Nuclear Technology

ISSN: 0029-5450 (Print) 1943-7471 (Online) Journal homepage: https://www.tandfonline.com/loi/unct20

Finite Element–Based Vibration Analysis of

WWER-440 Reactors

Eberhard Alstadt & Frank-Peter Weiss

To cite this article: Eberhard Alstadt & Frank-Peter Weiss (1999) Finite Element–Based Vibration
Analysis of WWER-440 Reactors, Nuclear Technology, 128:1, 46-57, DOI: 10.13182/NT99-A3013

To link to this article: https://doi.org/10.13182/NT99-A3013

Published online: 10 May 2017.

Submit your article to this journal

Article views: 1

View related articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=unct20
FINITE ELEMENT–BASED VIBRATION THERMAL HYDRAULICS

ANALYSIS OF WWER-440 REACTORS KEYWORDS: finite element mod-

eling, fluid-structure interaction, vi-
EBERHARD ALTSTADT and FRANK-PETER WEISS* bration monitoring
Forschungszentrum Rossendorf, Institute of Safety Research
Postfach 510119, D-01314 Dresden, Germany

Received August 14, 1998

Accepted for Publication April 20, 1999

European WWER-type reactors. In some cases these vi-

brations resulted in serious and safety-relevant damage.
A finite element model describing the mechanical During core barrel motion at the Greifswald power sta-
vibrations of the whole WWER-440 primary circuit was tion, relative amplitudes of up to 5 mm between the core
established to support the early detection of mechanical barrel and the reactor pressure vessel ~RPV! ~see Fig. 1!
component faults. A special fluid-structure module was were found.4,6 These high amplitudes were a conse-
developed to consider the reaction forces of the fluid in quence of the plastic deformation of the so-called spring
the downcomer upon the moving core barrel and the pipe segments that are to transmit the clamping force of
reactor pressure vessel (RPV). This fluid-structure in- the reactor vessel head upon the upper flange of the core
teraction (FSI) module is based on an approximated an- barrel. Insufficient upper fixing of the core barrel also
alytical two-dimensional solution of the coupled system led to guide lug damages. Because core barrel vibration
of three-dimensional fluid equations and the structural
equations of motions. By means of the vibration model,
all eigenfrequencies up to 30 Hz and the corresponding
mode shapes were calculated. It is shown that the FSI
strongly influences those modes that lead to a relative
displacement between the RPV and the core barrel. More-
over, by means of the model, the shift of eigenfrequen-
cies due to the degradation or to the failure of internal
clamping and spring elements was investigated. Com-
paring the frequency spectra of the normal and the faulty
structure, one could prove that recognizing such degra-
dations and failures even inside the RPV is possible by
pure ex-core vibration measurements.

I. INTRODUCTION

At many nuclear reactors, on-line vibration monitor-

ing became a standard procedure of plant surveillance.1,2
This development was stimulated by the occurrence of
abnormal vibrations. For the first time, excessive core
barrel motion was observed and diagnosed at the Pali-
sades reactor.3 Control element and core barrel vibra-
tions are known to have occurred in the former East
German power plant Greifswald 4,5 and also in other East

*E-mail: F.P.Weiss@fz-rossendorf.de Fig. 1. Scheme of RPV of the WWER-440 with internals.

