17th European Conference on Fracture

2 -5 September,2008, Brno, Czech Republic

On the fatigue stress range calculations with on-line monitoring

systems in nuclear power plants

R. Cicero1-2, a, I. Gorrochategui3, b, S. Cicero1-2, c, J. A. Álvarez1

1
Departamento de Ciencia e Ingeniería del Terreno y los Materiales, Universidad de Cantabria,
Santander, Spain.
2
INESCO INGENIEROS S.L., Santander, Spain
3
Centro Tecnológico de Componentes., Santander, Spain
a
[email protected], [email protected], [email protected]

Keywords: monitoring system, fatigue, stress intensity, ASME, coupling stress

Abstract. Nuclear power plants are generally designed and inspected according to the ASME Code.
This code indicates the stress intensity (SINT) as the parameter to be used in the stress analysis of
components. One of the particularities of SINT is that it always takes positive values, independently
of the sign of the stress (tensile or compressive). This circumstance is relevant in the Fatigue
Monitoring Systems used in nuclear power plants, due to the manner the different variable stresses
are combined in order to obtain the final total stress range. This paper describes some situations
derived from the application of the ASME Code, shows different ways of dealing with them and
illustrates their influence in the evaluation of the fatigue usage through the application to a practical
example.

1. INTRODUCTION

Monitoring Systems (MS) constitute an alternative for the fatigue assessment of components and
structures. These systems allow performing such type of assessments in an automated real time way
and, for this purpose, require the record of all those parameters affecting the stress state of the
component being assessed.
When using MS, the coupling of the different stresses acting in the assessed component is a key
factor. Commonly, the different stress states (coming from the different loads) are added using the
so called Stress Intensity (SINT) parameter, as indicated by the ASME Code [1]. This practice can
lead to errors in the calculation of the total stress state, as shown below. Therefore, despite this
methodology simplifies the analysis, it sometimes provides results that are not always optimised.
The main objective of this work is to determine the effect of this simplification by comparison
with an assessment in which the appropriate stress coupling is performed, and its corresponding
effect in the fatigue evaluation of components.

2. STRESS INTENSITY CONCEPT

In conventional fatigue MS the different loads are analysed independently. Therefore, the stress
analysis requires the combination of the corresponding stresses. The direct addition of the different
stress components can lead to errors in the determination of the total stress of certain transients,
given that such addition is usually performed through the SINT parameter, as indicated by the ASME
Code [1]. The SINT is defined as the maximum absolute value of the differences between the
principal stresses (Tresca Criterion). This parameter is always positive, so it is not possible to
distinguish between tensile and compressive stresses:

1964
 17th European Conference on Fracture
2 -5 September,2008, Brno, Czech Republic

SINT = max(S1-S2, S2-S3, S3-S1) (1)

S1, S2 y S3 are the principal stresses.

If the sign of the stresses is not known, errors in the coupling process (and then in the total stress
calculation) can be produced.

3. STRESS COUPLING

An adequate stress coupling methodology must be applied when using MS due to their specific
stress evaluation methodology, which consists in the independent analysis of the different loads
acting in the component. In a conventional assessment, stresses are obtained through the analysis of
all the loads acting simultaneously, without considering any coupling effect. This methodology can
not be applied when using MS, given that they are based on transfer functions that are particular for
each given load and location. On one hand, assessments using MS methodology allow the
calculation of the stress state in a component subjected to different loads acting (or not)
simultaneously, and independently of the variations in the different loads. On the other hand, it
requires performing the stress coupling in order to determine the effect of the combination of the
different loads acting at the same time.
Furthermore, stress calculation using MS is performed through their separation in components,
depending on their respective origin, (i.e, pressure stresses, thermal stresses, differential pressure
stresses, thermal stratification stresses…).
The appropriate and precise methodology here proposed, relies in the calculation of the six stress
components of the different loads (Sxi, Syi, Szi, Sxyi, Syzi and Sxzi). These stress components have
their corresponding sign and a fixed direction. Therefore, the stress component of each load (i) in a
given direction (i.e. Sxi) can be coupled as scalar quantities, and then, the different components of
the total stress (i.e., SxTotal, SyTotal, etc) can be derived. With such six total stresses (SxTotal, SyTotal,
SzTotal, SxyTotal, SyzTotal and SxzTotal), the principal stress components (S1, S2 and S3) corresponding to
the combined effect of all loads acting simultaneously, can be obtained. Finally, this principal
stresses allow calculating the actual stress intensity (SINT). This methodology will be defined here as
“actual coupling”.
Nowadays, the most extended calculation methodology is based on the addition of the stress
intensities (SINT) from different loads. This procedure simplifies the analysis, but can lead to
overconservative results and therefore, noticeable reductions in the fatigue life of the components
being assessed. The low accuracy of this methodology, called here “positive coupling”, comes from
the addition of the stress intensities (SINT), a parameter derived from the difference between
principal stresses. The principal stresses generated by the different loads have not the same
directions, and that is the reason why a direct sum can produce significant errors. The “positive
coupling” only provides suitable results when the principal stresses from the different loads have the
same direction and sign, but that is not probable in any case.
Therefore, a correct stress coupling methodology is required for a consistent realistic fatigue
evaluation. In this work, “positive coupling” and “actual coupling” will be compared.

