-
^ /
FLOW-INDUCED VIBRATION OF COMPONENT COOLING WATER
HEAT EXCHANGERS*
CONP-900617—1
^ DE90 003789
Y. S. Yeh and S. S. Chen**
Nuclear Engineering DepartmentTaiwan Power Company
Taiwan, Republic of China
**Materials and Components Technology DivisionArgonne National
Laboratory
Argonne, Illinois USA
SUMMARY
This paper presents an evaluation of flow-induced vibration
problems of
component cooling water heat exchangers in one of Taipower's
nuclear power
stations. Specifically, it describes flow-induced vibration
phenomena, tests to
identify the excitation mechanisms, measurement of response
characteristics,
analyses to predict tube response and wear, various design
alterations, and
modifications of the original design. Several unique features
associated with the
heat exchangers are demonstrated, including energy-trapping
modes, existence
of tube-support-plate (TSP)-inactive modes, and fluidelastic
instability of TSP-
active and -inactive modes. On the basis of this evaluation, the
difficulties and
future research needs for the evaluation of heat exchangers are
identified.
* Paper to be submitted for presentation at the ASME Pressure
Vessel and Piping
Conference, Nashville, Tennessee, June 17-21, 1990.
M A S T E R ^DISTRIBUTION OF THIS DOCUMENT IS UNLIMITED
The submitted manuscript has been authoredby a contractor of the
U. S. Governmentunder contract No. W-31-109-ENG-38.Accordingly, the
U. S. Government retains anonexclusive, royalty-free license to
publishor reproduce the published form of thiscontribution, or
allow others to do so, torU. S. Government purposes.
-
1. Introduction
The Maanshan Nuclear Power Station contains two trains of
component
cooling-water (CCW) heat exchangers for each of two reactors.
Trains A and B of
Unit 1 were placed in service in December 1982 and January 1983,
respectively.
Tube leakage in both heat exchangers was discovered in March
1983, after only a
few months of operation. Damage was noticed at tube supports and
midspan
regions. This was the beginning of a series of steps taken by
Taiwan Power
Company (Taipower) to resolve the problem. A brief history is
presented in
Table 1.
The CCW heat exchangers are one-pass, shell-and-tube heat
exchangers
with sea water in the tube side and hot demineralized water in
the shellside. A
schematic diagram of the heat exchanger is given in Fig. 1. Each
exchanger
contains 3,488 tubes contained in a 66-in. (i.d.) circular
cylindrical shell. The two
tube sheets are 1.75-in. aluminum-bronze plates, and two types
of baffle plates are
0.625-in. carbon steel plates; four are segmental baffle plates
with a single
horizontal cut, and ten are support plates with double
horizontal cuts as shown in
Fig. 1.
A 36-in. (o.d.) circular impingement plate was installed at the
inlet to
distribute the incoming flow to the first three passes. All heat
exchanger tubes
are cupronickel with 0.75-in. o.d. and 0.049-in. wall thickness.
The tubes are
arranged in a rotated triangular pattern with a pitch of
15/16-in.
-
2. Tube Damage
2.1 Initial Inspection
In March 1983, after a few months of operation, 14 tubes were
found
damaged in Train A and 4 in Train B. All damaged tubes had flat
spots
developing in the first and second baffle plates next to the
inlet nozzle. These flat
spots are 6-24 inches long. A damaged tube in Train A developed
a 4-in. long
crack along a 24-in. flat spot, and a damaged tube in Train B
was severed at the
second support plate. All damaged tubes were located in the
inlet area.
Apparently, the high flow velocity in the nozzle caused large
tube motion that
resulted in tube damage. The flat spots were caused by tubes
colliding with each
other.
In February 1984, the manufacturer of the CCW heat exchangers,
Struthers-
Wells Corp. (SWC), completed initial modifications of both CCW
heat exchangers
in Unit 1. The modifications included removing 420 tubes from
each heat
exchanger. No damage inspection was performed when they were
removed and
placed in groups in an open field.
In March 1984, the first five spans were inspected for flat
spots between the
supports and ring-type markings at tube supports. If no damage
was found in the
first five spans, the tube was considered undamaged. If damage
was observed in
these five spans, inspection continued to the subsequent spans
until no further
damage was found. Figure 2 shows the results of the damage
inspection. Among
396 inspected tubes in Train A and 400 in Train B, 36 tubes in
Train A and 28 in
-
Train B were damaged. Twenty tubes in Train A and 16 in Train B
contained flat-
spot damage between the first two baffle plates.
2.2 Damage Inspection of Selected Tubes
In July 1984, 28 tubes from each heat exchanger in Unit 1 were
selected for
evaluation of the extent and distribution of tube damage. Figure
3 shows the
location of the tubes selected from each heat exchanger. In
contrast to the tubes
near the nozzle, most damage occurred at the supports and
baffles. Only one tube
from Train A showed flat-spot damage. In general, most of the
damage was in
the first three supports. However, several tubes showed damage
only at the
middle supports, e.g., supports 5 through 9. This damage pattern
shows that
tubes located in regions other than near the nozzle are also
subjected to
unacceptable vibration.
From the damage inspection, several general conclusions can be
made:
• Excessive flow-induced vibrations were noted in both Trains A
and B.
• Tube-to-tube impacts appear between the first two baffles.
• Damage at supports occurs in various regions of the tube array
and at
different axial locations.
It is apparent that the original design of the heat exchanger is
inadequate
from the standpoint of flow-induced vibration. The mechanism
causing such
large tube vibration is due to fluidelastic instability.
-
3. Vibration Tests of CCW Heat Exchangers by Bechtel Group,
Inc.