46 NUCLEAR TECHNOLOGY VOL. 128 OCT. 1999

 Altstadt and Weiss
was accompanied by a quasi-static axial shift of the core
barrel inside the vessel, a considerable part of the guide
lugs was eroded. Up to 18 mm of material was worn off.
Quasi-static axial shift is an especially safety-relevant
event because this can impede control rod insertion into
the reactor.
To detect the faults at an early stage, one needs a
theoretical vibration model of the whole primary circuit,
which permits
1. description of the normal vibrations of the com-
ponents, especially to assign the measured vi-
bration frequencies in neutron noise, pressure
fluctuations, or mechanical displacements to vi-
bration modes of the whole coupled mechanical
system
2. determination of the physical limits for frequency
shifts and amplitude changes as alarm thresholds
for on-line vibration monitoring
3. assessment of mechanical loads connected with
the failure of a certain component.
Correct description of the mechanical vibrations of
the RPV and its internals requires that the fluid-structure
interaction ~FSI! in the downcomer region be consid-
ered. The finite element technique has currently reached
the point where vibration modeling of the whole primary
circuit of a nuclear power plant ~NPP! is allowed.1,7
From the point of view of vibration analysis, the
WWER-440 is not comparable to standard Western-type
pressurized water reactors ~PWRs!. The most significant
differences are as follows:
1. The center of gravity of the RPV with all inter-
nals and the upper block is above the ring foundation.
2. WWER-440 reactors have six coolant loops, with
loop seals and main isolation valves in each leg.
3. Further, they dispose of six horizontal steam gen-
erators that are hinged at the ceiling of the so-called steam
generator box.
4. The removable core barrel is horizontally fixed
by eight guide lugs and in the vertical direction by six
spring pipe segments.
5. The fuel elements are arranged in a separate core
basket that is inserted in the core barrel.
6. Moreover, the special construction of the control
elements, consisting of the absorber part and the fuel fol-
lower, requires a lower core structure below the lower
core support plate to fix the fuel followers when the ab-
sorbers are inserted in the core.
Operational experience with mechanical damage has re-
vealed that the spring pipe segments, the guide lugs, and
the control elements constitute the weak points from the
point of view of flow-induced vibrations. Comparable
NUCLEAR TECHNOLOGY VOL. 128 OCT. 1999
VIBRATION ANALYSIS OF A WWER-440 REACTOR
weak points and due consequences, such as core barrel
vibrations with such large amplitudes, are not reported
regarding Western PWRs.
There were about 31 WWER-4400213 and WWER-
4400230 reactors in Finland, East Germany, and Central
and Eastern Europe. There are 27 units still in operation
~6 in Russia, 2 in Ukraine, 4 in Bulgaria, 4 in Hungary, 4
in the Czech Republic, 5 in Slovakia, and 2 in Finland!.
The utilities in these countries can benefit from the pre-
sented results.
II. THE FINITE ELEMENT–BASED VIBRATION MODEL
The model comprises the whole primary circuit, in-
cluding steam generators, loops, coolant pumps, main iso-
lating valves, and the RPV with its internals. It was
developed using the ANSYS finite element code on a
Hewlett-Packard workstation platform. The model has a
modular structure, so that various operational and assem-
bling states can easily be taken into account. The model
represented in this paper is for the WWER-4400213, but
a model for the WWER-4400230 is also available.
Experimental experience showed the mechanical in-
tegrity of the system not to be endangered by shell mode
vibrations but by beam mode vibrations in the fre-
quency range up to 30 Hz. Additional finite element cal-
culations yielded that all shell modes occur beyond
30 Hz, with the two lowest core barrel frequencies at 18
and 26 Hz excepted. However, these core barrel shell
modes are not coupled to the bending modes. Thus, it is
sufficient to assemble the model from one-dimensional
pipe elements. Figure 1 shows a topology scheme of
the finite element model for the RPV and all the inter-
nals: core barrel, core barrel support skirt, core basket,
and upper core structure. The model of the RPV head
considers the upper callotte, the control element stand-
pipe frame, and the standpipes themselves. Each of the
71 nodes connecting two elastic pipe elements has 6 de-
grees of freedom ~DOF!: ux, uy, uz, rotx, roty, and rotz.
The material properties of the pipe elements depend on
temperature. Different reactor components are mutually
connected by stiffness matrices ~12 3 12!, which, e.g.,
represent the core barrel guide lugs ~nodes 20 and 39!,
the ring foundation ~nodes 43 and 71!, or the spring pipe
segments between the upper flange of the core barrel
and the RPV head ~nodes 31 and 50!. The parameters of
those stiffness matrices were estimated in separate three-
dimensional calculations. Extreme stiff connections are
modeled by constraints.
In analogy to the RPV with its internals, the loop
model consists of pipe elements, a pipe tee and pipe el-
bow elements, structural mass elements ~which are for-
mulated as 6 3 6 inertia matrices!, stiffness matrices, and
constraints. The loop model includes the steam genera-
tor, the hot and cold legs, the main coolant pump ~MCP!,
and the hot and cold isolating valves. Special attention
47
Altstadt and Weiss VIBRATION ANALYSIS OF A WWER-440 REACTOR

