You are on page 1of 43

Dimensioning of equipment

Economical utilization of material


1. Original method – ideally elastic material:

Equipment or structures dimensioning is on base of stress that is


determined providing that material is ideally elastic (linear section of Hookes
law).
- Advantage: Simple calculations
- Disadvantage: It is problematic to specify such load, when an equipment
or structure collapses or its deformations are too big.

2. Method taking into account elastic – plastic deformations


of material:

What is a behavior of a structure (equipment) subjects to the elastic-plastic


deformation?
External loading causes firstly elastic deformation. When is the load
higher in some structure places (with the highest stress) plastic deformation

starts to form (σ loc = σ K). The yet higher load causes that the number of these
places and their extent escalates. Final state is the structure collapse (total
collapse or too big deformations). The state of stress in the structure just before
the collapse has name „ limit state (extreme state, stress limit)“.

For structures, equipments and parts appreciation from the point of view of
their operational reliability and safety following types of limit states are used:

1. Elastic limit state – when is the load higher plastic deformations start to form
somewhere in a structure

58852921.doc 1 / 43 Print date: 22.5.2011


P. Hoffman
2. Strength limit state – when it is reached a material consistency is broken
(local or total); a fracture rises (brittle failure, plastic failure, ductile, mixed,
fatigue fracture or creep fracture)
3. Deformation limit state – when it is reached structure starts to have non-
permissible deformations
4. Load limit state – when it is reached a structure with firm shape loses its
connections and changes into vague structure or mechanism (structure
connections are lost by a failure of some structure part or parts – types of
failures see above)
5. Limit state of adaptation – till the state elastic deformations do not exceed
some value even if the overloading is repeated
6. Limit state of stability – whet it is exceeds non-permissible deformations
and/or failures rise very quickly

The theory of limit states is applicable for tough materials with marked yield

point σ K .

3. Basic models of tough materials:

1. Ideally plastic σσ
K

Theoretical material that does not


exist in praxis (after loading it has plast. deformation)
Fig.1 ε ε
perm
2. Ideally elastic-plastic ε
σ Plast
σ
An assume is, that till reaching the yield point
K > °C
material behaves ideally elastic (σ = E * ε ).
When the yield point is reached the stress

is constant (σ ) but deformation rises. ε


K ε
ε
K per
Fig.2
m
58852921.doc 2 / 43 Print date: 22.5.2011
P. Hoffman
The model is often used for engineering calculations

3. With linear hardening

σ
σ α
When yield point is exceeded the material
σ =σ + E´*(ε –
K K
is harder (it withstands higher stress).
ε K
)
Dependence of the hardening on extension
is done by a line that slope α ´ is defined by ε
ε
the modulus of hardening E´. K
Fig. 3

4. With exponential hardening


σ
σ
When yield point is exceeded the material σ =σ + a*(ε –
K K
is harder according exponential ε K
)b
dependence.
ε
ε
K
Fig. 4

As E´ << E difference of results of calculations according models 3. or 4.


compared to model 2. are negligible

In engineering praxis ideally elastic-plastic model is used.

For pressure vessels steels with low carbon content are used. Such steels
have very low hardening. For such steels is the model valid. The simplification
goes on the side of higher safety (calculated maximal load is a bit lower than real
with the hardening).
Models 3. a 4. are used for more exacting calculations with using of
numerical methods.

58852921.doc 3 / 43 Print date: 22.5.2011


P. Hoffman
On the material plasticity have effect temperature and time too (time of
external loading, its uniformity or variation).
In praxis exists in structures and/or apparatuses internal stress (effect of
welding, thermal treatment, forming, cutting, mounting etc.). It can worsen
strength conditions of such structure (sometimes for example pre-stress improve
conditions in structure).

4. Economical utilization of material plasticity

For all examples in the chapter we suppose an ideally elastic-plastic model (it is
without hardenning).

4.0 Plasticity for uni-axial tensile load of bar

ε
σ 4Plast

cross section S σK 1 2 4

S
σ Fmax
=
σ K
0
F 3 5 ε
F σ F
=F/S ε 5perm

Steps of the bar loading, overloading and unloading:


0–1–0 F≤σ K *S only elastic deformations (after unloading ε p = 0)
1–2 F>σ K *S plastic deformation ε p = 1-2 = 0-3
2–3 unloading after unloading is permanent deformation ε p = 0-
3
3–2–3 F≤σ K *S only elastic deformations (after unloading ε p = 0-
3)
3–2–4 F>σ K *S plastic deformation ε p = 2-4

58852921.doc 4 / 43 Print date: 22.5.2011


P. Hoffman
4–5 unloading after unloading is permanent deformation ε p = 0-
5
etc.
In the case all material “strings” are plasticized and all lengthen in one time.
When is the overloading and unloading repeated a fatigue failure comes →

It is impossible to overload such bar (or structure) repeatedly!!