3.1 Phase-1 Tests
Two modifications were made in accordance with the
manufacturer's
recommendations after the initial report of tube damage: (1)
removing 420 tubes
from each train. [One hundred fourteen removed tubes were not
replaced and the
holes were left in the baffles; supports and tube sheets were
plugged. The rest of
the tubes were replaced by BWG 12 tubes (0.75-in. o.d. and
0.109-in. wall
thickness.)]; and (2) replacing the original impingement plate
at the inlet (a 36-in.
diameter, 3/8-in. thick stainless steel plate welded to tie bars
and a support plate)
by a flow deflector. Bechtel Group, Inc. performed tests on the
modified heat
exchanger [1].
Eighteen tubes were selected for measurement. Two accelerometers
were
embedded in each pod and inserted inside the selected tubes to
measure tube
accelerations in two perpendicular directions at midspans. Three
additional
accelerometers were placed on the outside shell.
After properly checking all accelerometers, flow tests were
performed at a
series of flow rates, varying from zero to the rated flow of
127,000 gpm. A 14-
channel FM-type recorder was used to store the
accelerations.
Analysis of the data includes time-domain data, frequency
spectra, and
vibration amplitude.
-
3.1.1 Time-Domain Data
The data on acceleration-time plots revealed that
• Low-level background vibration of about 0.25 g peak-to-peak
existed at
zero flow due to other equipment.
• Tube accelerations increased with flow velocity, reaching a
magnitude
of 10 g peak-to-peak (p-p) at 127,000 gpm.
• Impact signrls were noted for flow above 7,000 gpm.
• The impact signals, occurring right at the peak of the
vibration
waveform, indicate a direct impact at the accelerometer pod
location.
In addition, the accelerations were doubly integrated to obtain
tube
displacements. Figure 4 shows the tube displacements (p-p) as a
function of flow
rate for nine tubes.
3.1.2 Frequency Spactra
Typical frequency spectra are shown in Fig. 5. The following
general
characteristics are noted:
• Tube responses at low flow rate are broadband, indicating that
the
excitation is turbulent buffeting.
-
• As the flow velocity increases, the major frequencies are
located in two
regions, i.e., at 21-24 and 26-32 Hz.
3.1.3 Monitoring of SheU Vibration
Three shell accelerometers were installed on the external
surface of the heat
exchanger to supplement the internally mounted accelerometers.
At 6,000 gpm,
faint impact sounds could be heard at the first baffle location.
At 7,000 gpm,
clean, but still faint, impact sounds could be heard at the
second support and first
baffle. At 8,000 gpm, sound intensities increased noticeably. At
12,000 gpm, the
impact sounds became more frequent and much stronger.
The heat exchanger was tested again with the flow deflector
rotated 180
degrees. The objective was to obtain vibration response under a
new shellside
inlet condition. The test was conducted in February 1984 with
flow rates at 7,000,
10,000, and 12,000 gpm. Tube responses were comparable with
those of the
original test. However, the flow deflector did change the
details of tube response.
3.1.4 Conclusions and Recommendations
When the shellside flow rate exceeded 7,000 gpm, the modified
CCW heat
exchanger experienced severe tube vibration. Some of the tubes
are subjected to
fiuidelastic instability at the rated flow of 12,700 gpm. The
modifications
recommended by the manufacturer are inadequate. Meanwhile, it
is
recommended that the following steps be taken before a final
resolution of the
problem is reached.
-
8
• Operate the heat exchangers at flow rates not to exceed 6,000
gpm
whenever possible.
• Implement a surveillance program to detect leakage and inspect
tube
damage before final modification.
• Investigate measures to resolve the problem, (e.g., improve
flow velocity
distribution by adding inlet and outlet nozzles, and increase
tube natural
frequencies by adding tube supports, in particular, in the inlet
and outlet
regions.
• Evaluate the feasibility of replacing the existing units with
a better
design.
3.2 Phase-2 Tests
Additional modifications to the heat exchangers were made as a
result of the
Phase-1 tests. The modifications included staking the bottom 14
tube rows in
Bays 1, 2, and 3, and installing perforated plates in Bays 2 and
3. Two large cuts
were made to implement these modifications in June 1984.
• The first bay was staked at midspan and the second and third
bays were
staked at the 1/3 and 2/3 points of each span. Staking strips of
carbon
steel with a width of 0.75-in. and a thickness of 60 mils were
wedged
horizontally between adjacent tube rows.
-
• A plate, 43 x 45 x 3/8 inches, with 3/8-in. diameter holes on
a 13/16-in.
staggered pitch, resulting in 19% open area, was installed in
Bajr 2.
• Two perforated plates were installed in Bays 2 and 3; a top
plate
contained 1/2-in. diameter holes on a 1-in. staggered pitch (22%
open
area) and a bottom plate contained 7/16-in. diameter holes on a
1-in.
staggered pitch (17% open area). All plates were stainless steel
sheets
3/8-in. thick.
The Phase-2 vibration tests were a continuation of the Phase-1
test. The test
procedure, instrumentation, and data analysis are the same as
those in Phase 1.
Accelerometers used for Phase 1 were used, to the extent
feasible, for Phase 2
tests. Twenty-eight additional acceleronieters were procured for
Phase 2 [2].
To guide the selection of suitable locations for installing
accelerometers, and
to assist the interpretation of test results, natural
frequencies of selected tubes
were calculated.
The test results were summarized as follows:
• Tube impact signals are noted at high flow velocity.
• Response of the unstaked tubes in the frequency range of 20 to
55 Hz was
excited at flow rates of 6,100 gpm and above. The only tube that
was
instrumented and staked detected prominent tube response at 129
Hz.
• The largest tube motion was about 15 mils.
