, an external potential is applied to the cathode to drive the oxygen reduction reaction.
The high-temperature fuel cell type oxygen detectors are used for measurements of gaseous oxygen, whereas the galvanic and polargraphic oxygen detectors are used for both gaseous and dissolved oxygen. All three types of electrochemical oxygen detectors measure the partial pressure of gascous oxygen and require temperature control or compensation.
4.3 Extractive Zirconium Oxide Oxygen Analyzer Among the three types of detectors, paramagnetic, catalytic combustion and electrochemical, the electrochemical oxygen detectors are rugged in design and can be used for high temperature and corrosive gas media. The high temperature zirconium oxide oxygen analyzer is suitable for steam media. and can be used for continuous sampling.
The Rosemount OXA 1(XX) oxygen analyzer [4.3] is designed for continuous measurement of oxygen concentration in a non-combustible or low level combustible Dowing gas sample.
4.3.1 Operation Principle The operation of the OXA 1(XX) oxygen analyzer involves the ionization of oxygen in both a sampl" and a known reference gas stream. The sensing cell consists of a small yttria-stabilized zirconium oxide disc, which is coated with porous metal electrodes (Figure 1). The cell operates at a temperature of 8(XTC. Sample gas is supplied to sampic side of disc by an external source. Reference gas, of known concentration, is provided to opposite side of disc.
When the sampic and reference gas streams come in contact with the electrode surfaces, oxygen ionizes into 0 2 ions. The oxygen ion concentration in each stream is a function of the partial pressure of oxygen in the stream. The potenti d n each electrode will depend on 4-3 NUREG/CR-557fs
~. _.. _ _ _ _ _. _. _ _ _. _. _ _ _.. _ _ _ _. _ _ _ _ _ _ _ _ _. _ _ _ _ _
the partial pressure of oxygen in the stream. The electrode with higher potential (higher oxygen concentration) will generate oxygen ions, whereas the electrode with lower potential
( lower oxygen concentration) will convert oxygen ions into oxygen molecules. The cell reaction at the two electrodes can be expressed as :
0 + 4c' -4 20-2 (at the cathode); O-29 O + 4 e'(at the anode).
2 2
The open-circuit voltage is related to the oxygen partial pressure by the following equation known as Nernst equation:
E = KT logio (P /P ) + C i 2 where :
E = the open-circuit voltage developed, P = partial pressure of oxygen in reference gas, i
P = panial pressure of oxygen in sample gas, 2
T = absolute temperature,
^
C = the cell's constant, and r
K = an arithmetic constant.
The output voltage is proportional to logarithm of the partial pressure of sample oxygen when the partial pressure of reference oxygen is constant. Because of this, output signal increases as oxygen concentration of sample gas decreases. This characteristic enables the oxygen analyzer to provide exceptional sensitivity at low oxygen concentrations. The maximum detectable oxygen concentration in the sample is equal to the oxygen concentration in the reference stream; at this point the open-circuit voltage will be zero volt.
If the oxygen concentration in the sample stream exceeds the oxygen concentration in the reference stream, the oxygen ions will move in the opposite direction, and the open-circuit voltage will be of the opposite polarity.
NUREG/ Cit-5578 4-4
4.3.2 Analyzer Specifications The OXA 1000 Oxygen Analyzer measures net oxygen concentration in the presence of all products of combustion, including water vapor. The unit consists of two major components, the oxygen sensor and the electronics, mounted in the same enclosure. The controller part of the electronics converts this voltage signal to a 0-20 mA or 4-20 mA output, and displays measurements on 3-1/2 digit LCD.
In Table 4.1 the analyzer specifications are given.
4.3.3 Calibration 4
The oxygen analyzer is factory calibrated. However it is recommended by the supplier that the sensor be calibrated every six months or more often if the signal shows zero shift of 0.1% and span shift of 3%. A two point procedure is used to field calibrate the OXA 100 Oxygen Analyzer. Two gas cylinders of different oxygen concentration are required, one for a low point standardization (ZERO) and one for a high point standardization (SPAN). The low point oxygen concentration must be greater than 1% O2 of system being measured, high point O2 concentration should not be greater than 100% of system being measured.
For calibration Portable Rosemount Oxygen Test Gas Kit [2] was used. This kit has ttvo portable tanks containing 8% and 0.4% oxygen in nitrogen. The calibration setup is shown in Figure 2. The sample gas line was attached to the test gas of choice and the test gas was irjected with pushbutton regulator. For ZERO point, the calibration gas of concentration 0.4% O was supplied at least five minutes for the display reading to settle. The meter 2
reading is then adjusted with keyboard to show the 0.4% O at ZERO point. Similarly for 2
setting the SPAN, the 8% O calibration gas was injected to the sample port and the meter 2
reading was adjusted to read 8% O at SPAN point.
2 4-5 NUREG/CR-5578
l 1
Table 4.1 Specifications of OXA 1000 Oxygen Analyzer Meawes 02 fmm 0-25% with linear output or Standard Measurement Rages 0.25-25% w,th logarithmic output i
Current Output-Filed select 0-20 mA DC or 4-20 mA DC.
i 0.1% O ori3% of reading whichever is greater 2
Accuracy Sample Specifications Temperature 10-700T Operating Pressure Limit 2.5 kPa (10 inches of water)
St nple Flow Rate 14 cc/ min (5 scfh)
Reference Flow Rate 5.6 cc/ min (2 scth)
Material in Contact with Sample 316L Stainless steel 0 /N, minimum 1% O2 Calibration gas Mixture 2 2 Supply Voltage 115/220 Vac i10% at 50/60Hz Power Consumption 250 watts maximum,75 watts nominal Analyzer Housing NEMA 1 Mounting Standard Rack Mounting,0.482 m (19 inches)
Environmental Specifications:
Location Non-hazardous, weather protected area Ambient Temperature Range 40"F to 120"F (5 C to 50"C)
Humidity 95% Maximum Relative Humidity Approximate Shipping. Weight 11.4 kg NUREG/CR-5578 4-6
4.4 Measurements in the PUMA facility 4.4.1 Installation In PUMA experiments, air is the non-condensable gas. The concentration of oxygen is used to estimate the air concentration. The locations of oxygen concentration measurements are: gas space of vessels (drywell, suppression pool, GDCS tank, reactor pressum vessel) and condenser inlets.(supply lines of PCCS and ICS). The sample gas is extracted from these locations through a stainless steel 6.35 mm (% -inch) tubing. The test pressure in vessels and piping is large (~200 kPa), but the required sample gas pressure for the analyzer is very small (2.5 kPa, Table 4.1). Hence a pressure-reducing valve was used to reduce the sample pressure down to 2.5 kPa. Laboratory dry air was used as the reference sample to the analyzer. Typical installation of the analyzer is shown in figure 4.3.
Two oxygen analyzers are used in PUMA experiments. One oxygen analyzer is dedicated to measure the oxygen concentration in the supply line of a PCCS condenser. The other analyzer is used to measure the oxygen concentrations from the other 11 locations. To sample the gas from these 11 locations, individual tubes are connected to a solenoid valves.