was put on the hinge of the steam generator at the ceiling

and on the connection of the steam generator to the main
steam line. Both connections are represented by a sepa-
rately calculated stiffness matrix between the steam gen-
erator center node and the ground. Neglecting these
couplings, one would drastically underestimate espe-
cially the first eigenfrequency of the loop. The bearings
of the MCP and of the hot-leg valve are represented by
stiffness matrices too. The parameters of the loop model
could be adjusted using results from modal analysis ex-
periments performed at coolant loops at the Greifswald
NPP, Germany, and in the Dukovany NPP, Czech Repub- Fig. 3. Finite element model of the whole primary circuit.
lic, as well.
The vibration behavior of the RPV and the six cool-
ant loops cannot be strictly separated. The RPV is not a
fix clamping for the loops, and the loop inertia cannot be
last nodes of all six loops, respectively. Figure 3 shows
neglected for RPV motions. The total mass of the RPV
an element plot of the whole primary circuit.
with all internals is ;600 t, and that of one loop is
To consider the FSI in the downcomer, a fluid-
;290 t. So, in general, it is to be expected that the mode
structure element ~FSE! that can be implemented into the
shapes will be coupled. The connection of the RPV and
finite element model was developed.8,9 The model upon
the coolant loops is realized by rigid area constraint equa-
tions between the RPV ~node 47, Fig. 2! and the first nodes which the FSE is based is shown in Fig. 4. It is particu-
of all six loops and between the RPV ~node 44! and the larly tailored to incompressible fluids streaming through
a narrow annular gap, which is formed by two concentric
cylinders. The fluid flow is driven by a pressure gradient
p1 2 p2 . The cylinders can perform small motions ~two
displacements and two rotations!. Thus, the FSE has 8
structural DOFs. The fluid equations are simplified and
can be solved analytically considering the moving inner
and outer cylinders.10,11 Assuming a potential flow, the
continuity equation can be written as

F~r, w, z! 5 F~r, z!{e jmw

and

1 m2
]rr F 1 ]r F 2 2 F 1 ]zz F 5 0 , ~1!
r r

Fig. 2. Topology of the finite element model of the RPV and

its internals. Fig. 4. Fluid-structure element.

48 NUCLEAR TECHNOLOGY VOL. 128 OCT. 1999

 Altstadt and Weiss VIBRATION ANALYSIS OF A WWER-440 REACTOR

with the boundary conditions: The whole model consists of ;450 finite elements with
;2300 active DOFs.
]r F6r5R I 5 ]t wI 1 vz0 ]z wI
and
III. EIGENFREQUENCIES AND MODE SHAPES
]r F6r5R 0 5 ]t wO 1 vz0 ]z wO , ~2!
where The coupling of all six coolant loops with the RPV
results in a great number of eigenfrequencies of the com-
F 5 velocity potential plex system because of the coupling of different DOFs.
r 5 radial coordinate In the frequency range up to 30 Hz, .100 eigenfrequen-
cies are obtained. Many of them are close together but
z 5 axial coordinate exhibit different mode shapes. Most of the mode shapes
m 5 number of periods of F and w in circum- are characterized by various loop displacements in many
ferential direction different phase relations, with the RPV being almost in
the rest position. Some mode shapes exhibit only large
vz0 5 mean axial velocity displacements of the loops and of the RPV as well. For
wI , wO 5 radial displacements of the inner and the example, Fig. 5 shows the lowest mode shape at 3.4 Hz
outer structural boundaries that can be ex- with significant displacements of the RPV and of the in-
pressed by the motion of the centers of the ternals ~see also Table I, mode 1!.
cylinders. For the investigation of the RPV and its internals, it
is useful to reduce the model size by means of the sub-
The fluid velocity distribution obtained in this way structure technique to exclude such modes that differ from
is inserted into the Navier-Stokes equations to calculate each other only in the phases of motions of the loops.
the pressure that reacts on the cylinder surfaces: The six coolant loops are represented in the reduced model

