2000 ASME PVP Conference, Seattle

THE USE OF GASKET FACTORS IN FLANGE CALCULATIONS

Jaroslav Bartonicek Manfred Schaaf

GKN Gemeinschaftskernkraftwerk Neckar AMTEC Advanced Measurements
Neckarwestheim Lauffen
Germany Germany

Friedrich Schoeckle
AMTEC Advanced Measurements Hoher Steg 13
Lauffen 74348 Lauffen / N.
Germany Germany

ABSTRACT INTRODUCTION
A flange calculation has to fullfill three major tasks. First the GKN I is a nuclear power plant that was built in the early 1970s;
prestress value for assembly has to be determined. Then a tightness the plant is in operation since 1976. Using the steam generator
proof and a stress analysis have to be performed for every relevant handhole flanged joints of this PWR power plant as an example, the
state of operation. For reliable calculation results it is necessary that development in gasket factors in conjunction with flange calculation
all parts of the assembly (i.e. flanges, bolts and gasket) are regarded. methods are summarized and some "lessons learned" are extracted.
Using the example of a flanged joint of a steam generator of a In the first years of operation already, there were a few minor
nuclear power plant, the European development in gasket calculation troubles with leakage of handhole and manhole junctions of the
in recent years and the necessary gasket factors are summarized in steam generators. Fig. 1 gives an overview over the steam generator
this paper. Some general conclusions are drawn. and the handhole and manhole locations. Minor troubles in this case
The use of gasket factors plays an essential role in calculations for means small leakages, i.e. corrosion effects or some drops of water
flanged joints. Only if realistic gasket factors for the selected gasket near gasketed joints, especially after pressure tests. But as the leaks
of the joint (that have to be determined in tests explicitly) are used, were only small, the attempts to solve this problem did not get high
above tasks of the calculations can be met. In Europe realistic gasket priority.
factors for use in calculations are standardized in prEN 13555 and in Finally, in 1986 there was a major leakage at a handhole closure of
Germany in DIN 28090. one steam generator. The leakage caused a forced outage of the plant
Within the European standardization tasks there are two procedures for repair purposes, i.e. a time and thus money consuming procedure.
for the calculation of flanged joints, actually. One is based on the Additionally the autorized inspectors and the licencing authorities
ASME-code (incorporated in prEN 13445), the second one uses a became involved, thus it was necessary to provide a solution of the
limit load theory (prEN 1591). Whereas the ASME-procedure is only problem with adequate paperwork.
useful for formal stress analysis purposes, the EN 1591 procedure Fig. 2 shows the original version of the handhole construction,
provides the tools for stress analysis and tightness analysis (including Fig. 3 gives details about the gasket. 12 bolts (size M27) are used to
the output of a prestress value). The gasket factors that are necessary tighten this flange closure construction of nominal diameter
for this calculation are defined in prEN 13555. DN 200 mm. The gasket consists of a steel plate (outer diameter
In Germany there is another procedure incorporated in the 244 mm, thickness 7 mm) with kamm-profile; there is a 0.5 mm
standards for nuclear power plants (KTA 3201.2/KTA 3211.2); this silver topping on each side of the kamm-profile.
procedure is based on the German prestandard DIN E 2505. Stress Until that event mounting was done using torque wrenches; the
analysis and tightness analysis is possible with this procedure, if necessary torque was derived from design data; the calculations and
appropriate gasket factors (prEN 13555) are used. their results will be discussed in the next chapter. From the state of
Of course, FE-calculations are commonly used for more complex knowledge at that time, inaccurate mounting was assumed to be the
designs, too. Similar to the analytic calculation procedures, the use of reason for the leakage. Therefore the first goal was to optimize the
realistic gasket factors is essential in these calculations.