A total of 11 solenoid valves are then connected to a single pressure-reducing valve that is in turn connected to the sample port of the analyzer. The solenoid valves open one at a time introducing the sample gas to the analyzer from each of the locations. A timer circuit controlled the opening time and interval of the solenoid valves. Since the analyzers have to 1
be kept away from the hot piping, sufficient length of the sample extraction tubing is used. It is recommended that the sample gas tubing to the analyzer should be as short as possible.
This will reduce the sample's dead time, the time required for the sample to accumulate in the analyzer and develop a signal. In PUMA experiments the sample supply line lengths were quite long, typically between 4.3 m and 10.1 m.
Hence the analyzer renonse time is a function of the oxygen concentration and ;abe length.
4-7 NUREG/CR-5578
4.4.2 Analyzer Response A test was carried out to investigate the response of the oxygen analyzer with different line lengths and oxygen concentrations. The purpose of this test was to establish criteria for selecting sample times for the oxygen analyzer with gas sampling from multiple locations.
Obviously the flow of the gas to the analyzer is affected by two parameters, line length and oxygen concentration. Two supply line lengths were chosen, 4.3 m and 10.1 m. These lengths are the shortest and longest line lengths used in the gas sampling system. Three different oxygen concentrations were prepared for this test,3.5%,10% and 15% of O in 2
nitrogen. The oxygen concentration in the gas mixture was calculated based on its partial pressure. The supply of oxygen came from the shop compressed air supply and the nitrogen from a bottled cylinder. The mixture was prepared in an empty cylinder. Gas was supplied to the instruments at 273 kPa (25 psig), approximately the same pressure of the system during a typical PUMA test. The supply lines were Dushed with compressed air to ensure that no pockets of test gas and nitrogen were collected in the lines. The monitoring computer was set up and the gas supply solenoid was energized to start the test. This simultaneously start d the flow of gas to the analyzer and the data Die. After the analyzer response had stabilized the flow was stopped and the test file saved. The line was then changed and flushed with compressed air to establish conditions for a different line length at the same concentration.
The procedure was repeated with different oxygen gas concentration.
The results of the tests are shown in Ggures 4.4-4.7. The response of the analyzer at 30 seconds is indicated in figures 4.4-4.6. The data shows that the analyzer reaches 95% of the l
signal in less than 30 seconds. In figure 4.4-4.7, the response time (here denned as the time i
when the analyzer responds to change in concentration) is.s'..own as a function of line length and the oxygen concentration. It is clear that with different oxygen concentration and line lengths, the response time is different. The oxygen concentration has large effect on the response of the analyzer. The longest response time was 25 seconds. Based on this test the
]
sampling time for each line was set at 60 seconds. The Grst 30 seconds, the sample gas Hushes the line. The remaining 30 seconds the analyzer develops the signal corresponding to the sample gas. For gas concentration calculation the data taken during the last 5 seconds of this 60 seconds period is used.
'NUREG/CR-5578 4-8
4.4.3 Sample Measurement In PUMA Test Data on oxygen coacentration wen: taken during a bottom drain line integral test. The integral test was conducted for 16 hours1.851852e-4 days <br />0.00444 hours <br />2.645503e-5 weeks <br />6.088e-6 months <br /> (57600 seconds). Typical air concentration measured on a steam supply n.e of the PCCS condenser is shown in figure 4.8. The air concentration is calculated based on oxygen concentration in air at 21%. Except during the first 4500 seconds the air concentration is zero for the entire period of the test. In this test the RPV depressurizes by discharging steam into the drywell and the suppression pool The steam discharged into the drywell during the blowdown replaces most of the air in the drywell. However, some air is still trapped at lower drywell after the blowdown process. The PCCS takes steam from the upper drywell. When PCCS starts condensing the steam, the remaining air in the lower drywell makes its way to the upper drywell. Thus during first 300 seconds the air concentration at the upper drywell reaches a maximum of 20% as shown in figure 4.9. When air concentration in the drywell is depleted, the air concentration in the PCCS line decreases.
4-9 NUREG/CR-5573
.4 6
m,s 6
,,-s-
--,4,a--
s 4
4-a-laL-w,m----ek-
,+Pae-S A
-5< aw-4~>
a
--A--
-4 h
2
-wasM l
l 4.5 References 4.1 Buonauito, R. P., Oxygen Measurement in the Presence of Acid Gases, Measurement l
and Control, September 1984.
l 4.2 Gallagher, J. P., Oxygen Analyzers, Measurement and Data, May-June 1975.
4.3 Razaq, M., A new Sensor and a Microprocessor-Based Analyzer for Monitoring Low Parts-Per-Billion Oxygen Contamination in High Purity Process Gases, ISA Preprint 91-0535 (1991).
4.4 Rosemount Analytical, OXA 1000 Extractive Zirconium Oxide Oxygen Analyzer, Instruction Bulletin IB-103-80, Fisher-Rosemount (1994).
l l
l i
a l
NUREG/CR-5578 4-10
CONTROLLED HEAT [
h FOR CONSTANT 20
\\
HIGH TEMP
\\\\
4 e o';
i j
h.
8
,4 7
s,
'N
~
N p
s o
j R llhig i,q,n A
a
- 1 REFERENCE 2'
W
\\
i g
80.9' z 0.7 x
am F y K
$ 0.g,4 SAMPLE.I GAS j j jp p
o3 3
f
\\
0.2 l
ELECTRODES 0.1
\\
-20 0
20 40 60 80 100 120 CELL OUTPb T (MtLLIVOLTS)
U METER Figure 4.1 The flow of oxygen ions through the hot zirconium oxide electrolyte causes a voltage difference acros.s the element 4-11 NUREG/CR-5578
OO EXHAUST%c0
^
S^.MPLEd.r 3@
j j 2]4 c0
~
l
..EFERENCE/
i E
CAllBRATE IN PLACE CONNECTION PUSHBUTTON REGULATOR
/ WITH CONTENTS GAGE - SET 5 SCFH TEST HOSE CONNECTS 0.4 8.0 TO FITTING ON BACK OF ANALYZER (SAMPLE)
["
[*
P00030 Figure 4.2 Typical setup for OXA low oxygen analyzer with portable oxyger. calibration tanks 1
NUREG/CR-5578 4-12
1 Timer Circuit From PUMA Component Total 11 Lines w
?
To OXA 1000 Sample Side Pressure 4
?-
Reducing Valve Solenoid Valve 1
SAMPLING MANIFOLD TO DATA ACQUISITION SYSTEM OO D-.- J >
\\
n-,
fj SAMPLING l INE REFERENCE LINE 1 POWER (LAB AIR SUPPLY)
SUPPLY BACK VIEW OF ANALYZER TO EXHAUST AREA Figure 4.3 OXA 1000 oxygen analyzer installation on PUB 'A facility 4-13 NUREG/CR-5578
b 4.3 m Une 30 -
25 7 20 R
I E
\\
10 5
, 3.9 @ 30s J
0 0
50 100 150 Time (s) 10.1 m Line 30 -
25
\\
\\
20 u 'S
\\
a i
10 (4.03 @ 30s 0
0 50 100 150 Time (s)
Figure 4.4 Analyzer response to 3.5% oxygen concentration for line lengths of 4.3 m and 10.1 m.
