Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Determination of The Crack Resistance Parameters

Download as pdf or txt
Download as pdf or txt
You are on page 1of 7

Journal of Theoretical and Applied Mechanics, Sofia, Vol. 48 No. 1 (2018) pp.

69-75
DOI: 10.2478/jtam-2018-0006

DETERMINATION OF THE CRACK RESISTANCE PARAMETERS


AT EQUIPMENT NOZZLE ZONES UNDER THE SEISMIC LOADS
VIA FINITE ELEMENT METHOD

V LADYSLAV K YRYCHOK ∗ , VASYL T OROP


The E.O. Paton Electric Welding Institute of the National Academy of
Sciences of Ukraine, Kyiv, Ukraine
[Received 16 October 2017. Accepted 12 March 2018]

A BSTRACT: The present paper is devoted to the problem of the assessment


of probable crack growth at pressure vessel nozzles zone under the cyclic seis-
mic loads. The approaches to creating distributed pipeline systems, connected
to equipment are being proposed. The possibility of using in common differ-
ent finite element program packages for accurate estimation of the strength of
bonded pipelines and pressure vessels systems is shown and justified. The au-
thors propose checking the danger of defects in nozzle domain, evaluate the
residual life of the system, basing on the developed approach.

K EY WORDS : Seismic analysis, crack growth, nozzle, vessel, distributed pipelines,


response spectrum.

1. I NTRODUCTION
Nowadays, the problem of the evaluation of residual resource of pressure vessels
with defects is not fully solved. This procedure is performed in order to know, how
much time the equipment can be exploited, without repairing or changing the parts.
For qualitative estimation in common situation, it is important to know the geometry
of the part with real or postulate defect locations, the material and its properties,
the history and parameters of loads, postulate loads, which characterize abnormal
situations during exploitation.
The present paper is devoted to postulate crack behaviour under the seismic loads
in heatexchanger, the part of nuclear power plant (NPP). In common situations, ac-
cording to [1], this procedure is not mandatory in project. But in practice, as men-
tioned in [2] and [3], fatigue and corrosion can initiate macro cracks in pipelines and
equipment, most of all in welded joints zones. Exploit organization must know, if
these objects are safe, or if it is necessary to change regimes, repair defects, make
diagnostic more often, etc. Some consideration of this problem can be noticed in
standard for pipelines [4]. Absence of material constants and necessity of experi-
mental research are the disadvantages of this methodology.

Corresponding author e-mail: k0965976337@gmail.com

Unauthenticated
Download Date | 4/12/18 2:28 PM
70 Vladyslav Kyrychok, Vasyl Torop

Guideline “VERLIFE” [5] was created by international society for NPP with re-
actor WWER-1000. It includes recommendations for strength estimation of tubes,
equipment and support elements. The recent version of the guideline suggests using
specific conservative constants of some material for defect growth rate evaluation. It
also allows us partially to take into consideration temperature regime, thermal and ra-
diation embrittlement. This feature is especially important for assessment of vessels,
which work over the project period. Taking into account the optional character of
guideline “VERLIFE”, Ukrainian engineers have some problems with the estimation
of fatigue and corrosion crack growth. When such cracks arise, every organization
solves this problem individually and uses its own analytical methods to forecast the
residual system resource. In our opinion, the most qualitative methodology is pro-
posed by American standard [6], which gives conservative crack resistance properties
for most of steels, used at NPP, depending on operation temperature, environment and
other parameters.