F
p2 2 p 5 r{ ]t F 1
1
2
~¹F! 2 1
l
4s
vz0 F G , ~3!
by a substructure of 24 DOFs. Table I gives an overview
of the modes of the RPV and its internals up to 30 Hz.
The longitudinal modes ~z direction! excepted, generally
where each mode shape occurs twice ~the x-z and the y-z planes!.
To demonstrate how important the consideration of FSI
r 5 fluid density is, the eigenfrequencies are listed for two cases: with and
s 5 gap width without FSEs. The first eigenfrequency connected with a
significant relative displacement between the RPV and
l 5 Darcy friction coefficient. the core barrel was calculated to be at 26.3 Hz with-
The linearization of the pressure with respect to the struc- out FSI. FSI shifts that eigenfrequency down to 13.7 Hz,
tural DOF and the calculation of the resulting forces and which agrees with the measured value. This effect can
momenta acting on the cylinders provide additional in- also be seen in the forced vibration spectra ~Fig. 6!.
ertia, damping, and stiffness matrices for the structural Eigenmodes 9010, 13, and 14015 ~Figs. 7b, 7c, and
equations of motion, which represent the FSE ~Ref. 9!. 7d! are particularly important to detect possible degra-
The FSEs are located between the RPV and the core bar- dations of the internal clamping elements like the guide
rel @e.g., between nodes 37 and 16, 38 and 17, and 40 and lugs and the spring pipe segments. Mode 13 at 17.6 Hz
22 ~see Fig. 2!#. The equation of motion for the coupled is foremost a z vibration of all the internals, while the
fluid-structure system then reads as follows: vessel amplitude itself is rather small. The relative
@ Mw 1 SGu i Mw Fi # us ] 1 @ Bt 1 SGu i Bt Fi # us _ 1 @ Ku 1 SGu i Ku Fi # su
5 fr ~t ! , ~4!
where
Mw 5 structural mass matrix
Mw Fi 5 mass matrix of the i ’th FSE
Gu i 5 geometrical transformation matrix
t Bt Fi , K,
B, u Ku Fi 5 damping and stiffness matrices,
respectively
su, fr 5 vectors of structural displacements
and excitation forces. Fig. 5. Coupled mode of the primary circuit at 3.4 Hz.

NUCLEAR TECHNOLOGY VOL. 128 OCT. 1999 49

Altstadt and Weiss VIBRATION ANALYSIS OF A WWER-440 REACTOR

TABLE 1
Eigenfrequencies of the RPV and Its Internals at Operational Temperature

Frequency0Hz

Number FSE No FSE Verbal Description

1 3.34 3.34 In-phase pendulum motion of RPV, core barrel, and core barrel internals; first
2 3.43 3.43 beam mode of the upper unit in-phase; loops in-phase
3 4.63 4.63 In-phase pendulum motion of RPV, core barrel, and core barrel internals; first
4 4.70 4.70 beam mode of the upper unit in-phase; loops in antiphase
5 8.07 8.07 Small in-phase pendulum motion of RPV, core barrel, and core barrel inter-
6 8.1 8.1 nals; first beam mode of the upper unit in antiphase; large cold-leg displace-
ments
7 10.3 10.5 In-phase pendulum motion of RPV, core barrel, and core barrel internals; first
8 10.4 10.6 beam mode of the upper unit in antiphase; hot-leg displacements
9 13.7 26.3 Antiphase pendulum motion of the RPV with respect to all RPV internals,
10 13.7 26.5 elastic deformation of core barrel and core barrel internals, second modes of
standpipe frame and standpipes
11 15.6 15.5 Second beam mode of the CE standpipes
12 15.6 15.5
13 17.6 17.6 z-vibration of all internals, small z-amplitude of the RPV, in-phase pendulum
motion of RPV, core barrel, and core barrel internals, second beam mode of
the upper unit
14 22.8 43.7 Antiphase beam modes of RPV ~lower part! and core barrel; elastic deforma-
15 22.8 43.8 tion of core barrel internals

Fig. 6. APS of the horizontal RPV bottom displacement ~37uy!