4.1. Example of loading of rectangular beam with section b x h with


bending force (fig. 5a,b):

FF compressive stress

MBmax = F * L/4

L/2 L/2 tensile stress

Fig. 5a: Example of beam loading

By gradual increasing of the beam loading in its profile rises tension. The
tension increases till it reaches in outer strings yield point σ K. Capacity of the
beam according the theory of limit states is not fully utilized. For higher loading
parts of the beam profile starts to plasticize gradually from outer parts towards
to the center. When is the all profile plasticized in the axis arise so-called plastic
joint. Originally triangular stress profile changes into a rectangular profile –see
fig.5b.

Fp
Fe
h
Fe 2/3h Fp h/2

b σ σ σ plastic
K K K joint

58852921.doc 5 / 43 Print date: 22.5.2011


P. Hoffman
Beam Elastic Part. plastic Fully plasticized
profile deformation profile deform. beam profile

Fig.5b: Curves of stress in beam


Maximal bending moment in elastic region

Me = Fe * 2/3h =1/2 * h/2 * σ K * 2/3h = 1/6 * h2 * σ K

Maximal bending moment for fully plasticized profile

Mp = Fp * h/2 = h/2 * σ K * h/2 = 1/4 * h2 * σ K

Note:
Force acts in a gravity center of the area. For triangle is the gravity center in 2/3 of leg, for
rectangular in 1/2. Force size is done by triangle or rectangular areas (with sides σ K and
h/2).

Coefficient of plasticity Cp is ratio Mp / Me .

Cp = (1/4 * h2 * σ K) / (1/6 * h2 * σ K) = 6 / 4 = 1,5

From it follows that for a rectangular beam profile is the maximal plastic
bending moment 1.5 times higher than maximal elastic bending moment.
Therefore it is possible to increase the bending loading of the beam 1.5 times or
decrease the safety factor (for example from value x = 1,5 to x = 1).

When such fully plasticized profile is unloaded, in outer profile strings


stress arises (as strings are elongated). The stress that corresponds to opposite
bending moment with value Mpe2 (= pre-loading with this elastic moment)

Mpe2 = (1 - Ce2) * Me

58852921.doc 6 / 43 Print date: 22.5.2011


P. Hoffman
where Ce2 = 2/Cp – 1 ... coef. of residual deformation after plastic joint arise

For some profiles are values of Cp and Ce2 in following table.

For profiles with Cp = 2 after unloading in outer strings arise stress with value
equals to yield point with opposite sign (e.g. instead tensile is compression). For
profiles with Cp > 2 opposite plastic deformation arise after unloading in outer
strings.

(E.g. for Cp = 1,5; Ce2 = 0,33 → Me2 = 0,33*Me; Mpe2 = 0,67*Me;


for Cp = 2,2; Ce2 = -0,1 → Me2 = -0,1*Me; Mpe2=1,1*Me > Melast).

≤+σ +σ ≤+σ +σ
≤-σ K
σ K
=0
K K K K

≤-σ -σ ≤+σ -σ
K K ≤-σ K K
K
M ≤ Memax M = Mp M = -Mpe2 M = +Mpe2 M ≤ Memax +Mpe2 M = Mp+ Mpe2

Fig. 5c: Alternating stress of a beam (fatigue loading result is fatigue failure)

Tab.1: Values of elastic and plastic section modulus We and Wp, coefficient of
plasticity Cp and coefficient of residual deformation Ce2 for various profiles

Profile Woe Wop=Woe*Cp Cp Ce2


________________________________________________________________
Square b
1/6b3 1/4b3 1,5 0,33e
(side b) b

Square skew b3/(6*√2) √2*b3/12 2,0 0,0


(side b) b b

58852921.doc 7 / 43 Print date: 22.5.2011


P. Hoffman
Rectangular 1/6bh2 1/4bh2 1,5 0,33
h
(height h, side b)
b
Triangle bh2/12 2*√2/6*bh2 2,34 -0,145
(base b, height h) h

b
Profile Me Mp Cp Ce2
________________________________________________________________
Circle π d3/32 d3/6 1,7 0,176
(diameter d)

Annulus π d3/32*(1-k4) d3/6*(1-k3) (0,18+k3-1,18k4)/(1-k3)


(diameters de and di) 1,7*(1-k3)/(1-k4)

Annulus 0,78*de2*s de2*s 1,27 0,575


(diameter de, wall thickness s)

di s
de de

4.2. Plasticization for combine load (tensioning + bending)

- Limiting force for loading with only force FPO

FPO = h * b * σ K (area on what acts tension times max. tension


σ K)
σ
FPO K FPO Remember chapter 4.1.