NUREG/CR-5578 4-14
r 4.3 m Line 30 -
25 3 20 a
\\
- e. 5
(
10 10.15 @
5 30s 0
i i
i 0
50 100 150 Time (s) 10.1 m Line
?>0 -
25 20 No 15 10.14 @
30s 10 5
0 0
50 100 150 Time (s)
Figuw 5 / nalyzer response to 10% oxygen concentration for line lengths of 4.3 m and 10.1 m.
4-15 NUREG/CR-5578
4.3 m Line 30 -
25 20 t
8 15 16.62 @
10 30" 5
l 0
0 50 100 150 Time (s) 10.1 m Une 30 -
18.62 @
25
\\
30s 20 y
Rg 15 10 5
0 0
50 100 150 Time (s)
I Figure 4.6 Analyzer response to 15% oxygen concentration for line lengths Of 4.3 m and 10.1 m.
NUREG/CR-557tt 4-16
Time to Response 25 -
i g 15 e
E i
F 10 --
=
0-i i
5 5
10 10 15 15 170 400 170 400 170 400
[O2] and Line Length (in)
Figure 4.7 Response of oxygen analyzer as a function ofline length and oxygen concentration 4
i i
f i
4-17 NUREG/CR-5f 78
0 0
0
~
0 0
6 s
ruo h
6 1
r o
f tse t
la r
g e
tn i
I kae 0
r b
0 e
0 n
i 0
l n
0 ia 4
r d
mo t
t o
)
b s
a
(
g e
n i
m ru d
iT se in l
y lpp 0
us 0
S 0
C 0
C 0
P 2
n i
de rusae m
no i
tar tnecnoc r
iA I
8 0
4 0
e r
0 0
0 0
ug 0
0 0
0 iF 3
2 1
^# C8mbC8C8 mmO
$wmr O
4k
JI I
30.0-
-#~ 20.0
(
i C
h Y
t m
1:
f C
i 8
.CO l
?
O C l 10.0 a/ l G
m d '\\
O n'$
l tik k
k r
i g
y j
i
\\ \\,
s 2
Ah*-
0.0 l
h 0.0 2000.0 4000.0 6000.0 l
8 Time (s) i M,
Ma Figure 4.9 Air concentration measured in PCCS supply lines during a bottom drain line break integral test for first 4500 seconds i
[
i
i l
- 5. CONDUCTIVITY PROBES 4
5.1 Introduction There are six conductivity probes in the reactor vessel chimney region. The objective is to measure the local void fraction and local flow regime. These data complement integral measurements of the pressure drop along the chimney to provide information the local interfacial structure of the flow.
5.2 Probe design The six probes are located in the chimney of the reactor pressure vessel, and they are inserted in pairs through the three upper instrument ports (figure 5.1). For each pair one probe is' located at the peripheral chimney channel and the other at the center channel. All 1
' probes consist of two electroi: the inside electrode is a stainless steel wire at 0 508 mm in diameter and the outside electrode is a 3.175 mm stainless steel tube. The wire electrodes are 3.175 mm long. The key part of this design is the wetted area of the wire electrode, which cannot be too small because the deionized water has very low conductivity.
T'.e probes penetrate the vessel through Swagelok fittings. A schematic of the probe construction is shown in figure 5.2. Figures 5.3 and 5.4 are schematics of an instrumentation port and an instrumentation tube.
5.3 Electric Circuit Design i
The electronic circuits for these probes was originally designed by Koskie et al. [5.1]. A schematic is shown in figure 5.5. The circuit is driven by a 50 kHz sine wave from a signal generator. This provides high frequency response and it prevents electrode corrosion. The gain of the circuits has been increased to operate with low conductivity water. The circuit consists of an electrically isolated loop that is driven by the oscillator. The current through this. loop is measured. This produces an amplitude modulated signal. A 4.7 kHz low pass 5-1 NUREG/CR-5578
l l
filter removes the 50 kHz oscillator signal but leaves the envelope which is then amplified i
and sampled by the data acquisition system.
l 5.4 Data Acquisition System Hardware The data acquisition system consists of two Keithley-Metrabyte DAS-58 A/D converter boards with 1 Mbytes on-board memory. This provides a total capability of 16 data channels.
i At present only 6 channels are used. Each channel is sampled at 10 kHz (i.e.: twice the frequency response of the circuit) to prevent data aliasing. The data for all probes are stored
]
in the board memory before they are transferred to the dedicated personnel computer for processing. Data transferring occurs once every second. The void fraction data stored are averaged values of the indicator function over half a second period. The other half second is spent transferring the data from the DAS-58 board memory to the PC memory. The time partitioning is shown below:
CAL.MEAN l SPARE TIME CALMEAN
! SPARE TIME CPU i
i 1
SAMPLING TIME ! TRANSFER SAMPLINGTIME ! TRANSFER
)
ODM I
i i
t, UNIT INTERVALTIME J
r 7
5.5 Data Acquisition System Software A computer program [5.1] controls the data sampling from multiple channels and 1
multiple boards, and it also performs the real-time data processing. The program also displays the processed data after the experiment is done. The control of the program is j
performed through a graphical user interface. Furthermore, the program may display the instantaneous high and low voltages from each probe. This is useful for probe calibration.
NUREG/CR-5578 5-2 l
l l
4 The program is organized into modules for inputs, calibration, measurements and post processing. The flow chart describes the general program structure:
START u-t u
FUNCTION SWITCilINTERFACE u
1F 1F STARTDATA ACQUISITION CONVERT DATA FORMAT H
)
1Yes u
START SAMPLING DATA CONVERT DATA 1
I YeL No
- - USER INTERRU17T TO STOP ?
VISUALIZE DATA 7 No 1Yes u
TRANSFER DATA VISUALI733 DATA
'~T i
CALCLN. ATE MEAN VALUE 4%
SET UP CODE PARAMETERS _,
e-RUN TIME IS COMPLETE 7 1No CPU WAITS FOR NEW DATA 1
kmp STOP i
MEASURE CIRCUIT j
g,
+-
OtfrPtTT RANGE When the wire electrode is dry, the circuit output voltage is high. When it is wetted, the value is low. This value depends on the conductivity of the water. The conductivity changes with time because there is dissolved Morphiline and ammonia (theses chemicals were added to water to increase the conductivity) in the water which evaporate faster than the water.
Therefore the water chemistry changes as the experiment progresses. The high voltage on the 5-3 NUREG/CR-5578 t
other hand is constant. To correct this variation there is an algorithm to measure the low circuit voltage every 30 seconds. A threshold between the high and the low voltages is then selected and the indicator function is set to one, if the voltage is above the threshold; it is set to zero, if it is below the threshold as shown in the following figure.