~the bold curve is calculated with consideration of FSI
in the downcomer, and the thin curve is obtained with-
out FSI!.

displacements connected with the pendulum component

of this mode can be neglected. Modes 9010 and 14015
are related to antiphase motions of the RPV with respect
to all internal components. Furthermore, these modes are
characterized by elastic deformations of the internals. Fig. 7. Selected modes of RPV and internals.

50 NUCLEAR TECHNOLOGY VOL. 128 OCT. 1999

 Altstadt and Weiss
Fig. 8. Modes of a single coolant loop.
The modes of a single loop were calculated with the
loop finite element model assuming the coupling nodes
of the RPV to be in the rest position. The finite element
model of the loops has been adjusted with modal analy-
sis experiments performed at WWER-440 plants. In the
frequency range up to 30 Hz, ;20 eigenfrequencies were
found. Figure 8 depicts four mode shapes for an empty
loop ~without coolant! calculated with the model. This
model was used for the parameter tuning. For the simu-
lation of forced vibrations ~Sec. IV!, the coolant is con-
sidered by an additional mass per length. The mode shape
connected with the lowest eigenfrequency at 1.15 Hz is a
pure tangential displacement of the whole loop with re-
spect to the RPV. The mode shape at 2.1 Hz essentially
corresponds to a torsional vibration of the steam gener-
ator around the collector nozzles. At 4.2 Hz, the mode
shape is a radial steam generator vibration with respect
to the RPV. The eigenfrequency at 5.4 Hz is determined
by the beam mode vibration of the whole coolant pump
in the cold-leg plane.
IV. FORCED VIBRATIONS
To simulate forced vibrations, time-dependent exci-
tation is needed as input to the finite element model ~forces
or forced displacements at the nodes!. For the WWER-
440 model, four sources of excitation are relevant:
1. pressure fluctuations in the coolant caused by vor-
texes
2. residual unbalance forces of the MCPs
3. acoustic standing waves in the coolant fluid
4. small ground motions caused by neighboring units
or aggregates.
NUCLEAR TECHNOLOGY VOL. 128 OCT. 1999
VIBRATION ANALYSIS OF A WWER-440 REACTOR
The presented model does not consider the influence of
acoustic standing waves.
The pressure fluctuations in the downcomer were cal-
culated on the basis of downstreaming vortexes that orig-
inate at the six RPV inlet nozzles of the coolant loops.
The vortex field parameters ~take-off frequency, geom-
etry, and downstream velocity! follow from the model of
the Karman vortex street.12 The subpressure in a single
vortex was estimated using the two-dimensional Navier-
Stokes equation of an Oseen vortex.9 From this pressure
the resulting force acting at the instantaneous position of
the vortex is determined. The time-dependent nodal forces
in the x and y directions are determined superimposing
the downstreaming vortexes of all inlet nozzles.9 The
forces are applied to RPV nodes 37 through 43 and, with
the opposite sign, to the corresponding core barrel nodes
16 through 25 ~Fig. 2!. Figure 9 shows the auto-power
spectrum ~APS! of the force acting at nodes 43 and 25 in
the x direction.
The forces caused by the residual MCP unbalances
consist of three harmonic parts: the MCP revolution fre-
quency at 25 Hz and the subharmonics at 12.5 and
8.3 Hz. The subharmonics are caused by nonlinear MCP
shaft bearing effects ~e.g., the oilwhirl phenomenon due
to increased bearing gaps!. The MCP revolution fre-
quency can slowly drift in a narrow frequency range of
;0.5 Hz. Therefore, the force signals were modeled by
sliding harmonics. The forces are applied at the MCP
master nodes of the loop substructure. Figure 10 shows
the power spectrum of a generated force signal.