1
 2000 ASME PVP Conference, Seattle

steam nozzle
mounting procedure itself and as a second goal to provide training to
the mounting staff.
An optimal mounting can be achieved using hydraulic tensioners in
combination with a control of elongation of the bolts; therefore this
method was selected. It was necessary to modify the bolt design for
this purpose, Fig. 4. The new bolts have an additional tensioning manhole
thread at the free end and – for elongation measurement purposes - a
rod in their center. With a special hydraulic tensioning device it is
possible to prestress all bolts of the junction at the same time and
with the elongation measurement the force in the bolts can be
monitored during the entire prestress process (bolt elongation vs. feedwater
nozzle
force is calibrated). Principle schemes of the methods are shown in
Fig. 5 and Fig. 6
The selected mounting procedure has some significant advantages:
(a) All bolts are prestressed at the same time (one integrated ring of
12 tensioners). This results in a reliable prestress level and a small
scatterband of the bolt forces. With the verification of the forces by
elongation measurements the results are even better.
(b) The time necessary for prestressing of a joint with the hydraulic
tensioners in combination with the elongation mesurement (special
snap-on design of the transducers) is comparable to the time
necessary with torque wrenches (special tightening sequence in
several load steps). But it is not necessary for the mounting staff to handhole
stay near the joint; especially the loading procedure can be
supervised from a distance. Thus the radiation exposure of the
mounting staff could be reduced significantly.
(c) It is no problem to provide a prestressing protocol to the
authorized inspectors.
Parallel to the modifications to the tightening process, new
calculations were performed to verify the necessary prestress value of
the bolts and to demonstrate that the junction design is (still) in Fig. 1 : steam generator of GKN I
accordance with the standards (KTA 3201.2). The calculation will be
discussed in more detail in the chapter below.
O 370
The new tightening procedure was successful. In the following O 295
years there were no more leaks at these joints – until 1998. Corrosion
effects at the outer shell of a handhole near the gasket demonstrated,
that there was a (very small) leak, again. This time, mounting could
not be blamed because the responsible staff was able to prove the
correct mounting procedure with a protocol.
Finally, it was found that the metallic gasket plates were the reason
for the leakage. The plates showed tolerances in dimensions; the
thickness of the plate at the kamm-profile varied up to +/-0.25 mm
along the circumference whereas a tolerance of +/-0.1 mm was the
25

limit. O 43
2

The silver topping is not able to compensate for these tolerances O 30

47

completely. This was verified in full scale leak rate tests that were
performed similar to the gasket testing procedure as outlined in
7

prEN 13555. The test rig is shown in Fig. 7. Gasket stress was M27 O 200

controlled using bolts equipped with strain gages, the leak rates were O 204
O 244
measured using the pressure decay method.
85

The resulting leak rates vs. gasket stress are given in Fig. 8. The
R

gasket factor Qmin(L=0.01) (determined on the loading part of the leak

42

rate curve, see below) has a value of 65 MPa for the gasket with the
low tolerance wheras it is 10 MPa higher with the gasket with higher
tolerances. The unloading part of the leakage rate curve shows the
same behaviour, i.e. the curve of the gasket with the bad tolerances is
shifted to higher gasket stresses.
As a consequence of this event, the quality control procedures for
the parts of the flanged joints were modified. Meanwhile only
gaskets with low dimension tolerances are in the warehouse. Fig. 2 : handhole junction - old version (dimensions in mm)

2
 2000 ASME PVP Conference, Seattle

secundary - handhole (DN200)

7

tension nut
“X” “Y” O 190
O 204
O 244

1,5 Detail “X”

1x45°

F force form
F hydraulic
,5
R1 tensioner
7

nut
0,5

0,5
4
4°

5
O 244

closure
Detail “Y”

1 gasket
R
1
4°

5 O 190
O 204

flange
Fig. 3 : gasket for the handhole junction (dimensions in mm)