- Limiting moment for loading with only moment MPO

MPO = Wp * σ K = Wo * Cp * σ K = 1/4 * b * h2 * σ K

58852921.doc 8 / 43 Print date: 22.5.2011


P. Hoffman
(MPO=1/6 * b * h2 * 1,5 * σ K)

-σ K
See above + table 1

+σ W is section modulus for bending


K

The stress profile in a beam for the combine loading under limiting state is on
the following fig. 6. (superposition of both loadings has to be on the yield point =
limiting state – maximal total load is such that all profile is plasticized).

Fig.6
c Fp1
rFp1
h
M
plastic
F rFp2 joint
Fp2
σ -
b σ
K K

In a distance c from surface a plastic joint arises. We specify limiting forces and
their moments calculated to the profile axis and from them total limiting values:

FP1 = c * b * σ K FP2 = (h – c) * b * σ K

FP = FP2 - FP1 = (h – c) * b * σ K -c*b*σ K = (h – 2c) * b *


σ K

Moment arms of forces FP1,2


rFP1 = h/2 – c/2 = (h – c) / 2 rFP2 = c/2

Limiting bend moment for fully plasticized profile is

58852921.doc 9 / 43 Print date: 22.5.2011


P. Hoffman
MP = FP1 * rFP1 + FP2 * rFP2 =
MP = c * b * σ K * (h – c) / 2 + (h – c) * b * σ K * c/2

MP = c * (h – c) * b * σ K

Ratio of limiting values for separate and combined loading is after modification:

FP / FPO = 1 – 2 c/h MP / MPO = 4 c/h * (1 – c/h)

If we eliminate the ratio c/h from these equations we obtain this equation:

MP / MPO + (FP / FPO)2 = 1

For our example it is valid:

for elastic state M=σ M*W F=σ F*S


for plastic state MP = 1,5 * σ K * W FP = σ K *
S

and after substitution these values into the upper equation and its modification
we obtain following equation:
2
σM σ 
+ F  =1
1,5 *σ K  σ K 

If we introduce a total stress σ Σ =σ M +σ F and dividing the equation with


σ K we obtain this very important equation

58852921.doc 10 / 43 Print date: 22.5.2011


P. Hoffman
2
σΣ σ σ 
= 1,5 + F − 1,5 *  F 
σK σK σ K 

Dependence of σ Σ / σ K on σ F / σ K we can set in the following diagram


(fig.7).

Dependence of total stress σ Σ /σ K on tension stress


σ F/σ K
Reserve on x = 1,0
Reserve on x = 1,5
1,67

Economical utilization of plasticity

x = 1,0 x = 1,5

Region of elastic loading


1/3 2/3

Only bend Fig.7 Tension stress σ F/σ K Only tension


(compression)

From this dependence follows that for only bending loading is

2,0
total limiting stress equal 1,5 x yield point. For combined loading
(tensioning + bending) with prevailing bending stress the total limiting
stress (acceptable) rises up and the maximum is for the ratio σ F /σ K

= 1/3, when is the total limiting stress σ Σ = 1,67 * σ K. Than it falls


down and for only pure tensioning it reaches value σ Σ = σ K.

This finding we can use for choice of the safety factor .

58852921.doc 11 / 43 Print date: 22.5.2011


P. Hoffman
A structure in what are only tensioning or compression stresses collapses
(according the elastic limit state) when in some part (parts) the yield point is
reached it is when σ F = σ K. For membrane stress (= only tension or
compression) the safety factor x = 1,5 (σ DF =σ K /x= σ K / 1,5) is used. If in a
structure is only pure bending we can use safety factor x = 1,0 (we have reserve
50 % to limiting state - σ DM =σ K / x = σ K). For the combination of tensioning
and bending is this reserve higher compared with only bending. Therefore is for
the case till the ratio σ F /σ K = 0,67 possible to use the safety factor x = 1. For
higher ratio values we can use the safety factor value again x = 1,5.