The ground motion can be modeled by a forced dis-
placement of node 71 ~Fig. 2!. The corresponding signal
is simulated by a pseudo random sequence with a root-
mean-square value of 0.1 mm.
The structural responses ~displacements of the nodes!
are calculated with a time-step width of Dt 5 0.013 s.
These time signals can be processed as if they were dig-
itized measuring series. The generated time series are
transformed into fast Fourier transform auto and cross
Fig. 9. Auto-power-frequency spectrum of the force acting on
node 43 in x direction and on node 25 in negative x
direction.
51
Altstadt and Weiss VIBRATION ANALYSIS OF A WWER-440 REACTOR
Fig. 10. Residual MCP unbalance. APS of the generated force
signal.
spectra. Figure 6 shows the APS of the displacement of
node 37 ~RPV bottom! in the y direction. In accordance
with the modal analysis ~Sec. III, Table I! the mode se-
lective influence of FSI is proven, which considerably
shifts the resonance peaks of modes 9010 and 14015.
In addition to the nodal displacements, combina-
tions of the DOFs can be calculated to simulate measure-
ments at any position. The fluctuation of the external
neutron flux, for example, corresponds to the change of
the water gap in the downcomer because of the neutron
transmission effect.4,13 Thus, the relative displacement of
the RPV and the core barrel nodes can be regarded as
approximately proportional to the ex-core neutron flux
fluctuation. Figure 11 shows the APS of the relative x
displacement between nodes 42 and 24, which corre-
sponds to the signal of an ex-core ionization chamber at
the upper core region at 0 deg. Consequently, this signal
clearly exhibits those resonance peaks that correspond to
modes with a significant relative displacement between
the RPV and core barrel ~modes 9010 and 14015; see
Table I!.
Fig. 11. APS of the relative displacement 42ux-24ux ~ioniza-
tion chamber in the upper core region, azimuthal po-
sition 0 deg!.
52
Fig. 12. APS measured during commissioning at Unit 5 of the
Greifswald NPP. ~The upper curve is the horizontal
displacement of the RPV bottom; the lower curve is
the vertical displacement at the RPV flange.!
Figure 12 shows the APS of real displacement mea-
surements taken at Unit 5 of the Greifswald NPP during
the commissioning tests. In this case, five main circulat-
ing pumps were running, and the coolant temperature was
1058C. The displacements were measured at the RPV bot-
tom in the horizontal direction ~upper curve! and at the
RPV flange in the vertical direction ~lower curve!, re-
spectively, by means of photooptical sensors. A compar-
ison of Fig. 6 ~blue curve! and Fig. 12 ~upper curve!
proves that all measured resonance peaks are well de-
scribed by the finite element model.
V. SIMULATION OF MECHANICAL FAILURES
To reveal which modes are best suited to detect deg-
radations of the safety-relevant internal components, sim-
ulation studies were performed by varying parameters of
the finite element model. To recognize the basic tenden-
cies in frequency and amplitude shifts when RPV clamp-
ing elements degrade, it is sufficient to use the reduced
model with the approximated description of the loops.
The calculation of the forced vibration spectra is based
NUCLEAR TECHNOLOGY VOL. 128 OCT. 1999
 Altstadt and Weiss VIBRATION ANALYSIS OF A WWER-440 REACTOR

Fig. 13. APS of x, y, and z RPV bottom displacement. ~The light lines are the normal state; the dark lines are with one guide lug
broken.!

NUCLEAR TECHNOLOGY VOL. 128 OCT. 1999 53

Altstadt and Weiss VIBRATION ANALYSIS OF A WWER-440 REACTOR

Fig. 14. APS of x, y, and z RPV bottom displacement. ~The light lines are the normal state; the dark lines are with two guide lugs
broken.!

