See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/343112327

Design of Bolted Flange Connections -Status of EN Standards

Presentation · July 2020

CITATIONS READS

0 351

1 author:

Fernando Lidonnici
SANT'AMBROGIO SERVIZI INDUSTRIALI SRL
8 PUBLICATIONS   3 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

EPERC=European Pressure Equipment Research Council View project

EPERC=European Pressure Equipment Research Council View project

All content following this page was uploaded by Fernando Lidonnici on 21 July 2020.

The user has requested enhancement of the downloaded file.

 Design of Bolted Flange
Connections – Status of EN
Standards

16/05/2018 Sant'Ambrogio Servizi Industriali s.r.l. 1

 Which are the conditions to be fulfilled
in order to seal a gasketed joint ?

A gasket is made of a relatively soft and plastic material (soft and plastic
at least if compared to the material of the gasket seats). When this material
is compressed against the seats by the tightening load it has to fill all the
possible channels through which the internal fluid, subject to its
pressure, may reach the outside of the pressure chamber.

Therefore there are two conditions in order to get a tight joint:

• During the assembly of the connection the gasket must be subject to a

pressure y (MPa) sufficient to get its yield limit in the local areas in
contact with the seats, so that all the possible leak channels are closed
(this situation is generally known as gasket seating).

• During test and operation the residual pressure on the gasket must
be m times greater than the internal pressure in order to avoid the loss
of contact between gasket and seats.

16/05/2018 Sant'Ambrogio Servizi Industriali s.r.l. 2

 Parameters that influence the correct gasket seating (1)
• Physical state of the fluid to be sealed. While it is theoretically possible to have a perfectly
tight joint when the fluid to be sealed is a liquid, it is virtually impossible to get a tight joint
when the fluid is a gas or a vapour. In fact liquids have a much higher viscosity than gases, so
that the pressure drop of a possible flow through the microscopic leak channels at the surface
between the gasket and the seats is so high that a leak cannot occur. Gases and vapours, on
the contrary, tend to fill all the available space, whichever is the volume and the available flow
area, so that their leak tightness can be evaluated only on the basis of a maximum allowable
flow (generally given as mg/s/cm of gasket perimeter)

• Surface roughness. It has to be noted that the metallic surface of the seats (generally
obtained by mechanical tooling) is very far from being perfectly smooth. Its appearance, if
examined by means of a microscope, is like in the following figure:

x  lm

y 2  x  dx
1
Rq 
lm 
x0

The formula above gives the “roughness” of the surface, expressed in RMS degrees (= Root
Mean Square), generally given in microinches, but can also be given in μ (1 microinch =
0,0254 μ). On the construction drawings the roughness is generally indicated by means of
reversed triangles: 1 triangle up to 1000 microinches RMS, 2 triangles up to 400, 3 triangles
up to 60, 4 triangles up to 8).
16/05/2018 Sant'Ambrogio Servizi Industriali s.r.l. 3
 Parameters that influence the correct gasket seating (2)
• Difference between the hardness of the gasket and the hardness of the seats. In order
to fill the gaps between the gasket and the seats the compressive load acting on the gasket
must cause a plastic deformation of the same. This load will of course be higher if the gasket
is hard. Note also that in the case of metallic gasket this condition is difficult to be achieved,
particularly in the case of strain hardening materials (30 points of difference in the Brinell
hardness are generally recommended). Gaskets harder than the seats will cause a
permanent print on the seats which can compromise leak tightness in case of gasket
replacement.

• Effective area in contact with the seats. This area is influenced by the amount of
compressive load, which is causing an overall bending of the flanges during bolt tightening
and subsequent application of pressure. The contact surface will be therefore reduced and
displaced towards the outside because of the flange bending:

16/05/2018 Sant'Ambrogio Servizi Industriali s.r.l. 4

Loads and deformations on a flange assembly in service conditions

16/05/2018 Sant'Ambrogio Servizi Industriali s.r.l. 5

 Equilibrium of a flange assembly in gasket
seating and service (or test) conditions

Gasket seating (Bolt load = gasket Load)

Wat  H G

Service / Test (Bolt load = hydrostatic end load on flange

ID + hydrostatic load on flange ring + gasket load)

Wop  H D  H T  H G

16/05/2018 Sant'Ambrogio Servizi Industriali s.r.l. 6

 What is the relationship among the bolt load applied during
the assembly (“Bolting up”) and the bolt loads existing in
the subsequent operating and test conditions?
A bolted joint is a complex system composed by two flanges (or by a flange and a cover), a gasket and a
series of bolts. Bolts and flanges are metallic components whose behaviour under stress is
generally elastic, while the behaviour of the gasket is only partially elastic, may be different in first
time loading and subsequent unloading / reloading, and may also be subject to creep under
constant load. The figure shows the variation of the original gasket thickness eGt under the bolting up
load Qo and subsequent unloading to the bolt load Ql. The inelastic behaviour of the gasket and the
variation of the gasket surface under the bolt load are the main reasons why it is nearly impossible
to establish such a relationship.