-
10
Tube vibration amplitudes increase with increasing shell side
flow velocity.
At the maximum flow rate of 12,700 gpm, the peak acceleration
amplitude was
about 1.2 g, which was considerably lower than the 15 g in
Phase-1 tests. The
maximum tube displacement was about 30 mils, peak to peak; tube
collision at
midspan is unlikely. Impacts were noted for flow over 9,000 gpm.
Tube wear is
still possible, and the long-term effect is uncertain. Under the
described
conditions, it is recommended that flow rate be restricted to
9,000 gpm, a rigorous
surveillance and maintenance program be implemented, and sources
be located
for replacement heat exchangers.
After the Phase-2 test, Taipower decided to replace two CCW heat
exchangers
with a new design, and modify the other two trains. The new
design has a 15-in.
tube span and it is not expected to have any flow-induced
vibration problems. The
other two trains were modified as follows by Taiwan Machinery
Manufacturing
Corp. (TMMC) on the basis of a recommendation of a French
consulting firm:
• The inlet region was changed from the original three passes to
seven
passes.
• The outer shell diameter was increased.
The modified heat exchanger is shown in Fig. 6 After completion
of the
modification by TMMC, flow tests were conducted by Loutech
Taiwan, Inc. [3].
-
11
4. Analysis and Vibration Tests of Modified Heat Exchangers
4.1 Tests by Loutech Taiwan, Inc.
To verify the severity of vibration and structural integrity,
Loutech Taiwan,
Inc., performed a flow test in September 1988. Twenty-five tubes
were
instrumented with two accelerometers each to measure the
accelerations in two
perpendicular directions. The instrumented tubes and the fisial
locations are
given in Fig. 7.
Four types of plots were obtained from the data from each
accelerometer:
• Overall tube displacement versus time: the peak amplitudes at
each of
147 time intervals within 110 seconds.
• Average spectra of tube displacements: average spectra for 0
to 147 Hz.
• Peakhold spectra: the peaks of all spectra from 0 to 147
Hz.
• Frequency spectra versus time: frequency spectra of different
tubes as
function of time.
Figure 8 shows typical average frequency spectra for a series of
flow rates
and Figs. 9 and 10 show the frequency spectra as a function of
time.
One method to determine the response characteristics and the
critical flow
velocity is to plot tube displacement as a function of flow
rate. The most frequently
-
12
used tube response is the rms value of *abe response
(acceleration, velocity, or
displacement). Loutech did not provide rms values of tube
displacements.
Alternatively, the peak values are presented. Figures 11 to 13
show the peak tube
displacements as a function of flow rate in logarithmic scale
for selected tubes,
where H and V are used to indicate horizontal and vertical
components
respectively.
4.2 Natural Frequencies: ISP-active and-inactive Modes
The natural frequencies of the tubes are very important for the
interpretation
of tube response data. A free-vibration analysis was performed
for different cases.
CCW heat exchanger tubes, fixed to the tube sheets at the two
ends, were
supported by 14 tube-support plates (TSP), see Fig. 14. To
facilitate assembly and
relative motion caused by thermal expansion, holes in the TSP
were made larger
than the tube diameters. It is not uncommon for tube holes to be
drilled 0.4 to
0.8 mm larger than the outside diameter of the tubes. In this
case, the magnitude
of the clearance was not known. Owing to the clearance, TSP? may
not provide
support in some cases. Therefore, there are two types of modes:
TSP-active and
TSP-inactive
Current design practices consider the heat exchanger tubes to be
simply
supported, without clearances, at the TSPs. For small
clearances, this
assumption is expected to be applicable; the tube will respond
as a continuous
beam supported by all TSPs, and anti-nodes of all modes do not
appear at the
TSPs. This type of mode is called "TSP-active." Studies show
that this
assumption is indeed applicable if the clearance is small [4-7],
although there are
-
13
some variations of resonant frequencies. With this assumption,
the simple beam
theory can be used to predict tube natural frequencies, as well
as tube responses to
different excitations.
When the clearance is relatively large, the tube will rattle
inside some of the
TSP holes for small-amplitude oscillations. This type of mode,
in which some
TSPs do not provide effective supports, is called
"TSP-inactive.:i
For intermediate clearances, the tube response will depend
strongly on the
excitation amplitude. For low excitations, the tube will vibrate
within the
clearance gap (TSP-inactive modes). For large excitations, the
tube will be in
contact with all TSPs most of the time (TSP-active modes). In
reality, the tube
response will be composed of both TSP-active and TSP-inactive
modes.
A typical CCW heat exchanger tube was analyzed for the
TSP-active and
-inactive modes. Table 2 shows the natural frequencies of
TSP-active and -inactive
modes for different cases. The characteristics of Tube BWG 12,
presented in this
paper, are similar to those of Tube BWG 18 except that the
natural frequencies of
Tube BWG 18 are a little lower. As an example, the natural
frequencies of TSP-
active modes of two tubes are compared in Table 3. The
difference in natural
frequencies of the two tubes for the TSP-active mode is small;
this is also true for
TSP-inactive modes.
Two different situations are considered for TSP-inactive modes:
a particular
TSP, such as TSP 2, being inactive, and two TSPs, such as TSPs 2
and 4, being,
inactive.
-
14
Figure 14 shows the tube and all TSPs. The TSP-active modes are
given in
Fig. 15 for the first five modes. The results agree well with
those obtained by
Bechtel [2]. Figure 16 shows the first five modes when TSP 2 is
inactive, and Fig.
17 shows the first two modes when two TSPs are inactive.
Among the TSP-inactive modes, some of the modes are dominant in
specific
spans; these modes are called energy-trapping modes [8]. For
example, Fig. 18
shows the lowest TSP-inactive modes for different cases. These
modes are the
most important ones when one TSP is inactive.