+10V Sampled Signa Max.
MinishidiV5i6 U:" I.' " " ' " " " " " " "
Circuit output Data After Processing (indicator function)
S
. Signa Min (utiabic). _ _ _ _ _ _ _ ___
10V liquid gas liquid gas The void fraction is calculated for I second time span as GasDataNumber Void Fracn.on =
TotalDataNumber Furthermore, there is a Minimum Vapor Level above which the signal is considered to be all vapor. This is necessary because of the variable minimum voltage.
5.6 Probe Calibration Probe calibration has been performed in the air-water calibration loop as shown schematically in figure 5.6. The average void fraction was measured by measuring the pressure drop across the pipe with a manometer. The calibration was carried out in the NUREG/CR-5578 5-4
bubbly flow regime with the conductivity probe located in the centerline of the pipe. The probe was traversed across the pipe diameter and no significant variations in the void fraction were observed. The average void fraction was varied from 0 to 0.25. Bubble sizes ranged between 3 to 6 mm.
Figure 5.7 shows the calibration curves obtained in the air-water loop. The void fraction is calculated using the level threshold algorithm described previously for various threshold levels. Best results were obtained with a threshold level of 40% of the peak to peak signal (i.e.: the standard deviation is 0.026). The primary reason for the relatively large standard deviation is the large surface of the wire electrode, which is necessary because of the very low conductivity of the deionized water.
5.7 Sample Measurement Figure 5.8 shows the void fraction data from the six probes for the bottom drain line break experiment. Initially the water level is very low in the vessel and all the probes are uncovered and the void fractions are unity. After approximately 1000 seconds the GDCS refills the vessel and all the probes are submerged in a short period of time. Then the water begins to evaporate slowly and the probes become uncovered in pairs. Since the probe pairs are 50 mm apart there is a small time difference between the uncovering times of each pair.
l 5-5 NUREG/CR-5578
=.
.~
5.8 References l
1.
~ Koskie, J.
E.,
Mudawar, I. and Tiederman W.
G., Parallel wire probes for measurement of thick liquid films, Int.1. /.fultiphase Flow 15, No. 4, pp 521-530 (1989).
1 2.
He, J., Bertodano, M., Ransom, V. H. and Ishii, M.,
Puma RPV Void Fraction Measurement Code DA02 Document, Third Edition, School of Nuclear Engineering, Purdue University, West Lafayette,IN, August 18, 1995.
)
i i
1 I
l
\\
l i
l l
i l
- T NUREG/CR-5578 5-6 r
l
MN I
9 9 H ft it Q
lj Ft.ANGE l
4 5
2' RANGE C
SPECLAL (CAMERA PORT) 4 INSTRUMENT PORT SPECIAL INSTRUMENT PORT 2* ME E
SPECIAL l
(CAMERA PORT) 4 INSTRUMENT l
PORT SPECIAL 2' RANCE
'-".-..-.....-4 INSTRUMENT (CAMERA PORT)
PORT ts
()
is l
l i
()
I I
l i
l l
I I
I I
l l
l l
i l
l
\\
l
\\
l l
I i
I I
i k
I I
l l I
l l
I 1
i l
l g
2 nacE iE n (CAMERA PORT) g i
g i
m.
l
. 7>
rassmussm Figure 5.1 Location ofinstrumentation ports in reactor pressure vessel 5-7 NUREG/CR-5578
6
/// NUN /}////UN/4 l
s ssssws\\m\\\\\\ f s
+ frp-U, f
's
/,
0.125" f
0.700" 1/s" Stainlas meet tube L ofton spaghetti" tube T
~
_Stainlem meet wire
...r,-l M
._Duralco 4525 Epoxy 0.125"
~T.
, f 0.020" 0.125" 4
Figure 5.2: Schematic of conductivity probe NUREG/CR-5578 5-8
t
't l
r 0
} 2 J
8 l
=
h c
S
=r
-p I
i r
b s
e g
n a
Nh l
p F
4
?
t r
s op tn m
e
~
m 7
l i
u l
a l
p a
r
, w ts e
ms n
s i
e R
v fo 6
ll c
i t
a m
-2 e
h l
Fl a
c w
S
[
3 5
e r
~
ug i
F
,o 4
%p n
s y
it n v
.o.
it-
~
Q a
z o
c a
b n
h n
c o i
r ts u
n ar p
e t
n i
i p
syaNNhM
t ZCWmO
%O N.
uu 4oc 4~
T I %"
7" i
s m
5 O O R
A
....y.. v... _
s Y
g
.A
,.j
\\%"t, = to guide the
\\
i Bellows N
/
conductivity prok I" Stainies secel tube i
t Figure 5.4 Schematic ofInstrumentation tube.
M i
= i>
q
. = f-
>A w
v
=
d
.ir.
J g
-Y
- W hh
/+
u-4 3_
~11= v "L" #4 A'
^3-
=
M
-sc.
N us T=
N J
+= 1 1
~ f.
vi
)
4 }-
v.,1 1.,
[-[p=
==
w
-> B II.
r 11..
- 1
,,,f f.
0 v
_ vi ----
=
B)
Y
\\
et m
~1 1~
v
&f
,. - --;,2 v.
Figure 5.5 Circuit diagram for conductivity probe 5-11 NUREG/CR-5578
m w
J
- 1. WaterPump 10g 2 Orifice Plate 3 External Air Supply 4 Flow Meters 1/C 5 Lower Air Plenum 6 lower Water Plenum: Mixing Chamber 7 Air Pressure Cauge u W 9c 8 Clear Lucite 2" Test Section 9 Instrumented Flanges:
(a) L/D = 2 (b) L/D = 32 (c) L/D = 62 10 PhaseSeparator 11 WaterReturnLine 12 HeatExchanger 11 13 WaterBypassLine 14 Flow ControlValve 15 55 Callon Water Supply Tank n
U 9b 16 BubbleInjector si 16 K~
g l 'lll\\
'. - - f
- gu
$:::::'U p g..,[
l u u _
9a u
,2
_.....=-
e 2
CP-I 5 ',7 1
2 5
1II v
11 E
E E
4 i i i IIT 3
Figure 5.6 Air-water calibration loop NUREG/CR-5578 5-12
___ __~
i l
O.28 --
0.24 -~
-*-Threshold = 0.60
-*- 0.50
.8
-+- 0.40 o
-o-0.35
- 15. 0.2 --
-o-0.30 b
lE y
o 0.16 --
g 6
9.
8 0.12 --
u E
a.
3 0.08 --
/
0.04 --
-x
- X.
0%
0.00 0.05 0.09 0.13 0.18 0.20 Void Fraction (dp sensor) 4 Figure 5.7 Calibration curves for conductivity probe 5-13 NUREG/CR-5578
1.0
- c-T l~i
!t U
f O.8 i
Il l
ll' i.