2. A PPROACHES TO ASSESSMENT OF SEISMIC LOADS IN PRESSURE VESSELS


In common situations, two approaches for seismic analysis of constructions are pop-
ular. They are response spectrum method and dependent on time dynamic analysis
method. The first approach, based on response acceleration of supports, depends on
resonance oscillation. To implement the method, it is necessary to perform modal
analysis and then, apply the response acceleration to the support elements at reso-
nance frequencies.
Assuming, that the seismic reactions in common event are periodic and not si-
nusoidal, we know this load is the sum of harmonic functions (Fourier range). That
is why, the supports can assume the full spectrum of response accelerations. This
hypothesis induces us to determine the response loads conservatively, as the sum of
all forces under each resonance frequency. Guideline [8] suggests getting conditional
resulting value, as square root of the sum of squares (SRSS) of all response loads.
Taking this into account, it can be noticed, that the larger frequency range is, the
larger reaction forces and stresses are. In practice, the range 0.5-30 Hz is taken into
consideration, others are neglected. Response spectrum method is used only for lin-
ear systems. If the model has non proportional dependencies (different stiffness in
different directions, gaps, dissipative nodes etc.), these parts are tried to linearize. In
this case, the solving accuracy is losing.
The dependent on the time dynamic analysis method is based on dynamic equa-
tions and seismic accelerogram, defined at peak ground acceleration. This approach
is suitable for solving nonlinear systems, but it is more sophisticated and much more
expensive than the response spectrum method. In fact, most manufactory objects of-
fer only response spectrum acceleration as output data for modelling. If the floor ac-

Unauthenticated
Download Date | 4/12/18 2:28 PM
Determination of the Crack Resistance Parameters at Equipment Nozzle Zones ...71

celerogram in such case is needed, it is possible to get it conservatively, performing


the dynamic analysis of structure and taking for boundary conditions the accelero-
gram “CA-482” at zero mark. The analytical form of “CA-482” can be found in
[8].
According to [9], nozzle zones and support elements are the parts, which can
be destructed during earthquake. The main cause of equipment branch connection
domains disruption are reaction forces of joined pipelines.
The seismic resistance assessment procedure for standard parts is simplified. It
induces us to evaluate only general bending and general or local membrane stresses.
Using these criteria, it is enough to operate only with beam model. Usually, in order
to receive general stresses at nozzle domain, it is accepted to make two steps. The
first one is the studying of joined pipelines in suitable codes to determine the reaction
forces and moments at nozzle. In this stage, it is possible to use the most popular
distributed piping solving program “Intergraph Caezar II” (in eastern Europe such
codes are popular, as DPIPE, ASTRA-AES and others). The second analysis is based
on studying the equipment via finite element method with more universal programs
(Ansys, Abaqus, Nastran etc.), using the output reaction forces and moments from
the first step.
The very important factor when getting the seismic stresses is the load orientation.
We need to take it into account, because the direction of response forces and moments
change during the period of time. If we get the seismic stresses, all the components
of tensor by using the SRSS rule are positive. If we consider the combination of
seismic and static regimes, we know exactly the direction of the last, because it does
not change during the time period. In this occasion, to get conservative stress values
of summary condition, the directions of seismic and static components should be the
same. If we need to evaluate the crack parameters, using the mentioned before two-
stepped modelling is not correct. This way is wrong, because during this simulation
compressive stresses in some elements arise. Determination of stress intensity factor
(SIF) is allowed only under tension loads. If we apply the forces and moments to
the 3D model of nozzle, at the first domain tension stress will arise, at the second
– compressive one. But this condition is only one of the probable situations. If the
crack is situated in compressive zone, conservative evaluation of SIF is impossible.
The second problem of the two-stepped modelling is the presence of general and
local stresses. They can have different directions and this condition will not allow
to determine maximum tension across the full crack front. To avoid all those men-
tioned aspects, it is necessary to study joining pipelines and pressure vessel together
in conjunction.