43 Tube Response Characteristics
Among the information that is most important for the
interpretation of the
daca is the natural frequency of the tubes. Based on the results
presented in Table
2 and Figs. 15-18, we can construct frequency bands for tube BWG
12, as shown in
Fig. 19. These frequency bands are plotted on the assumption
that only one
support is inactive. Experimental data show that those modes
associated with
more than two inactive supports are not significant. Two types
of frequency
bands, passing and stopping, can be constructed. Within a
passing band, waves
can propagate freely through each tube support; however, within
a stopping band,
waves will be attenuated as they travel from one support to the
next. All natural
frequencies are in the passing bands only. In a stopping band,
since no natural
frequencies exist within that frequency range, tube response
will be very small.
Figure 19 shows the approximate boundaries of the first two
stopping and passing
bands:
First Stopping Band 0 - 9.5 Hz
-
15
First Passing Band 9.5 - 73 Hz
Second Stopping Band 73 - 95 Hz
Second Passing Band 95 - 176 Hz
The first passing band is approximately between 9.5 and 73 Hz;
therefore, the
dominant tube response is expected to be in this frequency
range. In addition, the
first TSP-active mode is at 23.9 Hz; any frequencies below this
are the energy-
trapping modes associated with TSP-inactive modes. The first two
stopping bands
are approximately between 0 and 9.5 Hz and 73 and 95 Hz; in
these two frequency
bands, we expect that the tube response will be very small.
Based on the experimental data and analytical results, some
general
observations can be made.
TSP-inactive modes contribute significantly to tube response. It
can be seen
from the frequency spectra that TSP-inactive modes, whose
frequencies are lower
than that of the first TSP-active mode of approximately 23 Hz,
contribute
significantly to the response.
Some of the tubes are subjected to fiuidelastic instability of
TSP-inactive
modes. For example, let us consider Tube 11C. The frequency
spectra for Tube
11C are given in Fig. 8 for various flow rates. When the flow
rate is increased
from 10,000 gpm to 145,000 gpm, the contribution of TSP-inactive
modes increases;
at 14,000 and 14,500 gpm, TSP-inactive modes are dominant.
Furthermore, Fig.
13 shows that the slope of the response curve is much larger
than 2, an indication
that the tube is subjected to fiuidelastic instability
associated with TSP-inactive
modes.
-
16
The response of various tubes is different and the response of
each tube
changes with flow rate. The key parameter in this case is the
clearance between
the tube and TSP holes. As flow rate changes, the tube
configurations also
change, i.e., the contact points of the tubes with TSPs changes
with flow rate due
to the change in drag and lift forces.
The motion of Tube 5B is the largest of all instrumented tubes
associated with
TSP-active and -inactive modes; however, the TSP-inactive modes
are not the
dominant mode of the response. The large amplitudes are believed
to be
attributable to the high local flow velocity. The tube is not
subjected to fluidelastic
instability.
Based on the frequency band presented in Fig. 19, the dominant
mode of each
tube at different flow rates can easily be identified from the
spectra.
The response of all tubes is contained in the first passing
band; this is in
agreement with the frequency band consideration. The modes
include TSP-active
and -inactive modes. In some cases, the frequencies of the
active and inactive
modes can be seen clearly in the spectra. As an example, the
frequency spectra of
Tubes 1A-H and 2A-H are given in Fig. 9, the lower frequency
peaks are TSP-
inactive modes and the higher peaks are TSP-active modes. In
some other cases,
only TSP-inactive or TSP-active modes dominate; see Fig. 10 for
typical examples.
Among the instrumented tubes, the displacements of the tubes in
the first
row in the inlet region are the largest. This can easily be seen
by comparing the
displacements of tubes in different locations.
-
17
The tube displacements in the inlet regions are larger than
those in the outlet
region. In addition, the spans close to the inlet nozzle
(accelerometer groups B
and C) have larger displacements than spans further av/ay from
the inlet nozzle.
The main characteristics of the tube responses are
• TSP-inactive modes are important since some of them are
subjected to
fluidelastic instability.
• Fluidelastic instability of TSP-active modes does not
occur.
• Tube characteristics (support conditions) change with flow
rate.
• The tubes located in the front row in the inlet region have
larger
displacements.
• Tube wear may be important owing to the impact/sliding
associated with
fluidelastic instability of TSP-inactive modes and turbulence
excitation of
TSP-active and -inactive modes.
4.4 Fluidelastic Instability and Wear
From the experimental data, it is noted that fluidelastic
instability may
occur, depending on the modes. The critical flow velocities for
different cases
were calculated. The flow velocity distribution is not known.
Assuming a
uniform flow velocity distribution, the flow velocity is 1.16
ft/sec at 12,700 gpm.
-
18
The mass-damping parameter is 0.24, assuming a damping ratio of
1%.
Therefore, the reduced flow velocity is 0.84 and 1.84 for
TSP-active and TSP-
inactive modes, respectively. On the basis of the existing
correlations [9,10], the
critical reduced flow velocity ranges from 1.1 to 2.2, On the
basis of the existing
correlations, we expect the tubes are not subjected to
fluidelastic instability of TSP-
active modes, but may be subjected to fluidelastic instability
of TSP-inactive
modes. This is in agreement with the experimental data.
In the original design, the incoming fluid passes across three
passes only;
the reduced flow velocity at 12,700 gpm is 1.96 and 4.29 for
TSP-active and TSP-
inactive modes, respectively. Therefore, the tubes are subjected
to fluidelastic
instability of TSP-inactive modes and some of the tubes are
subjected to fluidelastic
instability of TSP-active modes. The instability of the
TSP-active modes had
caused tube damage in the original design.