O 370
O 295
Fig. 5 : hydraulic prestressing (schematic plot)
25

O 43
2

O 30
47
7

M27 O 200
O 204
O 244
85

R
42

unloaded loaded

F = 0 F = 0
∆ l F
∆l = 0 ∆l = 0

Fig. 6 : elongation measurement with internal rod

Fig. 4 : handhole junction - new version (dimensions in mm) (schematic plot)

3
 2000 ASME PVP Conference, Seattle

difference pressure
transducer

temperature pressure
transducer transducer

reference volume

closure

bolts gasket deflection measurement

with strange gages transducers

Fig. 7 : test rig for leak rate measurements (pressure decay method)

gasket plate with kamm-profile (di: 204mm do: 244 mm)- silver topping thickness 0.5mm
full line: tolerance in plate thickness 0.12 mm / dotted line: tolerance in plate thickness 0.24 mm

100

10

p = 80 bar
leak rate in mg/m/s

0.1

0.01

0.001

0.0001
0 20 40 60 80 100

gasket stress in MPa

Fig. 8 : measured leak rates

4
 2000 ASME PVP Conference, Seattle

COMMENTS TO THE HANDHOLE PROBLEMS FLANGE CALCULATION

The steam generators of GKN I were designed in the late 60s/ early For reliable calculation results it is necessary that all parts of the
70s. For the secondary side of the shell non-nuclear codes as TRD, assembly (i.e. flanges, bolts and gasket) and their interaction are
AD and DIN were used. As shown in detailled analysis in 1986 (after regarded.
the leakage event) and 1996 (for a proof of component integrity) the
formulas actually used for dimensioning are almost the same as those
used during the design state. As a summary of these analysis it can be Tasks of a Flange Calculation
stated, that the design of the steam generators in general and thus the The first task of a calculation is to determine the prestress level of
handholes still meets the actual standard requirements, even if the the joint. A prestress value is necessary for every flanged joint,
(stress) limits have changed (to lower values) with the years. therefore, this part of the calculation has to be done in each case.
Focussing on the handhole junction, the procedures to analize The prestress level depends on the gasket used and on the change
flanged joints have been modified with the years. Starting with force of the loads between mounting and operation state. This change of
balances and very generalized gasket factors like k0,KD and k1 (as the loads depends on the external loads and on the stiffness of the
defined in the 1960s in DIN V 2505, see Tab. 1), the understanding parts of the joint. Therefore a calculation must be performed, even if
especially of the gasket behaviour and the necessary gasket factors standardized flanges and bolts are used. Within this task the
has changed. Additionaly, the calculation procedures were improved boundary conditions (respectation of the limits of the gasket in use)
with the years, the actual state is discussed below. have to be considered.
Nevertheless, even with the "simple" calculation procedures used The prestress level (assembly state) depends on the tightness
during the design state combined with a lot of experience choosing characteristics of the gasket (minimum necessary gasket stress after
the appropriate gasket factors (the gasket factors were not explicitly mounting and in operating state), on the stress limits of the parts of
determined in tests at that time), it was possible to analize the joint the flanged joint (flanges, bolts, gasket) and on the change of the
and to evaluate a prestress value for the mounting of the handhole gasket stress between assembly state and operation.
joints (bolt forcedesign: 70 kN). This force (transformed to torque) was The second task of a calculation is a stress analysis (prevent
applied to the bolts via torque wrenches during mounting until 1986. destruction for static loads), the third is a tightness analysis (to
After the leakage in 1986 a new analysis was demanded. As there control emissions, i.e. to maintain a demanded tightness class).
was a leakage, the old prestress value was doubted, additionally to The calculations for use in tightness analysis and stress analysis
the quality of the mounting procedure. Meanwhile more realistic have to
gasket factors were defined (similar to the actual definitions, see - use relevant and realistic gasket factors,
Tab. 1), and the calculation procedure was almost the same as it is in - regard stiffness of flanges, bolts and gasket,