5. Plasticizing of statically indeterminate structures with only


uni-axial tensioning

As example we have a structure from 3 beams that is in the fig. 8.

F2
0
30
2 h 3 F1´ F3´
1 F1 F3

F
Fig.8 F

The structure has following parameters:

Beams lengths: l2 = h l1 = l3 = 2*h


Beams sections: A2 = A A1 = A3 = 4*A

Note:
These dimensions are selected for better illustration – you will see it thereinafter.

58852921.doc 12 / 43 Print date: 22.5.2011


P. Hoffman
The structure is loaded with force F that is in range Fe < 0; 2*A*σ K> . The
middle beam 2 will behave during the loading and unloading elastic (its
deformation will be only elastic). Balance of power in the structure is:

F1 = F3 F1´= F3´ = F1 * sin α = F1 / 2


F = F1´+ F2 + F3´ = F1 / 2 + F2 + F1 / 2 = F1 + F2

a) For loading with force F = 2*A*σ K in the central beam 2 will be stress
(providing that F2 = F1 – it is specified from requirement that all beams have to
have the same extension in the F direction and from the beams stiffness):

F2 = F / 2 = 2*A*σ K / 2 = A*σ K → σ 2 = F2 / A = A * σ K /A=σ K

From it follows that till the force is reached the beam 2 has only elastic
deformations.

b) Now we suppose that the structure is loaded with higher force that will be in

the range FI = FpI < 2*A*σ K; 4*A*σ K> . Under these conditions the beam 2
starts to deform plastically, but beams 1 and 3 have still the elastic deformation.
After unloading beams want to go to their original position but the beam 2 is
elongated. Therefore in the beam 2 arise after the unloading compression pre-
stress and in beams 1 and 3 tension pre-stress. For a new loading with force
increasing from 0 to FpI the structure will behave like elastic system in all this
range (see fig. 9 and 10). For the new loading in the central beam 2 will be
following stresses: - compression stress (pre-stress) → no stress → tension stress

c) For loading of the structure with even higher force F that will be in range

(FII < 4*A*σ K; 5*A*σ K> ) plastic deformations in beams 1 and 3 rise too.
After unloading of the structure in the beam 2 arises pressure plastic

58852921.doc 13 / 43 Print date: 22.5.2011


P. Hoffman
deformation (alternating plasticizing tension + compression → fatigue failure of
the beam 2).

d) For loading of the structure with the force higher than FIII > 5*A*σ K are
plastic deformations in all beams and the structure comes to a region of
uncontrollable creeping. For hardening materials it is possible to load such
structure but in praxis (as I told before) is not the hardening calculated (higher
safety).

F III. Uncontrolable
5Aσ
creep
Κ II. Plasticizing of Start of plasticizing
4Aσ beams 1, 2 and 3 of beams 1 and 3
Κ
3Aσ I. Plasticizing of beam 2,
Κ
elastic deformation of 1 and 3
2Aσ
Κ
1Aσ Elastic deformation
Κ of beams 1, 2 and 3
0 σ / 3σ
0 Ε
Κ

Κ
4σ / ε Start of plasticizing of beam 2
2σ Κ Κ

/Ε Ε
Beam 2 elastic deformation for loading a)

Fig. 9: Loading of 3 beams structure according fig.8. State of the middle beam 2.

Now we will consider an ideally elastic-plastic material. It means that the


beam 2, when it reaches the yield point, starts to deform plastically (on line σ K

= const.). The course of this alternating loading and unloading is shown in so-
called Planck diagram (fig.10).
2 ε k
ε 2
σ fict
ε k ε p12 Max. adaptation of the structure to
the overloading with force F
58852921.doc 14 / 43 Print date: 22.5.2011
P. Hoffman
1 2 5 7
σ Kt

Alternating plasticizing of
beam 2
σ fictmax
0 3
σ res12 ε
4
− σ Kt Fig. 10: Loading diagram of the
6 8 8´ middle beam of the 3 beams
ε p8 ε p8´

In region of elastic deformation the state of beam 2 moves on the line 0 – 1.


When the yield point is reached the beam starts plastically deform on the line
1 – 2. After unloading (line 2 – 4) a compression pre-stress arises (residual
stress) in the beam.

σ res12 =-E*ε p12

If the structure is loaded again so that ε =ε 2 =ε k +ε p12 ≤ 2/E*σ K the


repeated loading will go in the elastic state on the line 4 – 2. In so doing it does
not depend on this how many times was the point 2 reached (independent on a
loading history).
In the case the structure is adapted to this overloading.