Once the tubes are subjected to fluidelastic instability of
TSP-active modes,
they are not acceptable from the standpoint of flow-induced
vibration and it is not
necessary to calculate wear rate because tube leaks will occur
within a very short
time of operation. On the contrary, when the tubes are subjected
to fluidelastic
instability of TSP-inactive modes, the wear rate is much smaller
and its
significance depends on many other parameters. In general, it is
difficult to
predict tube wear under such conditions. In CCW heat exchangers,
several
important parameters, such as clearance, flow velocity
distribution, and wear
rate, are not known. It is impossible to predict tube wear with
confidence.
However, as an exercise, tube wear was calculated by means of
the approach
taken by Connors [11]. If a wear coefficient of 600 and
clearance of 25 mils are
used in the calculations, it is concluded that it will take
about 16 years for a tube to
-
19
wear through. This wear calculation is not accurate and only
gives an estimate of
order of magnitude.
5. CONCLUSIONS
On the basis of this evaluation, we have concluded the
following:
In the original design, the flow velocity was larger than the
critical flow
velocity; therefore, the tubes were subjected to the
fiuidelastic instability of TSP-
active modes. The original design was not acceptable and the
observed damage
was due to fluidelastic instability.
In the extensively modified version, the tubes are not subjected
to the
fluidelastic instability associated with TSP-active modes.
Therefore, occurrence of
tube leaks caused by flow-induced vibration is not expected to
occur during short-
term operation. Based on the tube response characteristics, it
is believed that the
clearance between the tubes and support plate holes is
relatively large. Because of
the clearance, wear will result from the sliding and impacting
of the surfaces due
to the oscillations associated with turbulence excitation and
instability of TSP-
inactive modes. When the unit is put in service, a surveillance
program and
follow-up work should be conducted to ensure the safety and
predict the useful life
of a unit.
From the experience with CCW heat exchangers, it appears that
the state of
the art is such that more work remains to be done. Specifically,
the following
items should be emphasized:
-
20
• Flow Velocity Distribution. An efficient computer code to
compute the
flow velocity distribution for general heat exchangers would be
very
useful. Currently, few codes can be used for CCW heat
exchangers.
Unfortunately, more time and cost will be required to use these
codes.
• Wear Prediction. The state of the art is inadequate to predict
tube wear.
Very little information is available for wear associated with
TSP-inactive
modes and data for TSP-active modes are very limited. The
detailed
dynamics of tube/support interaction are still not well
understood.
• Tube Response and Prediction of Stability. Methods are
available to
predict the critical flow velocity of fluidelastic instability
and tube
response to various excitations. However, the methods require
further
improvement.
ACKNOWLEDGMENTS
Tests of CCW Heat Exchangers were performed by Bechtel Group,
Inc., and
Loutech Taiwan, Inc.
Many engineers in the Nuclear Engineering Department of Taiwan
Power
Company have contributed to various aspects of the program;
among them are Y.
H. Cheng, S. C. Cheng, T. K. Lee, P. C. Kao, F. Y. Lai, H. H.
Lee, and P. F. Wang.
This work was performed for Taiwan Power Company under an
agreement
with the U.S. Department of Energy under contract agreement
31-109-ENG-38-
85847.
-
21
This work was performed for Taiwan Power Company under an
agreement
with the U.S. Department of Energy under contract agreement
31-109-ENG-38-
85847.
1. "Test Report: Vibration of Tubes in Component Cooling Water
Heat
Exchangers, Phase 1 Tests," Bechtel Group, Inc., San Francisco,
CA,
(September 1984).
2. "Test Report: Vibration of Tubes in Component Cooling Water
Heat
Exchangers, Phase 2 Tests," Bechtel Group, Inc., San Francisco,
CA,
(October 1984).
3. "Final Report on CCW Heat Exchanger Tube Vibration Test,
Third Nuclear
Power Station, Taiwan Power Company," by Loutech Taiwan,
Inc.,
(December 1988).
4. Sebald, J. F., and Nobles, W. D., "Control of Tube Vibration
in Steam
Surface Condensers," Proc. Am. Power Conf., Vol. XXIV, 630-643
(1962).
5. Moretti, P. M., and Lowery, R. L., "Heat Exchanger Tube
Vibration
Characteristics in a 'No Flow1 Condition," Final Report:
Tubular
Exchanger Manufacturers Association Experimental Program,
Oklahoma
State University (1973).
-
22
6. Shin, Y. S., Jendrzejczyk, J. A., and Wambsganss, M. W.,
"Effect of
Tube/Support Interaction on the Tube Vibration of a Tube on
Multiple
Supports," Argonne National Laboratory, Technical Memorandum
ANL-
CT-77-5 (1977).
7. Rogers, R. J., and Pick, R. J., "On the Dynamic Spatial
Response of a Heat
Exchanger Tube with Intermittent Baffle Contacts," Nucl. Eng.
Des. 36, 81-
90(1976).
8. Chen, S. S., "Vibration of Continuous Pipes Conveying Fluid/'
in Flow-
Induced Structural Vibrations, ed. by E. Naudascher,
IUTAM-IAHR
Symposium, Karlsruhe, Germany, 1972, pp. 663-675.
9. Pettigrew, M. J., Sylvestre, Y., and Campagna, A. O.,
"Vibration Analysis
of Heat Exchanger and Steam Generator Designs," Nucl. Eng. Des.
48, 97-
115(1978).
10. Chen, S. S., Flow Induced Vibration of Circular Cylindrical
Structures.
Hemisphere Publishing Corp., 1987.