the actual KTA-standard. - regard realistic operation loads like intenal pressure, external
However, there were no standards for the determination of the forces and moments, temperature, temperature distributions,
gasket factors and thus – from todays point of view – the gasket deformations etc.
factors still were not reliable because they resulted from different test - determine the necessary gasket prestress level for assembly and
procedures (if tested). For the gasket used in the handhole junction - determine gasket stress in operation.
for example, σVU=125 MPa was given in standard tables (note that
this value did not depend on a leak rate class). As Fig. 8 shows, this
value is high compared to realistic values, determined in modern State of the Art regarding Calculation
tests. With the gasket factors of 1986 (Tab. 1) a new prestress value In European standards, there are two procedures for the calculation
was determined: bolt force1986: 110 kN. This force was (and still is) of flanged joints with the gasket floating between the flange plates,
the resulting prestress value after tightening with the hydraulic Fig. 9. One is based on the ASME-code (incorporated in
tensioners. prEN 13445), the second one uses a limit load theory (EN 1591). In
In conjunction with the latest event, the handhole joint was Germany there is another procedure incorporated in the standards for
calculated once again, using the EN 1591 procedure and the realistic nuclear power plants (KTA 3201.2/KTA 3211.2); this procedure is
gasket factors determined in above tests (Fig. 8). For a leak rate class based on the German prestandard DIN E 2505.
of L=0.01 (which is regarded to be "very good" for the given The ASME procedure is more or less a dimensioning guideline
medium) QSMIN(0,01)=40 MPa was determined (note that the actual only applicable for a simplified stress analysis; it is not possible to
gasket factors that characterize the tightening behaviour depend on a perform a tightness analysis on this base. The gasket characteristics
tightness class). This value can only be achieved if a related are included by use of formal gasket factors (not explicitly proved in
Qprestress=90 MPa is applied during prestressing. Therefore, QSMIN(0,01) tests). It is not possible to determine the necessary prestress value. It
is the value that determines the tightness proof and Qprestress has to be is not possible to perform a tightness analysis.
considered in the stress analysis (the gasket factor QMIN(0,01) is less With the EN 1591 procedure and with the procedure according to
important within the actual calculation procedure, it can be used a a KTA it is possible to perform stress analysis and tightness analysis
start point for an iterative calculation). The calculation with these for flanged joints. Additionally, the necessary prestress values are
gasket factors provided a bolt force of 80 kN. provided, even the mounting procedure can be taken into account. All
The actual calculation results in a lower necessary prestress value relevant loads (operation states) of a flanged joint are considered; it is
mainly due to the lower values of the gasket factors used. But as the possible to include external forces and moments (torsion moments
junction design is able to bear the higher prestress value (from a only in KTA). The tightness analysis depends to a high degree on the
stress analysis point of view), it was decided to stay with the old use of realistic gasket factors. The necessary gasket factors are
value. defined in prEN 13555 (most definitions are similar to DIN 28090).
The actual demands on a reliable calculation and the author´s
experience are summarized in the following chapters.

5
 2000 ASME PVP Conference, Seattle

prEN 13445 K T A 3 2 0 1 .2 EN 1591

(ASM E ) K T A 3 2 1 1 .2
loads: loads: loads:
- in t e r n a l p r e s s u r e - in t e r n a l p r e s s u r e - in te r n a l p r e s s u r e
- fo r c e s / m o m e n t s - fo r c e s / m o m e n ts - f o r c e s / m o m e n ts
(via fic tive in te r n a l - te m p e r a t u r e - te m p e r a t u r e
pressure) d iffe r e n c e s d iffe r e n c e s
- r e la x a tio n

results : results : results :

- stresses - stresses - s tre s s e s
- d e fo r m a t i o n s / - d e f o r m a tio n s /
rotations rotations
- tig h tn e s s /le a k r a t e s - tig h t n e s s / l e a k r a t e s