.6 0.6
,l y
E"-
'I 32 i.
y 0.4
(
5 Top of chimney j
Middle of cMmney 0.2 V
,{
i j IV Bottom of chimney l
i i d 0.0 l
0 10000 20000 30000 40000 Time (s)
I Figure 5.8 Conductivity probe data for bottom drain line break (Probes located in the RPV chimney)
NUREG/CR-5578 5-14
. - ~
- 6. VORTEX FLOW METER 6.1 Introduction There are few choices of flow meters available for the measurement of low steam flow rate [6.1,6.2]. For steam flow rate measurement, differential pressure based flow meters such as venturi, nozzle and orifice plate, as well as, turbine flow meter, sonic flow meter and vortex flow meter can be used. The venturi, nozzle and orifice plate flow meters introduce a flow restriction and require a certain minimum pressure drop across the meter. Hence these meters are not suitable for low steam flow (typically 5 kg/h in a 25.4 mm pipe) measurement. The turbine and sonic flow meters have a large dependence on fluid properties such as density, viscosity and conductivity. In wet or superheated steam applications, turbine and sonic flow meters have poor performance. The vonex flow meter can be used for the measurement of wet and superheated steam flow rates.
The vortex flow meter is a device designed to measure the velocity of a fluid passing an obstruction placed in the stream of flow. This is accomplished by measuring the rate of vortex shedding from the obstruction, or strut, located in the stream of flow. It can accurately measure relatively low gas flow rates. In the vonex flowmeter there exists no moving parts that may lead to failure or require regular repair. Also, the meter does not need to be recalibrated for different working environments and fluids. All of these i
features make the vortex flowmeter an excellent choice for steam flow measurement especially for k,w flow conditions.
6.2 Principle of Operation The principle behind the vortex 110wmeter is relatively simple. The basic concept of operation is the result of the Von Kdrmdn vortex street effect [6.3, 6.4]. When an obstruction (a non-streamlined object) is placed in the path of a flowing stream, the fluid boundary layer is unable to remain attached to the object on its downstream sides and will alternately separate and shed vonices from one side and then the other. The slow-4 6--I NUREG/CR-5578
l moving fluid in the boundary layer on the bluff body becomes detached on the downstream side and rolls into eddies and vortices (figure 6.1).
This field of vonices is commonly referred to as the Kdrmdn vortex street. The distance between the shed vortices is constant, regardless of flow ulocity. Stated in terms of a flag 11 uttering in the wind, the intervals between vorucer (1) (or the wavelength of fluttering) is constant and is only a function of the diameter of the flag pole (d). Therefore, the faster the wind, the faster the vonices are formed and the faster the flag flutters as a consequence, but without changing its wavelength. A dimensionless parameter known as the Strouhal number relates the velocity of the liquid to the shedding frequency. The Strouhal number is given by:
fd St =
V For a range of Reynolds number, the Strouhal number is generally a constant, as r
shown in figure 6.2. Therefore, if one knows the vortex shedder width d and has a l
detector that is sensitive enough to count the vortices and determine the vonex frequency f, one can measure the flowing velocity of any substance as:
flow velocity = (fx d)/St In building a flowmeter based on Kdrman's principle, the manufacturer usually selects an obstruction width d that is one-quarter of the pipe diameter (ID). As long as the obstruction is not eroded or coated, the pipe Reynolds number is high enough to produce vonices, and the detector is sensitive enough to detect these vortices (for gases such as hydrogen the forces produced by the vortices are very small), one has a flowmeter that is sensitive to flow velocity and is insensitive to the natore of the flowing media (liquid, gas, steam) density, viscosity, temperature, pressure and any other properties.
A flow-sensitive frequency detector can be either a heated thermistor element or a spherical magnetic shuttle (with the movement of the shuttle measured inductively).
Detectors that are sensitive to pressure use either metal diaphragms or vanes. Pressure exerted on diaphragms can be converted into a variable capacitance, or a variable strain on a piezo-resistive, piezoelectric, or inductive sensor. Pressure exerted on vanes can NUREG/CR-5578 6-2
similarly be converted into an electrical signal through any of the aforementioned sensors. Alternatively, the velocity components in the free vortices downstream of the bluff body can be used to modulate an ultrasonic beam diametrically traversing the meter housing.
6.3 Foxboro Vortex Flow Meter 6.3.1 Specifications The Foxboro vortex flow meter model 83W-C is a wafer body (flangeless) style meter. The flow-sensitive detector is a liquid filled, double faced diaphragm capsule with a piezoelectric crystal in the center which detects the vortex-produced pressure changes as they are transmitted through the filling liquid.
In Table 6.1, vortex flow meter operating conditions and perforrance specifications are given. The accuracy of the meter and the calibrated range is shown in figure 6.3. For steam, the accuracy is i 1% of reading for flow rates with Reynolds number of 10,000 or greater.
6.3.2 Calibration The Femboro vortex flow meter model 83W-A0151SRTNF-C has a factory calibration for steam flow rate. A typical calibration sheet is shown in figure 6.d The meter K-factor, which shows the relationship between volumetric flow rate and th peak rate is meter specific. The meter non-linearity is less than 0.5% for the whole range of Reynolds numbers.
6.4 Steam Flow Measurement in PUMA Test The vortex flow meter sizing is based on the minimum flow rate required in the PUMA facility. Table 6.2 shows the steam flow rate limits for different size meters.
Based on this table and the PUMA test requirement, the meters of size 19.1mm,38.1 mm 6-3 NUREG/CR-5578
and 50.8m are chosen.
The meters were installed in the pipeline according to manufacturer's recommendations as shown in figure 6.4.
Steady state steam flow rate measurements were carried out with vortex flow meter in an ICS condenser supply fine. The ICS has a intake from the RPV steam dome, and it has a condensate line that returns condensed water to the downcomer of the RPV. The steam
)
supply line has a 38.1 mm vortex flow meter and the condensate line has a 12.7 mm 1
magnetic flow meter. At steady condensation, the rate of the steam entering ICS i
condenser should be exactly equal to the rate of condensed water entering RPV j
downcomer through the condensate line. A steady condensation in ICS was established by keeping a constant heater power to the RPV. The inlet steam flow rate was measured by the vortex flow meter and the condensed water flow rate was measured by the 1
magnetic flow meter. By applying different heater power to RPV, steam flow rate was set
)
at different levels. The result of the measurements is shown in figure 6.6. The difference in the steam flow rate measurement by vortex flow meter was less than 2%. The vortex flow meter has good linearity at both high and low steam flow rates.
i l
4 NUREG/CR-5578 6-4 e
6.5
References:
4 6.1 Upp, E.L Fluid Flow Measurement; Gulf Publishing Co. (1993) 6.2 Choices Abound in Flow Measurement, ChemicalEngineering, April (1991).
6.3 Griffin, O.M, Ramberg, S.E.,The Vortex Street Wakes of Vibrating Cylinders, J.
' FluidMechanics 66, pp553-576 (1974) 6.4 Within W. G., Theory, Design and Application of Vortex Shedding Flow-meters Measurement Technology for 80's, ISA Symposium, Delaware (1979).