11. Connors, H. J., "Flow-Induced Vibration and Wear of Steam
Generator
Tubes," Nucl. Technol., 33, 311-331 (1981).
DISCLAIMER
This report was prepared as an account of work sponsored by an
agency of the United StatesGovernment. Neither the United States
Government nor any agency thereof, nor any of theiremployees makes
any warranty, express or implied, or assumes any legal liability or
responsi-bility for the accuracy, completeness, or usefulness of .
. , . information, apparatus, product, orprocess disclosed, or
represents that its use would not infringe privately owned rights.
Refer-ence herein to any specific commercial product, process, or
service by trade name, trademarkmanufacturer, or otherwise does not
necessarily constitute or imply its endorsement recom-mendation, or
favoring by the United States Government or any agency thereof. The
viewsand opinions of authors expressed herein do not necessarily
state or reflect those of theUnited States Government or any agency
thereof.
-
Table 1. History of CCW Heat Exchangers
1. Trains A and B of Unit 1 were placed in service in December
1982, andJanuary 1983, respectively.
2. Tube leakage was discovered in March 1983. Damage was found
at supportsand midspans.
3. The manufacturer's action taken after tube leakage:
* Damage was considered to be attributable to inappropriate
design of theinlet flow distributor.
• Modifications, as follows, were recommended to alleviate the
problem:
a. Replacing the original flow impingement plate by a flow
deflector.b. Replacing 306 tubes from each heat exchanger with
heavier tubes (8WG
12vs.BWG18).c. Eliminating 114 tubes from each heat exchanger
and plugging the holes
in the tubesheets and baffles.
4. The CCW Heat Exchangers of Units 1 and 2 were placed in
service in Januaryand February of 1984, respectively, after
modifications were made in responseto manufacturer's
recommendations.
5. Bechtel Group, Inc. performed Phase-1 test.
Test of Train A after modifications in February 1984. Six tubes
at the inletregion and three tubes at the outlet region were
instrumented.
Train A was tested again with the flow deflector rotated 90
degrees.
6. Additional modifications were recommended on the basis on the
results of thePhase-1 test.
Staking the bottom 14 tube rows in bays 1, 2 and 3.
Installing perforated plates in bays 2 and 3.
-
7. Modifications of Unit 2 were completed in June 1984, and
those in Unit 1 werecompleted later.
8. Bechtel Group, Inc. performed Phase-2 test in July 1984.
Train B of Unit 2 and Train A of Unit 1 were tested.
Train B of Unit 2 was retested.
9. Taipower decided to order two new CCW heat exchangers and to
modify two ofthe original CCW heat exchangers
10. The modification of the original CCW heat exchanger was
completed byTaiwan Machinery Manufacturing Corp. in 1988.
11. Loutech Taiwan, Inc., performed tests on the modified Train
B of Unit 1 inSeptember 1988.
12. Taipower and Argonne National Laboratory performed an
evaluation of CCWheat exchangers.
-
Table 2. Natural Frequencies in Hz of CCW Tubes
All TSPsActive
2.38B48E+012.47746E+012.61928E+012.80613E+013.02973E+013.28225E+013.55647E+013.84557E+014.14246E+014.43886E+014.72400E+014.98308E+015.19570E+015.33561E+015.416e8E+019.55004E+019.74398E+011.0O413E+02i,04170E+021.08504E+021.13254E+021.18294E+021.23513E+021.28796E+021.34005E+021.38959E+021.43402E+021.46972E+021.49222E+021.52036E+022.24451E+022.27477E+022.32095E+022.37882E+022.44457E+022.51509E+022.58773E+022.66008E+022.72974E+022.79427E+02
TSP 2Inactive
1.13053E+012.39328E+012.49449E+012.65018E+012.84452E+013.062Q9E+013.29786E+013.55885E+013.84635E+014.15166E+014.46243E+014.76359E+015.03460E+015.24556E+015.41605E+016.81634E+019.55977E+019.77727E+011.00993E+021.04873E+021.09095E+021.13550E+021.18346E+021.23522E+021.28941E+021.34393E+021.39611E+021.44234E+021.47717E+021.52035E+021.76043E+022.24604E+022.27996E+022.32972E+022.38855E+022.45133E+022.51733E+022.58778E+022.66092E+022.73303E+02
TSP 3InactJve
9.68389E+002.39275E+012.48674E+012.62322EI012.80663E+013.043B6E+013.32520E+013.63698E+013.96541E+014.28958E+014.54128E+014.72946E+014.99427E+015.22849E+015.41581E+016.49951E+019.55868E+019.76213E+011.00489E+021.04179E+021.08759E+021.14026E+021.19731E+021.25641E+021.31414E+021.35919E+021.39086E+021.43571E+021.47461E+021.52035E+021.70462E+022.24585E+022.27741E+022.32180E+022.37920E+022.44969E+022.52867E+022.61100E+022.69198E+022.76585E+02
TSP 4Inactive
9.53309E+002.39177E+012.47919E+012.62222E+012.B3474E+013.09982E+013.36640E+013.57155E+013.85633E+014.20758E+014.56757E+014.89305E+015.02601E+015.20820E+015.41540E+016.43080E+019.55670E+019.74738E+011.00469E+021.04707E+021.09784E+021.14751E+021.18557E+021.23706E+021.29952E+021.36271E+021.41933E+021.44225E+021.47154E+021.52034E+021.69312E+022.24553E+022.27520E+02
.?.32211E+022.38809E+022.4652BE+022.53571E+022.58937E+022.66472E+022.74758E-I-02
9222233334445,5,5,6,9.9.1.1.1.1.1.1.1.1.1.1.1.1.1,2.2.2.2.2.2.2.2.2.