54 NUCLEAR TECHNOLOGY VOL. 128 OCT. 1999

 Altstadt and Weiss VIBRATION ANALYSIS OF A WWER-440 REACTOR

Fig. 15. APS of x, y, and z RPV bottom displacement. ~The light lines are the normal state; the dark lines are with two spring pipe
segments degenerated.!

NUCLEAR TECHNOLOGY VOL. 128 OCT. 1999 55

Altstadt and Weiss VIBRATION ANALYSIS OF A WWER-440 REACTOR
on transient analysis ~simulation of time series! followed
by digital signal processing ~Sec. IV!. The FSI was re-
garded in all these calculations.
Figure 13 shows the RPV bottom displacements in
the x, y, and z directions for the normal state of the re-
actor ~light curves! and for the case where the guide lug
at the 0-deg azimuthal position is broken or is not in
contact with the core barrel ~dark curves!. In total, there
are eight guide lugs with 45 deg of azimuthal division.
As expected, above all, the eigenfrequencies pertaining
to antiphase pendulum motion do respond to that fail-
ure. There is almost no effect on the other eigenfrequen-
cies. Moreover, the APS of the x direction remains nearly
unchanged, whereas in the y direction a significant shift
of the resonance peak ~corresponding to mode 10! can
be seen. The reason for this is that a guide lug acts as a
tangential spring so that the 0-deg guide lug contributes
only to the y stiffness. Thus, there is no frequency shift
of mode 9.
Figure 14 indicates how the forced vibration spectra
of RPV bottom displacements in all directions would
change in the case of two broken guide lugs at 0 and
45 deg. The change of the spectra can clearly be ob-
served in the x direction as well as in the y direction.
Even the spectrum of the z vibration provides a signifi-
cant change, though the guide lugs do not have a stiff-
ness in that direction. This is a result of the coupling of
different DOFs.
Figure 15 shows the change of the spectra of the RPV
bottom displacement in all directions in case of the failure
of two spring pipes ~Fig. 1!. In total, there are six 60-deg
spring pipe segments. A significant downshift of the ei-
genfrequency of mode 13 at ;18 Hz and a rather small fre-
quency shift of modes 9010 are observed. The spectrum
change can most clearly be observed in the z direction be-
cause the pipe segments mainly act as a vertical spring, and
mode 13 is a large z vibration of the RPV internals.
VI. CONCLUSIONS
The results obtained with the vibration model for a
VVER-440 reactor hint at the opportunity to establish a
monitoring procedure that is capable of distinguishing
between guide lug and spring pipe failures because these
two types of failures are mapped in different and well-
separated frequency regions. Moreover, it seems to be
possible to detect these failures by pure ex-core vibra-
tion measurements, which is important since in-core vi-
bration measurements are difficult to implement over
longer time periods. When measuring the displacements
for the x and y directions, even the localization of the
failed guide lug might become achievable.
Considering FSI in the vibration model is necessary
because especially the modes related to a relative dis-
placement between the RPV and core barrel are strongly
affected by the flowing coolant.
56
At present a project, funded by the European Union,
aims at improving vibration diagnostics at the Dukovany
power station. Adjusting the theoretical model to that spe-
cial unit and generating the basic vibration signatures for
damage states are part of the project.
The developed model is also an appropriate tool for
considering the possible consequences for the primary
circuit after an earthquake or a chemical explosion. The
corresponding loads are simulated by accelerations or dis-
placements acting on the system at the ring foundation
and other contact points to the building.
ACKNOWLEDGMENT
The project this paper is based on was funded by the
Bundesministerium für Forschung und Technologie and is reg-
istered with No. 1500916.
REFERENCES
1. A. TRENTY, “Operational Feedback on Internal Structure
Vibration in 54 French PWRs During 300 Fuel Cycles,” Proc.
7th Symp. Nuclear Reactor Surveillance and Diagnostics
(SMORN VII), Avignon, France, June 19–23, 1995.
2. H. STÖLBEN and H. J. WEHLING, “Vibration Monitor-
ing of Kraftwerk Union Pressurized Water Reactors—Review,
Present Status, and Further Development,” Nucl. Technol., 80,
400 ~1988!.
3. D. N. FRY, R. C. KRYTER, and J. C. ROBINSON, “Analy-
sis of Neutron Density Oscillations Resulting from Core Barrel
Motion in the Palisades Nuclear Power Plant,” ORNL-TM-
4570, Oak Ridge National Laboratory ~1974!.
4. P. LIEWERS, W. SCHMITT, P. SCHUMANN, and F.-P.
WEIß, “Detection of Core Barrel Motion at WWER-440 Type
Reactors,” Prog. Nucl. Energy, 21, 89 ~1987!.
5. E. ALTSTADT and F.-P. WEIß, “Experimental and Nu-
merical Investigation of Control Element Vibration During Ab-
normal Core Barrel Motion at a VVER-440 Type Reactor,” Proc.
Int. Simulators Conf. (Simulators X), Arlington, Virginia, 1993,
Vol. 25, p. 4.
6. E. ALTSTADT, M. SCHEFFLER, and F.-P. WEISS, “A Vi-
bration Model for the Primary Circuit of VVER-440 Type Re-
actors Based on Finite Elements,” Trans. SMiRT13: Int. Conf.
Structural Mechanics in Reactor Technology, Porto Alegre, Bra-
zil, August 13–18, 1995.
7. V. BAUERNFEIND, “Vibration Monitoring of a Four-
Loop PWR: Model-Investigation of the Sensitivity of the Mon-
itored Signals on Mechanical Failures,” Prog. Nucl. Energy,
21 ~1987!.
8. G. GRUNWALD and E. ALTSTADT, “Analytical and Ex-
perimental Investigations for Modelling the Fluid-Structure In-
teraction in Annular Gaps,” Symp. Fault Detection Supervision
NUCLEAR TECHNOLOGY VOL. 128 OCT. 1999
 Altstadt and Weiss VIBRATION ANALYSIS OF A WWER-440 REACTOR