6-5 NUREG/CR-5578
Table 6.1 Operating conditions and performance specifications of Foxboro vortex flow meter i
OPERATING CONDITIONS (a)
Reference Operating Normal Operating influence Conditions (b)
Conditions Operating Limits Process Fluid Clear Water Liquid, Gas, and Steam Liquid, Gas, and Steam Process 20 to 30 C
-20 to + 430 C
-20 and +430 C Temperature (70 to 85' F)
(0 to +800* F)
(0 and + 800
- F)
Ambient Temperature 20 to 30 C
-40 to +85 C
-40 and + 85 C i
(Electronics Housing)
(70 to 85* F)
(- 40 to + 185
- F)
(-40 and + 185* F)
Relative Humidity 50 to 90%
0 to 100%
0 and 100%
Supply Voltage 24 V de 12.5 to 42 V de 12.5 and 42 V dc (see Figure 2)
(a) Umited to nonflashing, noncavitating conditions. Flow rate and temperature of process may induce flashing and cavitation which is dependent on pressure drop and process vapor pressure. A minimum positive back pressure is required for proper operation.
j (b) Assumes ANSI Schedule 40 process piping; flanges bored to interfacing pipe inside diameter; gaskets to be 3.2 mm (0.13 in) thick and not protruding into process line; and control valves to be a minimum of eight pipe diameters downstream of flowmeter.
Clear water is from from air or particles. Deviation in pipe bore etc. can be compensated for in the meter electronics.
PERFORMANCE SPECIFICATIONS Factory Calibration Conditions Factory-Calibrated Flow Range for Water Nominal Nominal Mean i
Meter K-Factor in Pulses /ft
- Range in Reynolds Number I
Size (Pulses /L)
US gpm (L/s)
Range in L/s Range 3/4 in (15 mm) 5580 (197) 6.9 to 34 0.43 to 2.1 30 000 to 150 000 1 in (25 mm) 2250 (79.5) 8.9 to 56 0.56 to 3.5 30 000 to 190 000 11/2 in (40 mm) 570 (20.1) 14 to 140 0.88 to 8.7 30 000 to 300 000 2 in (50 mm) 258 (9.11) 18 Io 230 1.1 to 15 30 000 to 380 000 3 in (80 rrm) 78.7 (2.78) 34 to 500 2.1 to 32 38 000 to 570 000 4 in (100 mm) 34.8 (1.23) 59 to 890 3.7 to 56 50 000 to 750 000 6 in (150 mm) 10.00 (0.353) 140 to 2000 8.5 to 130 76 000 to 1100 000 8 in (200 mm) 4.26 (0.150) 240 to 3600 15 to 220 100 000 to 1500 000 10 in (250 mm) 1.99 (0.0703) 390 to 5800 24 to 370 130 000 to 1900 000 12 in (300 mm) 1.16 (0.0410) 560 to 8400 36 to 530 160 000 to 2 300 000 NOTES: 1) The K-factor is the relationship between input (volumetric flow rate) and the output (pulse rate).
- 2) Mean K. factor: The arithmetic mean value of K-factor over a def.gnated flow rate range (reference conditions).
The mean K-factor is derived as:
Mean K-factor = (KMAX + KMIN) / 2 Where KMAX is the M2.imum K-factor and KMIN is the Minimum K-factor over the calibrated flow range.
- 3) The linearity for liquids within the calibrated Reynolds Number range is : 0.5%; outside the calibrated range the linearity is : 1%, except for flow less than 20 000 Reynolds Number - see " improved Accuracy for Uquids at Low Flow" on Page 4.
- 4) For gas and steam flow, the linearity is 21%, above 20 000 Reynolds Number.
NUREG/CR-5578 6-6 s
Table 6.2 Steam flow rate limits for different size vortex flow meter Process Process Pressure Te Mmenum and Manrrann Fkm Rates for the Foeowng Stres(b) in h
in kg/h for In kg/h for in koh for In kg/h for in kg/h for in kg/h for in keh for in kg/h for in kg/h for in kgh for kPa Gauge C
15 mm 25 mm 40 mm 50 nun 80 mm 100 mm 150 mm 200 nun 250 mm 300 nun 0
100.0 3.3 & 34.9 4.2 & 86.7 7.7 & 342 12.9 & 748 28.3 & 1640 49.9 &
114 &
200 &
327 &
475 F.
2890 6620 11 600 19 000 27 500 250 139 0 3.6 & 112 5.6 & 278 13.9 & 1100 23.1 & 2310 50.7 & 5070 89.3 &
204 &
359 &
586 &
850 &
8930 20 400 35 900 58 600 85 000 t
500 158.9 4.3 & 186 7.3 & 461 17.8 & 1780 29.8 & 2980 65.3 & 6530 115 &
263 &
462 &
754 &
1100 &
11 500 26 300 46 200 75 400 110 000 1000 184.1 5.8 & 330 9.7 & 818 23.8 & 2380 39.7 & 3970 87.1 & 8710 153 &
351 &
616 &
1010 &
1460 &
15 300 35 100 61 600 101 000 146 000 5000 265.1(c) 24.8 & 1240 41.5 & 2070 102 & 5100 171 & 8500 492 & 18 700 820 &
2160 &
4170 &
7460 &
11 400 &
32 800 75 200 132 000 215 000 313 000 10 000 311.7(c) 36.5 & 1830 61.1 & 3050 156 & 7500 286 & 12 500 717 & 27 500 1370 &
3610 &
6990 &
12 500 &
19 200 &
48 300 111 000 194 000 317 000 460 000 Process Pro ss v
f Pressure Temperature 4
hi h
in b/h for in b/h for in b/h for in b/h for in b/h for in b/h for In b/h for in b/h for in b/h for In b/h for psig F
3/4 h 1 in 11/2 in 2h 3 in 4h 6h 8 ht 10 in 12 h 0
212.0 7.2 & 77.0 9.3 & 191 17.1 & 754 28.5 & 1650 62.5 & 3620 110 &
252 &
442 &
721 &
1050 &
6370 14 600 25 600 41 800 60 700 50 297.6 8.3 & 309 13.9 & 766 34 2 & 3020 57 & 5700 125 & 12 500 220 &
504 &
884 &
1440 &
2100 &
22 000 50 400 88 400 144 000 210 000 100 337.8 10.9 & 531 182 & 1320 44.8 & 4480 74.7 & 7470 164 & 16 400 289 &
660 &
1160 &
1890 &
2750 &
28 900 66 000 116 000 189 000 275 000
- C SN 470.0(c) 45.3 & 2270 75.8 & 3790 186 & 9310 310 & 15 500 735 & 34 100 1410 &
3700 &
7160 &
12 800 &
19 600 &
60 000 137 000 241 000 393 000 571 000 sO 1000 546.3(c) 64.9 & 3250 109 & 5430 267 & 13 300 473 & 22 200 11 3 & 48 000 2270 &
5980 &
11 600 &
20 700 &
31 700 &
?