TSP 5Inactive
.51959E+00
.39064E+01
.47765E+0I
. 63923E-I01
.86230E101
.05253E+01
.29078E+01
.6315SE+01
.97486E+01
.16079E+01
. 4727IE+01
.84508E+01
.34234E+O1
.19770E+01
.41472E+01
.41603E+01
.55442E+0174437E+0100799E+02,05206E+0208902E+0213414E+0219654E+0225808E+0229097E+0234623E+0241092E+0246170E+0247017E+0252031E+0269085E+0224517E+0227486E+0232747E+0239495E+0244902E+0251894E+02
'61017E+0269207E40273149E+02
TSP BInactive
9.51825E+002.38852E+012.49456E+012.62225E+C12.85222E+013.04704E+013.34421E+013.60336E+013.90738E+014.23021E+014.48968E+014.84973E+015.01855E+015.30070E+015.40789E+016.41220E+019.55012E+019.77780E+011.OO463E+O21.05040E+021.08789E+021.14413E+021.19067E+021.24690E+021.30252E+021.35013E+021.41028E+021.44153E+021.4B675E+021.51990E+021.69032E+022.24452E+022.28015E+022.32154E+022.39318E+022.44784E+022.53491E+022.59662E+022.68039E+022.74624E+02
TSPs 2 & l\Inactive
B.29776E+001.25001E+012.39854E+012.50390E+012.65102E+012.85479E+01.3.11S83E+O13.38327E+013.57531E+013.85679E+014.21025E+014.57386E+014.91663E+015.19907E+015.41526E+016.05711E+017.06316E+019.57002E+019.79375E+011.01002E+021.05086E+021.10094E+021.15084E+021.1B641E+021.23711E+021.29995EH021.36372E+021.42288E+021.47022E+021.52033E+02I.63128E+021.80180E+022.24763E+022.28230E+022.32974E+022.39309E+022.468B9E+022.53823E+022.58943E+022.66510E+02
TSPs 3 & AInactive
4.66164E+001.32502E+012.39716E+012.49075E+012.62533E+012.83516E+013.11056E+013.42886E+013.76846E+014.C8698E+014.31113E+014.58220E+014.89624E+015.16138E+015.31758E+015.41642E+017.29944E+019.56835E+019.77076E+011.00526E+021.04718E+021.10005E+021.15936E+021.22136E+021.27918E+021.31901E+021.36535E+021.41980E+0?1.46557E+021.49476E+021.52036E+021.84664E+022.24733E+022.27853E+022.32271E+022.38848E+022.46978E+022.55829E+022.64736E+022.72761E+02
-
Table 3. Natural Frequencies ot the TSP-active Modes of Tubes
BWG 12 and 18
Modes
123456789
1011121314151617180920
Tube BWG 12Hz
23.924.826.228.130.332.835.638.541.444.447.249.851.953.454.295.597.4
100.4104.2108.5
Tube BWG 18Hz
21.522.323.625.327.329.632.134.737.440.042.644.946.848.048.886.187.990.593.997.8
-
oFig. 1. Arrangement of Baffles and Support Plates
-
I ' ' I ' ' I ' 'I M i l
I I I f f' ' I '
! • • 1+-»
• ' m 11 ' i • • i ' •
i i i i
15 13 11 9 7
TRAIN A
TUBE FRETTING DAMAGEATSUPPORT
FLATTENED TUBE SURFACE
2/3/4 FLATTENED SURFACESFOUND ON ONE SPAN
flff i
Fig. 2. Damage Inspection of Removed Tubes
-
CCW HXs, Unit 1(Both Train A and B)
Daaage Appeared at Supports as Follows:
EX A
EL-3-1£1-26-2E5-1-2/3/4/5/8£5-4-1/2/3/4/6/9E5-51-3/4/6/9£5-54-1/2/3/4/6/8E29-26-6£53-1-1
HX B
EL-3-2/4El-26-2/3E5-1-2/4/5/6/7E3-4-2/3/4/5/6/7/9/10/12/13/14E5-51-1/2/3/4/5/6/7E53-4-8/9E53-51-5E53-38-6
El-3-2/4Row 1 (from bottom), tube 3 (fron left) at support
plates 2 and 4
Fig. 3. Location of 28 Tubes Selected from Each Heat
Exchanger
-
Fiq. 4. Peak-to-Peak Displacements as Function of Flow Rate
5000 1.000e+04 1.500e+04FLOW RATE, gpm
-
1 0 0 200 300 400
Frequency500 100 200 300 400 500
Fig. 5. Frequency Spectra of Tube 1H, a Typical Tube
-
Fig. 6. CCW Heat Exchanger Modified by TMMC
-
13
/ > A / W V A / \ / V A A A A A A y A . V ^ . A y y A^ A
A/NyV̂ y V \A/W*V/**/ VVVW yVW VW\A Ay^ A ^ V A A / V W V V V W
\ y V . y v y y ^ y \ A y V / \ y y y v
VVVV^AyY^y^;Ay^Av^r,/^/^AAAA^AA7^y^y^y^yvX/V^/AAAXXA/A/.