Unauthenticated
Download Date | 4/12/18 2:28 PM
72 Vladyslav Kyrychok, Vasyl Torop

3. C OMPUTATIONAL MODEL
3.
TheComputational model
object of the investigation was to determine the probability of a leak appear-
anceThein the defected zone of
object of the investigationthe heatexchanger after the
was to determine the earthquake. Fora this
probability of leakaim,
appearance in the defected
it was created the modelzone of the
of the heatexchanger
equipment aftertubes
and joining the earthquake.
between theFor this and
vessel
aim,the
it was
fullycreated the model ofTothesimplify
fixed supporters. equipment and joining tubes
the sophisticated between
pipelines the vessel
modelling, it was
and written
the fullythe fixed
Python code for import the objects from DPIPE into Abaqus/CAE.it The
supporters. To simplify the sophisticated pipelines modelling,
was main
written the
algorithmPython codeon
is based fortranslating
import thetheobjects from DPIPE
information into Abaqus/CAE.
from solving DPIPE file (node
The coordinates
main algorithm is based on translating the information from solving
of tubes, boundary conditions, load parameter, material properties, DPIPE file sec-
(nodetion characteristics) to the Python script, readable in Abaqus. To decreasematerial
coordinates of tubes, boundary conditions, load parameter, the solving
properties, section characteristics) to the Python script, readable in Abaqus. To
time, the straight piping cells are modelled as Beam elements (the mesh consists from
decrease the solving time, the straight piping cells are modelled as Beam elements
(the linear
mesh line elements
consists fromoflinear
type PIPE31). To consider
line elements of typethePIPE31).
ovalizationTostiffness
considerreduction
the
ovalization stiffness reduction of elbows, they are created as Solid parts (thehexahedral
of elbows, they are created as Solid parts (the mesh consists from linear mesh
elements
consists from oflinear
type hexahedral
C3D8R). The fragment
elements of of theC3D8R).
type geometryThe model is represented
fragment of the on
Fig. 1.model is represented on Fig. 1.
geometry

Fig. 1. Fig.
Fragment of modelled
1. Fragment heatexchanger
of modelled andand
heatexchanger joining pipelines
joining pipelines.

The body of the vessel is made from tube 325x12 mm, the nozzles from
The body of the vessel is made from tube 325 × 12 mm, the nozzles from pipe
pipe 168x9 mm. The operation pressure is 2.4 MPa. The temperature of the
168 × 9 mm.
environment The operation
is ◦150˚C. pressure
The material is 2.4heatexchanger
of the MPa. The temperature
and joining of the
partsenviron-
is
ment is 150 C. The material of the heatexchanger and joining
08H18N10T. Young’s modulus is 194 GPa. Density is 7900 kg/m . One of the parts
3 is 08H18N10T.

Young’s modulus is 194 3


supports of heatexchanger is GPa.
fully Density
fixed, theis others
7900 kg/m
work .asOne of the
sliding supports
guide of heatex-
in order to
changer
avoid is fully fixed,
the exceeding the others work
of temperature as sliding
stresses. Modalguide in order
analysis wastoexecuted
avoid theon
exceeding
the
rangeofof
temperature
0.5-30 Hz.stresses. Modal spectrum
The response analysis was executed onand
accelerations thedamping
range of 0.5–30 Hz. The
were taken
according
responseto [1].
spectrum accelerations and damping were taken according to [1].

Unauthenticated
Download Date | 4/12/18 2:28 PM
Determination of the Crack Resistance Parameters at Equipment Nozzle Zones ...73
If the container is made from stainless steel, the most widespread crack
initiation cause is intergranular corrosion at the interior side of the wall near the
If the container is made from stainless steel, the most widespread crack initiation
welded cause
joint.is At the same
intergranular time, atwe
corrosion postulated
the interior side ofthe defect
the wall nearin
thedomain, where the
welded joint.
maximum seismic
At the stress
same time, arises (see
we postulated the Fig.
defect2). The crack
in domain, wherewas semielliptical
the maximum seismic with 12
mm lengthstressand 9 (see
arises mmFig.depth.
2). TheTocrack
create
was the optimalwith
semielliptical finite element
12 mm length mesh,
and 9 mm the defect
depth. To create the optimal finite element mesh, the defect
zone is modelled as a detached cell, meshed via quadratic tetrahedral elements zone is modelled as a of
detached cell, meshed via quadratic tetrahedral elements of
type C3D10 and linear wedge element of type C3D6, bonded with the shell of the type C3D10 and linear
wedge element of type C3D6, bonded with the shell of the main vessel part via tee
main vessel part via tee constraint (the mesh of main part consists from linear
constraint (the mesh of main part consists from linear quadrilateral element of type
quadrilateral
S4R). Theelement of type
fully created S4R).
object The
consists fully
from created
49138 elements. object consists from
The determination of 49138
elements.the The solution
all needed modelof the model
parameters is on
is based at Linear
crackElastic
domain, based
Fracture on Linear Elastic
Mechanics.
Fracture Mechanics.