4. CASE STUDY: APPLICATION TO A NUCLEAR POWER PLANT COMPONENT

The methodology outlined above for stress evaluation of components using MS is the procedure
here proposed instead of the positive stress coupling. Both methodologies are compared by means
of a stress evaluation in the critical location of the feedwater nozzle of nuclear power plant. The

1965
 17th European Conference on Fracture
2 -5 September,2008, Brno, Czech Republic

feedwater system provides warm-up water to the vessel, being the water pressure the same to that
existing in the vessel. The vessel has two feedwater loops, with two nozzles in each one.

4.1 Geometry and materials

Figure 1 shows a scheme of the nozzle geometry. The Safe End New (SA 508 CL1) and the Safe
End Old (A-105 Gr. II) are made of a carbon steel [2]. The vessel wall and the nozzle base material
is a low alloy steel (ASTM A 336 cc1332 [3]). The primary thermal sleeve (SB-167) is an Ni-based
alloy [4]. Finally, the outer thermal sleeve (XM-19) and the cladding (ER308) are made of a
stainless steel [4]. All temperature dependent material properties at an average temperature of 325qF
are shown in Table 1
Safe end new

Safe end old

Outer thermal sleeve Primary thermal sleeve

Nozzle
Sparger
Cladding
Vessel

Figure 1. Nozzle scheme and materials at different locations

Table 1. Material properties at 325 ºF

Nozzle Forging Primary
Safe Cladding Outer Thermal
& Thermal
Ends Stainless Sleeve
Vessel Wall Sleeve
Elastic Modulus (psi) 27.95E6 26.55E6 26.875E6 26.875E6 29.725E6
Coefficient Of Thermal Expansion
7.4E-6 7.4E-6 9.9E-6 9.15E-6 8.0E-6
(in/in/qF)
Thermal Conductivity, (Btu/hr-ft-
31.95 23.35 9.95 7.85 9.7
qF)(1)
Specific Heat (Btu/lb-qF) 0.119 0.121 0.126 0.126 0.123
Density (lb/in3) 0.283 0.283 0.283 0.283 0.283
Poisson’s (4) Ratio 0.3 0.3 0.3 0.3 0.3

4.2 Loads

Among all the applied loads produced by the plant transients in the component, only variable loads
are considered in a fatigue assessment. Then, the fatigue analysis in these locations will be
performed taking into account the following loads:
- Pressure loads: derived from the pressure in the vessel.
- Mechanical loads: interaction with the rest of the pipes of the feedwater system. Mechanical
loads are shown in Figure 2 and defined in [5] and [6]. These loads vary linearly between zero
(shutdown) and the values shown in Figure 2 (normal operation).
- Thermal loads: derived from the thermal changes during the different transients.
Differential pressure load will not be considered in this assessment, given that its low value has
negligible influence in the total stress. Seismic loads are not considered in fatigue analysis; in case
of earthquake a particular evaluation must be done.

1966
 17th European Conference on Fracture
2 -5 September,2008, Brno, Czech Republic

Figure 2. Mechanical loads in the nozzle

4.3 Critical location

In fatigue analysis, the critical locations are considered those points in the components where the
fatigue usage is the highest. Stress analysis of the feedwater nozzle has been performed using a
finite element model [7] with the aim of detecting this point. The critical point has been indentified
in the safe-end, in the location shown as point 39 in Figure 3.

x
z
39
Y
y Z X y
a) b)

x
Figure 3. a) 2D axisymmetric model of the nozzle. b) Critical locations in the feedwater nozzle

4.4 Stress analysis

The stresses produced by the different loads are analysed in the Safe end of feedwater nozzle
(Figure 3).