A/v^Ay^A/Vv^yV^yV^y•vVV^y^AAy^y^Ay^vVAAy^Ay^yvv^y^XAy\X\A,VVVV>AAyVV^AA/kyvV\Ay\AAAyNA/\/V\AAy\A/^yV^AA/VA/Vyv^A-AAy^Ay^A/\y^vv^y^y^/^V•\AyV^Ay\.AA^^y^yVV^AyV^y^
A X X A / /\.-••%Ayv\Ay^/VV
NAyV^yV^y^y^Ay^AAy^AyVV^y^y^Ay^AAA/v^/V^AAX
VV^AAXXAy^Ay•VVVV^AyXAyVN''
A.AyNy^cVAAy^yVVAyVAAy^AXy^AAAAAAAAj^ «- A ~ - - ' - AAAy%*
9
^y\Ay\yV
\A A * /s A X A A A \ X \ ' V V V ' - VV VV VVV V VVVN W V V W V
VV ' vf v v V V V-' VV V W v V W \AyOC^XYAYAi/VY^^^ 1 ,̂
AAv^Ay^Av^A.̂ A.VVVV^AAAv^A'
NAAA^y^AAy^y^AAAAXy^AOOOOvAA A ^ y ^ ^ ' ^ ' ^
\AAyVV^y\AA^%A/vV\AAAA^VAyV\A/Ay^WNyA^"A,A.AAA/WA-VWW\.XyVV\AAAA'WW/OO^^YXVVYV
, ' w ̂ '"^Wvv^A-V\/VVVV\Ay^AyvVAA-• AAAAyV^A,rtv^y\AyVV\A.XAA/Jr
ŷv̂ Aŷ A.-• •AyV^-^AAAAAAAAAAAy\AA.-.-AAA.\/V\A'5
AAAy\A_Ay\AA/\AA "AAA/
\A/VVVV\AvT^/\Ay\yv\AA AAyAA/\AAAA/\Ay\AAAAAA/\AAAy
\ A - - H p V ^ y
INLET INSTRUMENTATION DEPTH
A 25.5 !N. (TUBE 1-4 & 20, GROUP A)
• 65 IN. (TUBE 5-9 &19, GROUP B)
• 107 IN. (TUBE 10-15, GROUP C)
• 149 IN. (TUBE 16-18, GROUP D)
OUTLET INSTRUMENTATION DEPTH
• 25 IN. (TUBE 11 & 2', GROUP E)
X 64.25 IN. (TUBE 3'-5\ GROUP F)
• 106 IN. (TUBES 61 & T, GROUP G)
Fiq. 7. Instrumented Tubes
-
11C-H
4\; • • • ? * ! •
TV. 1 . . : . ,
Ullllu
J •:
fBjnriT
; i H 5 4 i
Liiiiiillllyi'iiiiiiiiiii,
' . • • • • ! •
gpm
145,000
11C-V
140,000
130,000
120,000
110,000
qj...j..4...j....j 100,000
90,000
76,000
•*3i:
J : : :
=..; (I
iuiOl
i! H
yil
T e l•••••!• •{••ii-
!G jn?
U : :"
..J....
.!*...:
iiiiliiibui....
60,000
140 0 140
Frequency, Hz Frequency, Hz
Fiq. 8. Frequency SDectra of Tube 11C
-
TSP-InactiveModes
mtnu
TSP-ActiveModes
Tube 1A-H10,000 gpm
140.
TSP-InactiveModes
ml nut
TSP-ActiveModes
Tube 2A-H145,000 gpm
P-Q.
140.
Fig. 9. TSP-Inactive and TSP-Active Modes
-
TSP-Inactive Modes
mlnuTube 5F-H1200 GPM
TSP-Active Modes
Tube 5B-V7000 GPM
Hz. 140.
Fig. io. Either TSP-Inactive or TSP-Active Modes Are
Dominant
-
TUBE 1A-V
0.1
EBUJQ
0.011000 10000
FLOW RflTE, gpm
100000
0.1
£
LUQ=3H
0.011000
TUBE 1A-H
10000
FLO ID RflTE, gpm
100000
Fig. 11. Displacement of Tube 1A
-
TUBE 5B-V
£UiQ
t o.i-Ja.z
0.011000 10000
FLOW RflTE, gpm100000
TUBE 5B-H
ESuuQ
to.i
z
0.011000 10000
FLOW RflTE, gpm100000
Fig. 12. Displacement of Tube 5B
-
0.1TUBE11C-Y
E£
o3
0.011000 10000
FLOW RflTE, gpm100000
0.1 TUBE 11C-H
10000FLOW RflTE, gpm
100000
Fig. 13. Displacement of Tube 11C
-
CM
CM
CO
-
1ST MODE, 23.9 HZ
3RD MODE, 26.2 HZ
4TH MODE, 28.1 HZ
5TH MODE, 30.3 HZ
TSP-flCTlUE MODES
Fiq. 15. The First Five Modes of TSP-Active Modes
-
1ST MODE, 11.3 HZ
2ND MODE, 23.9 HZ
3RD MODE, 24.9 HZ
4TH MODE, 26.5 HZ
5TH MODE, 28.4 HZ
TSP 2 SNflCTlUE
Fig. 16. The First Five Modes of TSP-Inactive Modeswhen TSP 2 is
Inactive
-
8.30 HZ
12.5 HZ
TSP2 RNDTSP 4 INRCTIUE
4.66 HZ
13.25 HZ
TSP 3 HND TSP 4 INflCTlUE
Fiq. 17. The Fi rs t Two Modes for TSPs 2 and 4 Inactive andTSPs
3 and 4 Inactive
-
TSP 2 INflCTlUE 11.31 HZ
TSP 3 INflCTlUE 9.68 HZ
TSP 4 INflCTlUE 9.53 HZ
TSP 5 INflCTlUE 9.52 HZ
TSP 8 INflCTlUE 9.52 HZ
Fig. 18. The First TSP-Inactive Mode for One TSP Inactive
-
FIRST
STOPPING
BANDFIRST
PASSINGBAND
SECONDSTOPPING
BAND
SECONDPASSING
BAND
'7/777.
OHZ 50 HZ 100 HZ 150 HZ
Fig. 19. Frequency Bands for Tube BWG12