Fig. 2.
Fig.Location
2. Location and shape
and shape of crack.
of the the crack

The proposed approach allows to estimate static and dynamic loads of each ele-
The proposed approach allows to estimate static and dynamic loads of each
ment with minimal operational memory and Central Processing Unit requirements.
elementThewith minimal operational memory and Central Processing Unit
solution time of such and similar models is about several minutes.
requirements. The solution time of such and similar models is about several
minutes.4. R ESULTS AND ANALYSIS
In Fig. 3, the normal stresses across the crack front in cut section of defects domain
4. are represented.
Results The maximum normal stresses and SIF KI were determined as the
and analysis
algebraic sum of static
In Fig. 3, the normal (the stresses
own weight + pressure
across + temperature)
the crack front inand
defects obtained
cut from
section of
response spectrum analysis dynamic component.
domain are represented. The maximum normal stresses and SIF KI were determined
as the algebraic sum of static (the own weight + pressure + temperature) and
obtained from response spectrum analysis dynamic component.

Unauthenticated
Download Date | 4/12/18 2:28 PM
74 Vladyslav Kyrychok, Vasyl Torop

Fig. 3. Distribution of maximum stresses across crack shape under the


Fig. 3. Distribution of maximum stresses
simultaneously across
acted crack
static and shape under
seismic loadsthe simultaneously acted
static and seismic loads.
Basing on the stress field on surface of crack front, it is possible to get SIF by
formula on
Basing (1):the stress field on surface of crack front, it is possible to get SIF by
(1)
formula (1): 𝐾𝐼 = 𝜎 ∙ √2𝜋𝑟,
where σ is the normal stress at point of crack tip domain, acting across defect
√ front and this point.
(1) shape, r is the shortest distance between
KI = crack
σ 2πr ,
We have calculated, that the maximum SIF is 8 MPa·m1/2. It is more than
threshold stress intensity factor of the stainless steel (according to [6] KIth≈5.5
where σ is the normal stress at point of crack tip domain, acting across defect shape,
MPa·m1/2 at 150˚C). The crack growth after the earthquake can be found from Paris
r is the shortest
equation (2).distance
Our taskbetween crackfor
was solved front and this stress
zero-based point.cycle. It was taken in
We have because
calculated, thatnormal
the maximum SIF is 8exceed
MPa m 1/2 It is more than thresh-
account, seismic tension stresses static. components at crack
tip andintensity
old stress ΔKI can change
factor of the0stainless
from to maximumsteelSIF
(according
during the [6] KIth ≈ 5.5 MPa m1/2
toperiod.
𝑑𝑎
(2)◦ C). The crack =
at 150 growth 𝑛
С ∙ (∆𝐾after
𝐼 ) , the earthquake can be found from Paris equation
𝑑𝑁
(2). where
Our task
da/dNwasis solved for zero-based
crack elongation stress
per cycle, cycle.
C and material
It was
n are takenconstants, ΔKI is be-
into account,
cause range. normal tension stresses exceed static components at crack tip and ∆KI
SIFseismic
can changeMethodology [10] suggests
from 0 to maximum considering
SIF during 50 seismic cycles by determination
the period
the fatigue damage of material. For the regions with high probability of earthquake,
this value ought to be increased. daThe calculationnof crack growth was performed,
(2) using material constants C=2.38·10= -9 C (∆K I ) ,
dN (C produces fatigue crack grow rates in the
units of mm/cycle when ΔKI is in the units of MPa·m0.5) and n=3.3, according to [6]
wherewithda/dN
taking is
into account
crack the operating
elongation C and The
temperature.
per cycle, n arecomputation shows that∆K
material constants, the I is
maximum
SIF range. defect increasing by using such parameters is 0.11 μm. We can be sure,
this value is very small and safe for present equipment.
Methodology [10] suggests considering 50 seismic cycles by determination the
fatigue damage of material. For the regions with high probability of earthquake, this
value ought to be increased. The calculation of crack growth was performed, using
material constants C = 2.38×10−9 (C produces fatigue crack grow rates in the units
of mm/cycle when ∆KI is in the units of MPa m1/2 ) and n = 3.3, according to [6]
with taking into account the operating temperature. The computation shows that the
maximum defect increasing by using such parameters is 0.11 µm. We can be sure,
this value is very small and safe for present equipment.