86 000 197 000 345 000 564 000 819 000 0:
w 1500 597.5(c) 82.2 & 4110 137 & 6870 354 & 16 900 648 & 28 200 1630 & 61 800 3110 &
8190 &
15 800 &
28 300 &
43 400 &
109 000 249 000 437 000 713 000 1 040 000 (a) Values hsted see for dry saturated steam (steam quahty = 100%) with the low Bow cuth set at Rs mmimum value. For steam quahty other than 100%. refer to the Foxboro FlowEspert 5169 Program or Tl 027%7.
(b) A mirumum upper range value (URV) of three times the minimum Row rate is recommended. The maximum URV equals the mdmum Row rate. For example, for a 25 mm analog output nowmeter at O kPa. gauge. the recommended minimum URV would be 3 K 4.2 or 12.6 kgte and the manimum URV would be 86.7 h AL D
(c) The Extended Ternperature Range sensors. T" or "T*, are required for these appilcations. see Model Code section.
. m
l i
l 1
rhN fv
~ ~ ~
~~
' ~ ~
~
10
- ~
I
\\
~
\\.
\\,
s-Figure 6.1 The Karman vortices behind a bluff body NUREG/CR-5578 6-8
I i
0.4 0.3 w
O
.o E
5 0.2
.c:
- 3
.h% 0.1 0
1E1 1E2 1E3 1E4 1ES 1E6 1E7 Reynolds Number Figure 6.2 Vortex shedding frequency as a function of Reynolds number b9 NUREG/CR-5578
l I
)
i r
9 K-FACTOR zo LINEARITY Q
b WATE R O
CAllBRATEC REGION 2 7. -
f
/
REFERENCE _
//,
\\\\'au ; \\ \\\\\\\\\\\\'
^ 0*5%
K-FACTOR
' \\\\\\\\\\\\\\\\\\ M 1
m-
- 2 7. -
(
CO.RRECTED u GUTPUT REGION O
10 000 100 000 1 000 000 REYNOLDS NUI,tBER 1
1 Figure 6.3 Typical accuracy and K-factor linearity of the vortex meter for the calibrated range i
NUREG/CR-5578 6-10
.- _.._..___._ __ _ _.-.~_.._._._..____._..
4 o
VORTEX METER CALIBRATION WITH MASTER METERS CUSTOMER NAME:
PURDUE UNIVERSITY CUST. ORDER NUMBER: 7Yl5736 CUST. TAG NUMBER:
SALES ORDER:
96F46387/1 SERIAL NUMBER :
96361898 p
MODEL CODE: 83W-A30SISRTNF-C
' METER SIZE:
.75 -INCH 4.
FLOW FLUID PULSES
% DEV RATE VELOCITY FREQUENCY TEMP.
PER FROM REYN0LOS RUN (FT3/ MIN) (FT/SEC.)
(HZ)
( DEG.F ) CU.FT.
MEAN K NUMEER t
1 4.29 23.82 375.9 80.6 5254.1699
.287 160314 2
2.85 15.79 249.6-81.4 5264.0357
.100 107337 3
1.83 10.18 161.5 81.5 5285.3312
.305 69280 4
1.08 6.02 94.9 81.3 5253.2322
.305 40853 i
REFERENCE K-FACTOR IS 5269.3 PULSES /CU. FT.
186.08 PULSES / LITER INSTRUMENT WITHIN CALIBRATION SPECIFICATIONS.
LINEARITY 15 +/-
.305 PERCENT FROM REYNOLOS NUMBER 40853 CUST. K FACTOR 14410 PULSES /LB MAXIMUM FLOW 50 LBS/HR CAL DENSITY.363 LBS/CU.FT.
FLOWING TEMP. 365.8
- F FULL SCALE PULSE RATE = 200.23 HZ (CALIB. PURPOSES ONLY)
LOW FLOW CUT-IN 86.31 HZ l
FILE # 0 BY: J. ZONFRILLO
)
DATE: 23 Sep 1996 The calibration test equipment used is traceable to the National Institute of Standards and Technology (NIST), formerly called N8S.
l j
Figure 6.4 factory supplied calibration sheet for 19.1 mm vortex flow rr.eter 6-11
12F 5.0 35p VERTICAL CONDutT Ct BY USER TO AVOID l
A ACCUMULATION OF.
l/
[y, DATA PLATE MOtSTURE IN t
-t TERMINAL BLOCK
'y O
ENCLOSURE.
l l'
l 1,.
U 116 46 1
l i
36 s
... i.4
-._.. T q
r_
C;
>.c
}
HOLE FOR NOMINAL 20 mm (CEE 23), PG16. OR
~
\\
n' in I
l 1/2 in CONDUlT FITTING.
PLUG HOLE ON
__ f _
X a
OPPOSITE SIDE n
IF NOTE USED.
C.__.
B t
C_
-- - --l A
t
./
I o
C Dimensions: mm or in, as Appreable Nominal U.no A
B Size O.D.
LD.(b)
C E
E(c) 20 mm 57 mm 18.8 mm 79.5 mm 236 mm 302 mm 0.75 in 2.25 in 0.740 in 3.13 in 9.3 in 11.9 in 25 mm 67 mm 24 mm 79.5 mm 246 mm 312 mm 1.0 in 2.63 in 0.957 in 3.13 in 9.7 in 12.3 io
/
40 mm 88 mm 40 mm 79.5 mm 265 mm 329 mm
[."-
s 1.5 in 3.38 in 1.50 in 3.13 in 10.4 in -
13.0 in 50 mm 104.6 mm 49.2 mm 79,5 mm 283 mm 340.4 mm o
2.0 in 4.12 in 1.94 in 3.13 in 1
- 2 in 13.4 in A
80 mm 136.7 mm 72.9 mm 95.3 mm 318 mm 376 mm o
3.0 in 5.38 in 2.87 in 3.75 in 12.5 in 14.8 in a
100 mm 174.5 mm 96.7 mm 121 mm 355 mm 404 mm 4 in 6.87 in 3.8 i in 4.75 in 14.0 in 16 in 150 mm 222 mm 146 mm 178 mm 399 mm 465 mm 6.0 h 8.75 in 5.8 in 7.0 in 15.7 in 18.3 in 200 mm 279 mm 194 mm 229 mm 452 mm 518 mm 8.0 in 11.0 in 7.6 in 9.0 in l 17.8 in 20.4 in Figure 6.5 Typical dimensions and installation of vortex 1: v meter on a pipe line NUREG/CR.5578 6--12
_ _ _ _ _ - _ -. - _. _. _. - ~ _ _ - -...... - _ _ - _ _ _
4 3.00E-02
^*
2.50E-02 3
2.00E-02 y
1.50E-02 f
1.00E-02
/
=g 5.00E-03 0.00E+00 0.00E+00 1.00E-02 2.00E-02 3.00E-02 Magnetic Row Meter Data (kg/s)
Figure 6.6 Comparison of steam flow fate measured by '.ortex flow meter (38.1 mm size) and condensed water flow rate measured by magnetic flow meter.