For a new reach of the yield point in the beam 2 (point 2) a following stress is
necessary (line 4 – 2):

σ =σ K + |σ res12 |=σ fictive

When is the structure loading so high that plastic deformation reaches the

value ε p = ε k (total deformation is ε = ε 5 = 2∗ ε k – point 5),

reaches the value of the residual stress after unloading σ res12 =-σ Kt (point 6).

58852921.doc 15 / 43 Print date: 22.5.2011


P. Hoffman
For next loading rising and resultant next rising plastic deformation (point 7)
the beam 2 starts to plastic deform after unloading in the compression region too
(point 8). For the next overloading and unloading plastic deformations in tensile
and compressive regions arise (points 7´and 8´).

Next alternating overloading and unloading (on lines 6–5-7–8-7) faces to the
alternating plasticizing of the beam profile with result of fatigue failure (with
always increasing permanent deformation).
The structure is not able to adapt itself to such overloading.

. σ fict.limit =2*σ K

The structure adaptation against an overloading can be only in the region


of deformations < ε k; 2ε k >.

About possibilities of the adaptation a dimensionless quantity called the


coefficient of adaptation kp decides (shake down, Einspieltheorem), that is done
by the ratio of loading in the second cycle F2, when the structure starts to
plasticize in tension and compression regions to loading F1 when the structure
reaches for the first the yield point in tension:

F 56
k = 2
= ≤2
F 01
p

Note:
These relations are valid only for uniaxial stress. E.g. for the triaxial stress are
these relations much more complicated, but the principle and results are similar.

Loading of statically indeterminate structures with sudden shape changes


or other discontinuities causes a rise of local secondary and peak tensions
(stresses). These tensions are in the wall section distributed unevenly. Tension
peaks have only a local character so that from it following plastic deformations
do not expand into the peak surroundings. From it follows that even a high local

58852921.doc 16 / 43 Print date: 22.5.2011


P. Hoffman
overrun of the yield point has not effect on the total strength of the structure
(system, pressure vessel etc.).

That is why a static loading peak tensions are not taken into account!

For steel shells loaded mainly with membrane stress following reach of
these transitional stresses L is specified:
4
R2 * s2
L= ≈ 0,78* R * s ≈ 0,55* D* s
3* (1− µ 2 )

For technical praxis a value with safety factor 3 is assumed

L K =1,65* D* s

where R = D/2 is a radius of curvature in a calculated place (D is a diameter)


and s is a wall thickness of the membrane (shell). μ is Poisson constant and for
steel
μ ≈ 0,3.

Example.: For R = 1000 mm and s = 10 mm is L = 0,78*√(1000*10) = 78 mm

6. Stress categories in structure (membrane wall)

1. Primary stress = tensile and compressive stresses (membrane s.) in wall


that are distributed uniformly or bending stresses that are caused by external
forces. After arising of a plastic deformation in some strings of a profile
(when in the place the yield point is exceeded) they do not decrease too much.
That is why these primary stresses tensile and compressive are limited with
the safety x = 1,5 (exceeding of the loading owing to working conditions or
incorrect dimensioning can cause a rise of a big plastic deformation with
following failure). Bending stresses and combined stresses (e.g. bending +

58852921.doc 17 / 43 Print date: 22.5.2011


P. Hoffman
tensile) have safety factor x = 1. In the case the value of an allowable stress
must not exceed the yield point (theoretically).

2. Secondary stress = statically indeterminate stresses that after exceeding the


yield point do not causes bigger plastic deformations and after a some
plasticization are able to adapt on an local overloading (effect of the statically
indetermination). In the group belong for example stresses caused by local
external forces, membrane shape change, temperature change etc. A value of
these secondary stresses themselves or with combination (sum) with primary
stresses is limited by the requirement that their maximal value in any
direction and any string has to be lower than is double yield point (see fig.
10).