Unauthenticated
Download Date | 4/12/18 2:28 PM
Determination of the Crack Resistance Parameters at Equipment Nozzle Zones ...75

5. C ONCLUSIONS
The review of standard methods for estimation of fatigue crack growth parameter is
represented. The approaches to seismic assessment of important equipment with de-
fects, using conservative material constants, are analysed. The principles of pressure
vessels with joining pipelines modelling are proposed. Basing on the example of
NPP heatexchanger, the opportunity of macro defect condition assessment in domain
of stress localization at the equipment after the earthquake is shown. The present ap-
proach can be suitable for damage estimation of equipment with large wall thickness,
containers manufactured from material with low fracture toughness, and vessels at
objects situated in seismic active regions.

R EFERENCES
[1] Rules and Regulations in Nuclear Power Engineering. Regulations for Strength. Anal-
ysis in Nuclear Power Plant Equipment and Piping. PNAE G–7–002–86, M., Energoat-
omizdat, 1989, 525 p. (In Russian).
[2] T OROP, V. M. The Pipelines, Containers and Pressure Vessels Limit Assessment, Based
on Developed Expert System, Dissertation of the Doctor of Technical Science, 05.02.09
/ NAS of Ukraine, – K., 2006. – 360 p. (in Ukrainian).
[3] D UB , A. V., V. A. D URYNIN , ET. AL . Development Ultrasonic Testing Procedures and
Determination of Performance of the Assembly of the Joints of Header to Steam Gen-
erator PGV-1000M, Technical Diagnostics and Non-destructive Control, No. 4, 2014,
36-51 (in Russian).
[4] DSTU-N B V.2.3-21:2008. Assessment of the Residual Strength of Magistral Pipelines
with Defects (In Ukrainian).
[5] Guidelines for Integrity and Lifetime Assessment of Components and Piping in WWER
Nuclear Power Plants (VERLIFE), 2013.
[6] ASME Boiler and Pressure Vessel Code, Section XI: Rules for In-service Inspection of
Nuclear Power Plant Components.
[7] Abaqus Scripting Reference Manual. – USA, Providence, RI., 2010, – 2825 p.
[8] RTM 108.220.37-81 Strength Calculation under the Seismic Load (In Russian).
[9] Seismic Resistance Analysis Technique of the Active NPP Elements via Seismic Margin
Assessment Method, MT-T.0.03.326-13 (In Russian).
[10] Technical Condition Assessment Technique, Basing on Strength Parameters and An-
alytical justifying of Pump Bodies Resource Remapping at Ukrainian NPP. MT-
T.0.08.158-05 (In Russian).

Unauthenticated
Download Date | 4/12/18 2:28 PM

You might also like