6-13 NUREG/CR-5578
i
- 7. - CONCLUSIONS l
l The measurements in PUMA include steam-water two-phase flow, low steam flow (typically 10-100 kg/h in 38 mm pipe), and non-condensable gas (air) concentration in steam in addition' to the standard pressure, temperature and differential pressure. To measure the two-phase flow rate, the void fraction or quality is required. However, no i
l standard commercial two-phase flow meters or void meters are available. Hence, the impedance void-meter and the local conductivity probe were developed for the void fraction measurement under the integral test conditions.
For low steam flow rate l
measurements the vortex flow meter was used. Improvements were made on the magnetic flow meter and the oxygen concentration meter for PUMA test conditions.
l A simple ring-type impedance void meter was developed for measurement of the area a
average void fraction. The electronics and the software were developed for transient pipe void fraction measurement. Cross-calibration between the impedance technique and other void measurement methods.such as the gamma densitometer. technique and the differential pressure method were performed.
Calibration relationships between impedance and void fraction in each flow regime (i.e., bub'aly, slug, churn, annular and i
stratified) for both vertical and horizontal flows were obtaited. In addition, flow regimes of vertical flows were identified from the characteristics of the impedance signals.
l The conunercially available magnetic flow meters can only measure liquid flow rate in single phase flow and they can not measure two-phase flow rate. Commercial l
magnetic flowmeters were modified and equipped with strip-type electrodes. Theoretical aspects of this technique were investigated extensively for single-and two-phase l
- conditions.
Cross-calibration between the magnetic flowmeter and other flowrate measurement instruments were conducted. Based on the cross-calibration results, it is concluded that original commercial magnetic flowmeters and modified magnetic flowmeters are applicable to vertical bubbly flow and horizontal stratified and intermittent flow, respectively if the proper signal processing software developed at
[
Purdue is used.
l 7-1 NUREG/CR-5578 l
l
.._..___.____.__._m The oxygen gas analyzer was used to measure the concentration of oxygen (or air) in the steam. A multi-port gas sampling system was developed for on-line measurement of air concentration in the steam during the test in various PUMA components. The response time of the analyzer was studied as a function of oxygen concentration and the sampling line length. The oxygen concentration has a large effect on the response of the.
analyzer. The longest response time was 25 seconds. Based on this test the sampling time for each line was set at 60 seconds. For gas concentration calculation the data taken during the last 5 seconds of this 60 seconds period was used.
A conductivity probe was developed to measure the local void fraction for conditions in the RPV up to 1 MPa and 180 C. He probe was made of a single point stainless steel electrode. The electronics and the data processing software were developed for the transient local void fraction measurement. The probe calibration was performed in the air-water loop. Sample data for the PUMA test conditions are presented and discussed.
i A commercial vortex flow meter was used to measure low steam flow rates in pipes.
l The vortex flow meter characteristics, installation and testing are discussed. Steady state steam flow rate measurements were carried out with the vortex flow meter in the PUMA ICS condenser supply lines. The inlet steam flow rate was measured by the vortex flow meter and the condensed water flow rate was measured by the magnetic flow meter. The difference in the steam flow rate measurement by the vortex flow meter was less than 2%. The vortex flow meter has good linearity at both high and low steam flow rates.
I i
l l
1 NUREG/CR-5578 7-2
. - ~
~.
NRC FORM 335 U.S. NUCUEAR REcutATORY COMhesslON
- 1. REPORT NUMBER E naa, DYRevY. "an"d AdQ %--
am, 3aca BIBLIOGRAPHIC DATA SHEET 6". "Y 1 NUREC/CR-5578
- 2. TITLE AND SU8 TITLE 3-ATE REP RT PUBLISHED Instrumentation for the PUMA Integral Test Facility l 1999 MONTw vEAR
- March
- 6. AUTHOR (S)
- 8. TYPE OF REPORT S. T. Revankar, M.
Ishii, Y. Mi, M. L, Bertodano, Y. Xu, S. Kelly, V. H. Ransom. Purdue University and J. T. Han, NRC Technical Report
- 7. PERICO COVERED (inclusive Dates) 06/01/97 - 05/31/98
- 8. PERFORMING ORGANIZATION - NAME APC ADORESS (if NRC provide DMslan. Off6ce or Region, U.S. Nuclear Regulatory Commission, and malling address; :f contractor, provide name and matting address.)
School of Nuclear Engineering Purdue. University West Lafayette, IN 47907-1290
- 9. SPONSORING ORGANLZATION = NAME AND ADORESS (if NRC, type 'Same as above"; if contractor, provide NRC Olvision. Office or Region.
U.S. Nuclear Regulatory Commission, and mailing address.)
Division of Systems Technology Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Washington DC 20555-0001
- 10. SUPPLEMENTARY NOTES J. T. Han, NRC Project Manager
- 11. ABSTRACT (200 words or less)
The flow conditions in the PUMA (Purdue University Multi-Dimensional Integral Test
~
Assembly) tests require special instrumentation.
In this report, special instruments used in PUMA, impedance meter, modified magnetic flow meter, oxygen analyzer, conductivity probe, and vortex flow meter are presented.
The impedance void-meter was developed to measure the average void fraction in pipes.
The commercially available magnetic flow meter electrodes were modified to measure the liquid flow rate in two-phase flow.
.A multi-port gas sampling system using oxygen analyzers was developed for on line measurements of air concentration in steam at various components of PUMA.
A commercial vortex flow meter was used to measure low steam flow rate in pipes.
A conductivity probe was developed to measure the local void fraction in the reactor pressure vessel Or water-steam at 1 MPa (150 psig) and 180 C (356 F) or below.
These meters were tested and calibrated in the laboratory.
- 12. KEY WORDs/DESCRIPTORS (List words or phrases that will assist researchers in locating the report.)
- 13. AVAILABILnY STATEMENT unlimited Ts.o phase instruments, PUMA, simplified boiling water reactor (SBWR)
H. 3ECURITY CLAsSIFICADON impedance raeter, modified magnetic flow meter, oxygen analyzer, N '*8')
conductivity probe, vortex flow meter, two-phase measurement, void fraction, non-condensable measurement unclassified (This ReporO unclassified
- 15. NUMBER OF PAGES
- 16. PRICE NRC FORM 335 (2-89)
l i
i 4
i i
1 Printed
- on recycled paper i
Federal Recycling Program i
UNITED STATES SPECIAL STANDARD MAIL '
POSTAGE AND FEES PAID NUCLEAR REGULATORY COMMISSION us WASHINGTON, DC 20555-0001 pggg OFFICIAL BUSINESS PENALTY FOR PRIVATE USE $300 2
0 9
l 4
a m
m m.
m a
_m, m
ves
=
ww ew -< m w-w-w
- -- m