+σ K

ε
-σ K

Example of the secondary stress in a shell (elliptical tube) with internal


overpressure p
A
Owing to the pressure the elliptical
tube shape wants to change into
p
the circular ones → secondary
B bending stress in wall
(primary stress = tension)

Place A tends to the lower radius Place B tends to the bigger radius
→ outer strings tensile, inner press → outer strings press, inner tensile

+ = + =
58852921.doc 18 / 43 Print date: 22.5.2011
primary
P. Hoffman secondary total primary secondary total
stress stress stress stress stress stress
For a proper design we can expect in places with the highest stress partial
plasticization and residual stresses after unloading (= adaptation on the
local overloading). But if these residual stresses (pre-stresses) reach the
yield point there is a danger of an alternating plasticization that causes to
the fatigue failure (with big contractions) → low-cycle fatigue.
The adaptation of a structure or pressure vessel to an overloading is used
for pressure test (elimination of possible stress peaks). A distance on what
these secondary stresses act depends on a vessel diameter and its thick-
ness.

3. Stress peaks (tertiary stresses) = stresses with local character that are only in
some profile strings (local external forces, sharp deviations of form, welds,
notches, material changes etc. Local plastic deformations practically have not
effect on a structure (shell ...). They are taken into account only for low-cycle
fatigue (e.g. number of cycles of a structure loading and unloading).

7. Conditions of plasticity for stress in more axis

In the case we have to specify an equivalent stress σ e, that will compare


with the yield point σ K. The equivalent stress is given by 6 components of a
stress tensor. Conditions of plasticity determined by various theories differ in a
way of a specification of the equivalent stress. As the plastic deformation takes
place usually by shear there are used various forms of shear stresses for the
determination of the equivalent stress.

58852921.doc 19 / 43 Print date: 22.5.2011


P. Hoffman
In engineering praxis following theories for biaxial or triaxial stress are
used (in next text we suppose that σ 1 > σ 2 > σ 3). The equivalent stress

determined according these theories is compared with the yield point (σ e <
σ K).

A reason of this way is that tensile tests for yield point and tensile strength
specification are made on steel rods with uniaxial stress but in praxis in
structures etc. is bi- or triaxial stress. That is why we have to found an
equivalent value that can be compared with the yield point.

1. Hypothesis of maximal normal stress σ max (tensile or press)

Lameé 2 axial stress σ e=σ 1

3 axial stress σ e=σ 1 (tension) or σ e = σ 3 (press)


(depending on σ ktension and σ kpress )

2. Hypothesis of maximal shear stress

σ max − σ min
Guest 2 axial stress τ max = ≤τ K
2

(remember Mohr´s circles = stress circles – see p. 35); where σ max = σ 1; σ min =
0

According this hypothesis a structure failure comes when a maximal shear


stress τ max reaches value equal to a limiting shear stress for uni-axial tensile. In
normal stresses is the condition expressed by the following relation σ e =σ 1 -
σ 3 ≤ σ K. In the case of bi-axial stress is σ 3 = 0. Than is a region of safe
loading given by following conditions:

58852921.doc 20 / 43 Print date: 22.5.2011


P. Hoffman
σ −σ ≤ σ
1 2 K
σ ≤σ1 K
σ ≤σ
2 K

For satisfying of these conditions of a structure (shell ..) a maximal value of


a resulting stress must lie inside so-called Tresc´s hexagon (fig.11).

Now I show you some examples of various loadings and them equivalent stresses
in the Tresc´s hexagon.

a) Uni-axial tension in direction of axis σ 1/σ K; σ 1 ; σ 2 =0


in direction of axis σ 2/σ K; σ 2 ; σ 1 =0

b) Uni-axial press in direction of axis -σ 1/σ K; -σ 1 ; σ 2 =0


in direction of axis -σ 2/σ K; -σ 2 ; σ 1 =0

c) Simple shear σ 1 = σ 2; σ 1 = -σ 2

d) Thin-walled spherical shell with inner overpressure pi


Tangential (circumferential) stress = axial stress → stress in all directions
is the same
σ o =σ 1 =σ t=σ 2 = pi * r / 2s

Similarly it is for outer overpressure (but for the case we have to take into account not
only the stress but conditions of stability too – see later)


e) Thin-walled cylindrical shell with inner

 overpressure pi
 
σ σ 
 
Tangential stress = 2 * axial stress (see p. 34)
 spherical shell (pi)
σ

σ o =σ 2 = pi * r / 2s; σ t = σ 1 = pi * r / s = 2*σ 2


  cylindrical

 shell (pi)
     
 
σ σ
      
σ  


  

c σ  
  
    
 
 


spherical shell (pe) 21 / 43  
58852921.doc
P. Hoffman

  Print date: 22.5.2011

 
 σ
Obr.11. Conditions of plasticity for bi-axial stress

3. Hypothesis HMH (energetic hypothesis Huber – Mises – Hencky)

According this hypothesis the biggest effect on a structure failure has a


specific energy of stress needed for a shape change (elongation, compression,
contraction ...). For the case of a plane stress (bi-axial) it is valid that