4.4.1 Pressure stress

The resulting stresses (psi) in the critical locations caused by the pressure loads are gathered in
Table 2 (axes following Figure 3). A constant pressure of 1000 psi (easy scaled up or down to
account for the different pressures occuring during the different transients) was applied along the

1967
 17th European Conference on Fracture
2 -5 September,2008, Brno, Czech Republic

inside surface of the feedwater nozzle and the reactor vessel wall. Sx, Sy, Sz, Sxy, Syz and Sxz are the
six stress components referred to the mentioned coordinate system, and SINT the corresponding
stress intensity (1).

Table 2. Stresses (in psi) caused by pressure load

Sx Sy Sz Sxy Syz Sxz SINT
7187 4510 -501 -101 -379 -98 7723

4.4.2 Mechanical stress

In order to be conservative enough, the  angle defined in Figure 2 has to be selected as that one
that produces the highest SINT in the critical locations [6]. Table 3 shows the stresses (psi) obtained
for the critical locations corresponding to different  values. It can be seen that the maximum stress
is obtained when =0.
Table 3. Mechanical stresses (psi) for different  values
H Fx Fy (Fz in the SINT (psi)

(ft-lbs) (lbs) model)(lbs) Safe end
24500 0 -24500 0 4802
24500 22.5 -22635 9375 4550
24500 45 -17324 17324 3840
24500 67.5 -9375 22635 3704
24500 90 0 24500 3624

The stresses caused by the mechanical loads in the critical locations are gathered in Table 4.

Table 4. Stresses (in psi) caused by mechanical loads

Sx Sy Sz Sxy Syz Sxz SINT
72 267 8 2398 1 -49 4802

In order to demonstrate the conservatism of the “positive coupling”, it was calculated by means of
finite element analysis the stresses caused by the combination of mechanical and pressure loads.
The stress results (Table 5) reveals how mechanical and pressure loads add their effects in the Safe
End location (positive stress coupling).

Table 5. Stresses (in psi) caused by the joint action of pressure and mechanical loads
Sx Sy Sz Sxy Syz Sxz SINT
7259 4778 -493 2297 378 -147 9162

On the other hand, the conservatism of stress coupling is very noticeable in the Safe end location:
SPRESS.+MECH. < SPRESS.+SMECH.  9162 psi < 7723+4802 psi  9162 psi < 12525 psi

4.4.3 Thermal stress

The stress response (Green’s Functions) to a one degree temperature change (unit step change) is
used to determine the thermal stress for any temperature change occurring during transients.

1968
 17th European Conference on Fracture
2 -5 September,2008, Brno, Czech Republic

Applying the Green’s Function in the convolution integral is possible to obtain the thermal stress
in any condition. A thermal shock from 182.78°C (361°F) to 132.78°C (271°F) was applied to the
feedwater nozzle model, given that these values cover almost all the temperatures that occur during
most of the transients. The most representative transients are: start-up, shutdown, scram and power
reduction.
The stress history caused by thermal loads (thermal shock from 182.78°C to 132.78°C) is shown
in Figure 4.
12000

10000

8000

Sint
SX
stress (psi)

6000
SY
SZ
SXY
4000
SYZ
SXZ

2000

0
0 100 200 300 400 500 600 700 800 900 1000

-2000

time (sec)

Figure 4. Thermal stresses in the Safe End conditions (T = -90 ºF)

5. FATIGUE MONITORING SYSTEM APPLICATION

5.1 Fatigue evaluation in MS

The maximum allowable number of cycles for the component being analysed can be obtained using
the fatigue curves of the design Code [1] once the values of the applied stresses are known. When
the component is subjected to stress cycles of different amplitudes (due to the different transients
occurring in the plant), the fatigue damage produced by each amplitude level (ui) is calculated as the
relation between the number of cycles actually applied (ni) and the number of maximum cycles
allowed (Ni) at such amplitude.
ni
ui (2)
Ni

The failure criterion is given by the Miner´s law (3):

m
ni n1 n2 n n
¦N
1
  3  ...  m
N1 N 2 N 3 Nm
1 (3)
i

The total stress in a component is given by the stresses caused by the different loads occurring
during the transients. In the feedwater nozzle here analysed, the different types of stresses
considered and the corresponding methodologies used for their evaluation are briefly described in
Table 6. All the parameters (temperature, pressure, flow…) needed for the stress calculations are
obtained by means of the different instrumentation installed in plant.

1969
 17th European Conference on Fracture
2 -5 September,2008, Brno, Czech Republic

Table 6. Stresses considered. Expressions and some of their characteristics.