Fig. 9 : calculation procedures layed down in European and German standards

AD-Merkblatt B7 DIN V 2505 DIN E 2505 prEN 13555 description

1977 1964 1986 / 1990
minimum assembly gasket
k0, KD k0, KD σVU QMIN(L) stress
minimum operation
k1 k1 σBU = m * p QSMIN(L) gasket stress
maximum assembly gasket
σVO QSMAX
stress
maximum operation
k0, KDδ σBO QSMAX gasket stress
elastic recovery
ED ED E0, KI intercept, slope
creep
− /∆V gC factor

Tab. 1 : gasket factors used for calculation

6
 2000 ASME PVP Conference, Seattle

With none of above calculation procedures, it is possible to perform CONCLUSIONS

a fatigue analysis, because neither of above calculation methodes Not only with the presented example, but also with the experience
allows to calculate local stresses explicitly. This can be done with and the development within the last 15 years (replacement of
appropriate FE-calculations, for example. asbestos, more rigid environmental demands etc.) there are a few
There are no standards for flanged joints with metal-to-metal lessons learned:
contact of the flange plates, neither for gasket factors (that are A flange leakage demonstrates deficits in preventive action. It is
different to those of the floating gasket types) nor for calculation much more efficient to prevent the reasons for a failure than to try to
(analysis) procedures. control the results of a failure. In other words, it is economically
highly reasonable to invest in measures to prevent leakage; thus the
cost of repair and of forced outages as well as the effort for control of
Gasket Factors the emissions can be reduced to a minimum.
The gasket factors as defined in DIN 28090 and prEN13555 (that Part of the preventive action is the analysis of the joint.
are used in EN 1591) are summarized in Tab. 1. These gasket factors In the analysis of flanged joints it is necessary to take all parts
can be classified in factors describing the tightening characteristics (bolts, flanges, gasket) into account. Every approach that neglects the
and factors describing deformation characteristics. interaction between the parts involves uncertainty.
The most important tasks of a flange calculation are determination
Tightening Characteristics of the prestress level, stress analysis and tightness analysis.
For every gasket there is a certain minimum gasket stress in the There are tools for the reliable analysis of flanged joints. In Europe
state of assembly (QMIN(L)), that is necessary to reach the requested there is the flange calculation standard EN 1591, that allows stress
leak rate (or tightness class). This minimum gasket stress is analysis as well as tightness analysis explicitly; the bolt forces for the
determined using the loading part of the curve e.g. in Fig. 8. mounting state can also be calculated. In a similar way this can be
During service, it is necessary to maintain at least a sufficient done with the calculation procedure provided in the German KTA-
minimum gasket stress in every relevant operating state (QSMIN(L)). standard.
This minimum gasket stress depends on the applied predeformation The use of realistic gasket factors that are determined in
of the gasket during mounting of the joint. The highest value of standardized tests is highly important in calculations.
QSMIN(L) equals QMIN(L); with an increase in predeformation of the Part of the preventive action is a qualified mounting, too. The
gasket during assembly the QSMIN(L) -value decreases. mounting procedure must meet the demands on tightness. The
Regarding tightening characteristics, QSMIN(L) and the related mounting procedure has to be considered in stress and tightness
Qprestress -value are the crucial input data into calculations. analysis, the necessary prestress value has to be a result of a
calculation. The mounting staff is involved in a quality process;
Deformation Characteristics therefore qualified personnel is necessary.
To prevent destruction of the gasket or drastic changes in
tightening capabilities, the upper limits of the gasket stress in the
state of assembly (QSMAX(RT)) and in operation (QSMAX(T)) have to
be regarded.
To determine the changes of the gasket stress between the state of
assembly and operation, the stiffness of the gasket - described using
the elastic recovery (representated by the slope KI and the intersept
E0) - is a necessary gasket factor.
Finally, creep and relaxation of the gasket under operating
conditions must be known, because this can result in a drastic
unloading of the joint. gC is the gasket factor, that takes this
characteristic into account.