σ 2
=σ 2
+σ 2
-σ *σ σ = σ + σ − σ *σ
2 2

e 1 2 1 2 or e 1 2 1 2

If we divide both sides of the equation by σ K and bring results in the fig. 11
we obtain an ellipse that goes through apexes of the Tresc´s hexagon.
Comparing these two hypothesis we can see that the Guest hypothesis lies on the
side of higher safety. The HMH hypothesis is more economical (it better utilize a
material strength – for simple shear and cylindrical shells; not for simple tension
/ press and spherical shell).

In the case of tri-axial stress is a range of a safe stress a “hexagonal prism


with pyramids” for the Guest´s hypothesis, in the case of HMH hypothesis it is a
spheroid (circular ellipsoid).

Example:

58852921.doc 22 / 43 Print date: 22.5.2011


P. Hoffman
Let us suppose a thin-walled cylindrical shell with inside radius r = 500 mm and
wall thickness s = 20 mm, that is loaded with an inner overpressure pi. Yield
point of the shell material is σ K = 230 MPa. Safety factor x = 1.5. Our task is to
specify maximal allowed overpressure.

It is valid that (see p. 34)

tangential stress σ t = p*r/s; axial stress σ o = p*r/2s;


radial stress σ r = -p and σ t> σ o >σ r

According the condition σ max (Lameé) is

σ t = p * r / s ≤ σ K/x →
pD = σ K * s / (x * r) = 230*20/(1,5*500) = 6,13 MPa

According the energetic condition HMH is

1
σ = * (σ − σ ) + (σ − σ ) + (σ − σ )
2 2 2

2
e t o o r r t

For easier calculations we introduce symbols


a = σ o / σ t = 0,5;
b = σ r / σ t = p / (p*r/s) = -230*20/(230*500) = -0,04

and after substitution and modification we obtain

σ
σ = t
* (1 − a ) + (a − b) + (−b − 1) = ... = 0,901* σ
2 2 2

2
e t

58852921.doc 23 / 43 Print date: 22.5.2011


P. Hoffman
p*r σ
σ = 0,901* ≤ K
x = 1,5
s x
e

Than the maximal inner overpressure calculated from the hypothesis (for
plasticity conditions) is
σ *s 230 * 20
p = K
= = 10,21 .MPa
0,901 * r 0,901 * 500
p

and allowed overpressure is pD = pp / x = 10,21 / 1,5 = 6,81 MPa


(x according Guest 6,13 MPa)

8. Basic methods of equipment dimensioning


(vessels, parts etc.)

Presumptions for the structures dimensioning:

• Description of processes that are in an equipment


• Mass and energy inlets and outlets = balances
• Control system (regulation, sensors, actuating devices, way of control,
parameters fluctuation etc.)
• Location in a line (effects of surroundings, footings, lifting necks or eyes,
connecting flanges and pipes etc.).

Fig.12.: Chart of a way of an equipment dimensioning


 

     
     


 
  

      
   

        ≤ 
          
      
     
calculation 
   
58852921.doc
P. Hoffman
24 / 43
       
Print date: 22.5.2011

feed back  


A) Specification of external loading

During this specification a designer has to solve following problems:

• Possibility of separation of a combined loading in simple loadings and


subsequent utilization of laws superposition of stresses and shifts (elongations
etc.).
• Character of external forces (surface f. = inner and outer pressure, local
forces ...; volumetric f. = mass of equipment, internal tensions owing to
welding or fabrication, thermal tensions owing to dilatation ...)
• What is an effect of the loading = monotonous or changing (periodical or
casual); way of operation.
• Conditions of a loading character and transients states (way of operation and
its limiting conditions, way of starting and breakdown (after every shift or in
longer intervals), equipment maintaining etc.).
• Further operational conditions or restrictions (operation with lower or higher
output, allowed variation of operating temperatures and pressures, heating or
cooling speeds and number of their changes, corrosive and abrasive effects of
working media, fouling on working surfaces during operation, ways of
cleaning and sanitation etc).