16/05/2018 Sant'Ambrogio Servizi Industriali s.r.l. 7

 Is it really possible to control the bolt load in order to
have the reasonable assurance that the gasket (at least
at bolting up) has been subject to a predefined load?
The reality is that all the devices used to tighten a gasketed joint have a positive and a
negative tolerance. If bolts are tightened by a manual wrench (or even a torque wrench)
the relationship between the torque and the load is strongly dependent on the friction
factor (the one between the bolt threads and the nuts and the one between the nut and the
flange rear face). Supposing that the two factors have the same value μ, the following
formula gives the torque Mo (Nmm) needed in order to develop a load Fo (N) on a bolt
having a nominal diameter dBO (mm) :

M o  1,2  d BO Fo
Note that μ may have values between 0,15 and 0,25, so that the range is not negligible,
also in the cases where you are able to measure the real value of the torque. But also
those devices (generally hydraulic) which put the bolts under direct load have a
positive or negative tolerance. Although the scatter value of the load on a single bolt
will decrease if you consider a bolted assembly made of several bolts, you should
consider a minus tolerance on the applied bolt load for the purpose of gasket calculation
(a gasket not sufficiently compressed will leak), and a plus tolerance for the purpose of
determining the bolt stress.
16/05/2018 Sant'Ambrogio Servizi Industriali s.r.l. 8
 The Taylor Forge method – Conventional gasket width
The Taylor Forge method was developed in U.S.A. in 1930, but it is still used in the great
majority of Pressure Vessel standards in the world (ASME Section VIII division 1 and 2,
PD 5500, CODAP, ISPESL VSR and EN 13445.3 - Clause 11). It starts from the simplified
assumption that the minimum required bolt load has to be calculated separately for all the
possible service conditions (Gasket seating, Operation, Test), thus ignoring any relationship
among them. The calculation starts from the determination of a conventional effective
gasket contact surface. The loads on such surface have to be calculated using conventional
gasket factors m and y (MPa). The table of the gasket factors is still basically the same
developed by Taylor Forge in 1930. In the following equations the symbols and the equations of
EN 13445.3 Clause 11 have been used.

• Basic conventional gasket width for flat gaskets bo  w / 2

(w= effective geometric gasket width)
• Basic conventional gasket width for ring joint gaskets bo  w / 8
• Gasket width to be used for calculations:

b  bo when bo  6,3 mm else b  2,52 bo

Note: this formula is valid for S.I. units only

• Effective gasket sealing diameter for flat gaskets: G  Gasket OD  2b

(for ring joint gasket it is the average gasket diameter)
16/05/2018 Sant'Ambrogio Servizi Industriali s.r.l. 9
 The Taylor Forge method – Calculation
of the minimum required bolt area

• Bolt load for Gasket seating Watm   bGy

G2
• Bolt load for Service Wop   P  2 bGmP
4
G2
• Bolt load for Test Wtest  Ptest  2 bGmPtest
4

 Watm Wop Wtest 

• Minimum required bolt area Ab  max  , , 
 f f f 
 bo b b test 
• Maximum recommended bolt y De2  Di2 
area (in order to avoid gasket Ab 
crush) Note: for flat gasket without compression stop 2 f bo
De and Di = Gasket OD and ID
16/05/2018 Sant'Ambrogio Servizi Industriali s.r.l. 10
The Taylor Forge Method – Calculation of flange moments

 Ab  Ab min 
• Bolt load for flange Bolting-up: W 'atm   f bo
 2 
Note: this load is greater than the bolt load Watm used to determine the bolt area in gasket seating

• Bolt loads for Service/ Test:

G2
H G  H T  H D  Wop HT  H D   P H G  2 bGmP
4
• Flange moment for Bolting-up: M atm  W 'atm hG

• Flange moment for Service/Test: M op  H G hG  H T hT  H D hD

Note: use design pressure for service, test pressure for test conditions

16/05/2018 Sant'Ambrogio Servizi Industriali s.r.l. 11

 The Taylor Forge Method – Calculation of flange stresses

• Correction factor for bolt spacing δB:

Note: dB is the bolt diameter, e is the flange thickness, B is the flange ID

M atm / op / test
• Reduced calculation moments: M
B
• Stresses for integral flanges (H=Hub, r = radial, θ = tangential):

Note: for factors φ, βF, βγ, λ, K, lo or equivalent (different standards use different symbols) see
the relevant formulae or graphs. In PD 5500 and EN 13445.3 Clause 11 the nominal flange
design stresses are reduced by a factor k=2/3(1+B/2000), however not smaller than 0,75 and not
greater than 1. fH is the nominal design stress of the shell.