and Safety for Technical Processes (SAFEPROCESS 94), Hel-
sinki, Finland, 1994, IFAC.
9. E. ALTSTADT, G. GRUNWALD, M. SCHEFFLER, and
F.-P. WEISS, “Analytische Modellierung mechanischer Schwin-
gungen von Primärkreiskomponenten des Druckwasserreak-
tors WWER-440 mit finiten Elementen,” FZR-172, Final Report
of the BMBF Research Project No. 1500916 ~Mar. 1997!.
10. S. S. CHEN, “Dynamics of a Rod-Shell System Convey-
ing Fluid,” Nucl. Eng. Design, 30, 223 ~1974!.
11. D. MATEESCU and M. P. PAIDOUSSIS, “The Unsteady
Potential Flow in an Axially Variable Annulus and Its Effect
on the Dynamics of the Oscillating Rigid Center-Body,” J. Flu-
ids Eng., 107, 421 ~1985!.
12. Y. N. CHEN, “60 Jahre Forschung über die Kámánschen
Wirbelstraßen,” Schweizerische Bauzeitung, 91, 44 ~1973!.
13. J. A. THIE, “ Power Reactor Noise,” 81-65375, American
Nuclear Society ~1981!.

Eberhard Altstadt ~MS, applied mechanics, 1985, and PhD, identification

of mechanical systems, 1989, Dresden Technical University, Germany! is a senior
scientist at Forschungszentrum Rossendorf. His background includes mechani-
cal vibration analysis of nuclear reactors, nonlinear finite element analysis of
reactor internals exposed to accident loads, mechanical vibration tests at NPPs,
and theoretical work on FSI.
Frank-Peter Weiss ~MS, physics, Dresden University of Technology, 1973;
PhD, reactor physics, GDR Academy of Sciences, Germany, 1986! is the director
of the Institute of Safety Research of Forschungszentrum Rossendorf. At the same
time he is a professor at the Dresden University of Technology. His background
includes nuclear reactor physics, light water reactor thermal hydraulics, materi-
als safety, and diagnostics of nuclear reactors.

NUCLEAR TECHNOLOGY VOL. 128 OCT. 1999 57