Usual way
Load / Type Particular
Origin of General expression
Stress load expression
calculation
Sp = A* Press
Pressure -Sp = pressure stress intensity SINT
Pressure in vessel I FEM Sp= f(pressure)
stress -A = constant calculated from stress report
-Press = reactor pressure
Sm = [(T - T0) / (T100 - T0)] · m
-Sm = mechanical stress intensity (SINT)
Interaction with the -Vm=_mechanical_stress_intensity_at
Mechanical rest of the pipes of temperature T100
I FEM Sm = f(local temperature) -T = temperature in the location of interest
stress the_feedwater
system -T0= reference temperature (mechanical
stress is null (0%)
-T100=temperature that produces the highest
mechanical stress (100%)
W W
St t ³ xT t  W ·'T W ·dW ST t ³ Grt (W )·'T W ·dW
Thermal load due to 0 0
Thermal
variations of the II FEM
stress - St(t) = thermal stress intensity (SINT) along time
water temperature
- T(W ) = temperature change in the interest location
- xT t  W = Grt(W) = response to a unit step thermal change (Green Function)

As seen in Table 6, the methodology used for the stress calculation varies depending on the type of
the stress: Type I or Type II. Type I stresses are calculated by means of transfer functions and the
stress intensity is linearly proportional to the applied load. Type II stresses needs the calculation of
the Green’s Function, that is, the response to a unitary change of the load (i.e. temperature in
thermal stresses). Using the Green’s Function, and applying the Duhamel (convolution) integral, the
stress associated to any other load is obtained.

5.2 Stress and fatigue damage assessment.

The stress assessment and its corresponding fatigue damage analysis (Miner´s law) in the critical
location is here evaluated by using two types of stress coupling:
a) Positive stress coupling (most common)
b) Actual stress coupling (considering the coupling between the stress components).
The fatigue damage is calculated for Start up + Shutdown transients, given that they are the most
representative and critical in terms of fatigue damage. The stresses caused by these transients in the
Safe end are shown in Figure 6. The fatigue damage result is shown in Table 7. The stress intensity
calculated by means of positive coupling is higher than the actual stress, and therefore it produces
more fatigue damage (12-13 times higher), as shown in Table 7.

Table 7. Fatigue damage, U

Positive coupling Actual coupling
Start up + Shut down 6,573 E-4 5,191 E-5

As shown in the table, the damage associated to positive coupling is more than ten times the damage
associated to the actual coupling.

1970
 17th European Conference on Fracture
2 -5 September,2008, Brno, Czech Republic

60000

50000

40000

30000
psi

20000

10000

0
0 5000 10000 15000 20000 25000 30000 35000 40000 45000 50000 55000

-10000

-20000

time (sec)

REAL COUPLING POSITIVE COUPLING

Figure 6. Total stresses applying different procedures of calculation

7. CONCLUSIONS

“Positive coupling”, which directly adds the stress intensity associated to the different loads, is the
most extended methodology for the stress assessment in MS of nuclear power plants. Its application
is a conservative method for the calculation of stresses and can lead to noticeable reductions in the
fatigue life estimation of the components of the nuclear power plant being assessed.
The “actual coupling” is the methodology here proposed for the stress assessment in the
evaluation of fatigue damage. This methodology is based on the calculation of the stress
components of the different loads, their addition, and the subsequent obtainment of the actual stress
intensity.
The application of both coupling options to a real case has demonstrated that the conservatism
associated to “positive coupling” can be noticeable.
The effort in the initial implementation of the “actual coupling” is bigger than the “positive
coupling”. However, after that, the stress and fatigue damage calculations are analysed with
computers and the computation time differences between “actual” and “positive” coupling is
negligible.

References
[1] ASME Boiler and Pressure Vessel Code, Section III. 1986 Edition
[2] GE drawing No. 107E6465, Sheet 1, “Feedwater Safe End & Thermal Sleeve,”
[3]_DE ROTTERDAMSCHE DROOGDOK MU. N.V. AFD. KERNENERGIE, TEK. No. G.214,
“STOMPEN_NOZZLES N.4 A_D,”
[4] GE drawing No. 107E6465, Sheet 2, SI File No. GAR-02Q-206.
[5] Equipos Nucleares, S.A. Design Report No. AR-4401, Rev.0, “Detailed Analysis of Feedwater
Nozzle,” 4/4/94
[6] GE Certified Design Specification No. 25A5481, Rev.1, “Reactor Vessel – Feedwater Nozzle,”
14/3/1994.
[7] ANSYS ® Release 11.0
[8] Structural Integrity Associates Calculation No. GAR-02Q-301, Revision 0, “Feedwater Nozzle
Finite Element Model and Green’s Function.”

1971