• Stresses for a loose flange:

Note: in a loose flange σr = σH = 0

16/05/2018 Sant'Ambrogio Servizi Industriali s.r.l. 12

 Gasket factors m and y for different gasket types (source: EN 13445.3 clause 11)

16/05/2018 Sant'Ambrogio Servizi Industriali s.r.l. 13

 Nominal design stresses for bolting in the Taylor Forge method

In the Taylor Forge method extremely low nominal design bolt stresses compensate the
very low gasket seating stresses. In reality the actual bolt loads are much higher than
the loads predicted by the method. ASME VIII division 1 recognizes that the actual bolt
stresses obtained in a normal tightening procedure is about
1564
f bo 
d bo (for a bolt M20 made of SA 193 B7 it gives 349 MPa against a
nominal design stress of 172 MPa). However there is no
obligation to apply bolt loads using a controlled torque wrench. Pay attention in case torque
or load values must be prescribed!
16/05/2018 Sant'Ambrogio Servizi Industriali s.r.l. 14
 The flange design method of EN 1591.1:2001 and EN 13445.3 Annex G
The harmonised Pressure Vessel standard EN 13445.3, in addition to the standard design
method of Clause 11 (based on Taylor Forge) gives also in its Annex G an alternative
flange design method, based on EN 1591:2001. The intention of this alternative method
is to overcome all the simplified assumptions made in the preceding Taylor Forge and DIN
methods using a more detailed model of a bolted assembly. The main features of the
method are the following.
1. A stress analysis of a more complex model is performed. The model is composed
by a gasket, a series of bolts, and by two (possibly different) flanges, or by
flange and a cover.
2. A more precise study on the elastic-plastic behaviour of the gasket is performed.
This behaviour is now described by means of 6 different gasket parameters instead
of the 2 parameters considered in the preceding methods.
3. Through these gasket parameters it is possible to study the actual deformations of
the entire assembly, thus establishing a correlation among the bolt load created
at bolting-up and the bolt loads in all the subsequent pressure situations (test
and service).
4. Analysis of the tightening device used, so that its characteristic plus or minus
tolerances (scatter values) may be considered in gasket and bolt calculations.
5. Additional loads are also considered.
6. Different temperatures for the various components (flange 1, flange 2, bolts and
gasket) may also be considered.
7. All the components are checked using the rules of limit analysis.
16/05/2018 Sant'Ambrogio Servizi Industriali s.r.l. 15
 Gasket parameters in ENV 1591.2:2001 and EN 13445.3 Annex G

• Qomin (MPa) is the minimum seating stress (equivalent to y in the Taylor Forge method)
to be put on the gasket in order to achieve the initial plasticization;

• Qmax (MPa) is the maximum allowable compressive stress on the gasket (as a function
of the operating temperature);

• Eo (MPa) is the gasket elastic modulus (as a function of the operating temperature) at
zero compressive load (see further explanation);

• K1 (plain number, function of the operating temperature) is the coefficient of variation of

the gasket elastic modulus; to understand the meaning of this coefficient and of the
preceding parameter Eo, we can say that each gasket which has been compressed up to a
specified load Qm, when compression is released behaves as its elastic modulus were given
by the formula EG= Eo + K1Qm; in other words, the more is the compressive load on the
gasket, the more is its rigidity; this phenomenon doesn’t take place on metallic gaskets
where it is always K1 = 0;

• m (plain number, function of the operating temperature) is the same as in the Taylor Forge
method (however it is not possible to make a direct comparison, because in the Taylor
Forge method m is referred to a conventional fixed contact surface, while in Annex G
the effective gasket contact width is calculated each time iteratively);

• gc (plain number, function of the operating temperature) is the “creep factor”, that is the
correction factor to be applied to the gasket elastic modulus EG in order to take into account
the creep that takes place in some gasket materials even at room temperature; use of this
16/05/2018
factor Sant'Ambrogio
avoids further complications given byServizi Industriali
a hypothetic s.r.l.dependent stress situation. 16
time
 Example of tabulated gasket parameters to be used in the alternative flange design
method of EN 13445.3 Annex G (source: ENV 1591.2:2001 and Table G.9.6 of EN 13445.3)

16/05/2018 Sant'Ambrogio Servizi Industriali s.r.l. 17

 The new approach in EN 1591.1:2013 considering tightness classes
(= guaranteed leakage rates) for the case of gases and vapours

• Qmin(L) (MPa) is the minimum required seating stress (equivalent to y in the

Taylor Forge method) to be put on the gasket in order to achieve leak tightness for a
specific Tightness Class L;
• QSmin(L) (MPa) is the minimum required gasket surface pressure for tightness
class L in service (this characteristic depends on the gasket surface pressure QA
applied during assembly – this is the equivalent of the m x p value of the Taylor
Forge method);
• Qsmax(To) (MPa) is the maximum allowable compressive stress on the gasket
(at room temperature);
• Qsmax(T) (MPa) is the maximum allowable compressive stress on the gasket (at
service temperature);
• ∆eGC (mm) is the creep relaxation of the gasket
• EG (Mpa) is the modulus of elasticity of the gasket for unloading/reloading
• μG is the friction factor between the gasket and the flange facing
16/05/2018 Sant'Ambrogio Servizi Industriali s.r.l. 18
Sealing gasket parameter when no leakage rate is specified (EN 1591.1:2013)

16/05/2018 Sant'Ambrogio Servizi Industriali s.r.l. 19

Example of tabulated gasket parameters with specified leakage rate (EN 1591.2:2008)
(these tables are referred to commercial gasket types used in the PERL project)

16/05/2018 Sant'Ambrogio Servizi Industriali s.r.l. 20

, for 80 bar underlined, for 160 bar double underlined
Leakage rate during unloading of a gasket tested at 10 bar Helium pressure

16/05/2018 Sant'Ambrogio Servizi Industriali s.r.l. 21

 Scatter of initial bolt load on a single bolt (Table B.1 of EN 1591.1:2013)

16/05/2018 Sant'Ambrogio Servizi Industriali s.r.l. 22

 Positive and negative scatter values for nB
bolts starting from the scatter values of a
single bolt (EN 1591.1:2013 Annex B.2)

 3   3 
1   1  
 n   n 
 n   1    n   1  
B B

4 4

W
Nominal load to be specified
WA 
for bolt tightening
(1   n  )
Total load to be used for bolt
checking
WB  WA (1   n  )

(W = theoretical bolt load for proper gasket compression)

16/05/2018 Sant'Ambrogio Servizi Industriali s.r.l. 23
 Advantages and disadvantages of the alternative flange design
method of EN 13445.3 Annex G (same as EN 1591.1:2001)

• More reliable calculation of the bolt loads, which are generally higher than in any
one of the other methods

• The bolt loads are strictly dependent on the tightening device used: if an
alternative device having different scatter values has to be used, a new checking of the
assembly is needed

• Tightening load of any flange connection has always to be specified and the user is
not allowed to use higher loads

• The nominal design bolt stress is the same of the other product types: therefore
the bolts have no additional safety margin against excessive tightening (note also that
the safety factor for bolting-up is 1,1 against the yield point as in the DIN method)

• It is not possible to design the bolts independently from the design of the
other parts of the connection (gasket and flanges): any modification made on gasket
and flanges will have an influence on the bolt loads

• The 6 gasket parameters recommended in the tables should be confirmed by the

gasket manufacturers

16/05/2018 Sant'Ambrogio Servizi Industriali s.r.l. 24

 Cross comparison of bolt load values for gasket seating
and service conditions as a function of gasket width in
different calculation standards (Gasket: Mineral fibre
soft iron jacketed ID=1000 mm - PS=25 bar)

16/05/2018 Sant'Ambrogio Servizi Industriali s.r.l. 25

 Main problems in extending the new design method of EN 1591.1:2013 to the Unfired
Pressure Vessel standard EN 13445.3

• EN 1591.1:2013 contains a table of gasket characteristics to be used with liquids,

however in case of gases or vapours for which a specific tightness class is specified,
gasket parameters shall be either supplied by the gasket manufacturer or resulting
from tests in accordance with EN 13555:2014.
• All the data contained in EN 1591.2:2008 are referred to tests performed on the gasket
used in the PERL project only. Similar gasket types of different manufacturers may have
different characteristics. In EN 1591.2:2008 there is a reference to the data contained in
European gasket data bases (e.g. http://www.gasketdata.org)
•Testing procedure and gasket data are always referred to the standard testing procedure
of EN 13555, that is testing with Helium normally at 40 bar with a standard test gasket
having specified dimensions.
• The results in terms of leak rates are therefore valid only for Helium: if other fluids are to
be sealed, it is necessary to convert the leak rates for Helium into leak rates for the
specified fluid (see Annex I of EN 1591.1:2013)
• Even if the use of EN 1591.1:2013 is certainly the best possible method to be used for the
sealing of gaseous fluids on the basis of specified leakage rates, the lack of generally valid
gasket characteristics in the standard itself is a big problem for the designers.
• For this purpose a general revision of EN 1591.2:2008 is needed.
• As for the case of the previous edition of EN 1591.1:2001 and for the case of Annex G of
EN 13445.3, the results are only valid with specified values of the torque and of the relevant
tightening devices to be used for bolt tightening.
16/05/2018 Sant'Ambrogio Servizi Industriali s.r.l. 26
View publication stats

THANK YOU VERY MUCH FOR

YOUR ATTENTION!

16/05/2018 Sant'Ambrogio Servizi Industriali s.r.l. 27