B) Internal stress

For a specification of the internal stress it is necessary to solve following


problems:

58852921.doc 25 / 43 Print date: 22.5.2011


P. Hoffman
• It is possible to replace a component with a simple geometric form (sphere,
cylinder, plate, beam, strut-frame, frame etc.) or their combination?
• It is possible to analyze the component or equipment as single elements for
what we know solution and calculation of internal forces (stresses) and
moments and we are able to determine boundary conditions for their
connection?
• It is possible or we have to use a given method of calculation according
standards, rules etc.?
• Unless, it is possible to use simplifying presumptions and so the problem
approximately solve? What is an error of such approximate solution?
• Is the structure etc. statically determinate (internal forces can be calculated
from balance with external forces) or indeterminate (for specification of
internal forces we need deformation conditions in addition)?
• It is known what an accuracy of our calculation is? Are results on a side of
higher or lower safety?
• Will calculations perform from designed operational parameters or from fault
(extreme) conditions?
• Do internal stresses rise during a part (equipment etc.) manufacturing
(forming, machining, jointing ...)? What are their values and what is their
effect on internal stress (together with external forces)?
• What are boundary conditions it is a way of placing or join with other parts
(supporting, hanging, way of fixation etc.)? It is very important parameter
that has a big effect on results accuracy.
• Are available data for solution of problems of fatigue, creep, determination of
safety factor or calculations according stochastic theory of operational
reliability?
• It is possible to use (for more complicated cases) a finite elements method or
some computer program?

C) Defining of equivalent stress

58852921.doc 26 / 43 Print date: 22.5.2011


P. Hoffman
Calculated tri-axial stress is compared with material strength that is a
result from a tensile test of uni-axial stress. Results must be comparable and
reproducable. From tensile tests we obtain a start of shear for these
characteristic parameters (yield point etc.) according following hypotheses:

• Main axial stress reaches yield point (σ K; σ 0,2 )

Lamée ..... σ omax ≤ σ K

• Maximal shear stress reaches a limit of elastic shear of material providing


that

Guest ....... τ max = 0,5 * σ K

• Axial tensile elongation ε omax reaches a maximal limiting value ε K on the


yield point, it is that

St. Venant ....... ε omax =σ K /E


• Total energy needed for elongation that can a volume unit absorb reaches its
maximal value on the yield point

Haigh-Belrami ........ (E∆ )


e max = 1/2 * σ K
2
/E
• Total energy needed for a shape change (elongation, compression,
contraction ...) that can a volume unit absorb reaches its maximal value on the
yield point

HMH ..................... (E)max = (1+µ )/2 * σ K


2
/ E ≈ 0,65 * σ K
2
/
E (Huber-Mises-Hencky) Poisson´s constant for steel is µ = 0,3

• Octohedral shear stress reaches its maximal value that is expressed in form

τ max = √2/3 * σ K = 0,47 * σ K

58852921.doc 27 / 43 Print date: 22.5.2011


P. Hoffman
Every of these 6 quantities forms a base for special hypothesis of
specifying of an equivalent stress for uni-, bi- or tri-axial stresses. Its proper
choice has a big signification from the point of view of safety in operation and
economical utilization of material.

D) Allowed stresses

These stresses are given as a verified material characteristic given by


tensile tests (e.g. yield point) divided by the safety factor x.

Safety factors are different for various engineering components, structures


and equipments and are given by standards and regulations for their design,
fabrication and operation. For common engineering calculations tabulated
mechanical properties of materials are used together with given constant
standard safety factors.

For more exacting calculations we can use for example an effect of a


change of material properties with time etc. (theory of reliability taking into
account cumulative growth of damage, mathematical models of time change of
material properties etc.). Pressure vessels must be made from atested materials.

In some cases is not a part dimensioned from a point of view stress but
according technological needs (e.g. casts wall or a weldment thickness, tubeplate
thickness etc.).

Why calculated values differ from designed (equipment over- or undersizing)?

• Variability of strength characteristics of materials compared to tabulated


average values resulting from tests → for more exact calculations (e.g. for
pressure vessels) we must have certificates for every used material.

58852921.doc 28 / 43 Print date: 22.5.2011


P. Hoffman
• Changes of mechanical properties in a part profile (for profiles with thicker
walls is the difference higher than for thin ones → see material tables).
• Change of material strength depends on a character of loading (e.g. speed of
loading, material hardening or fatigue).
• Difference between calculated and actual working conditions of a structure,
equipment etc.
• Effect of other parts of a structure that are connected with the part etc.
(rigidity or elasticity of a system and its parts, their relationships,
relationship among forces etc.).
• Effects of stress raisers (concentrators) in fixations, joints, shape changes
(footings, supports, necks, beams, holders etc.).
• Effect of additional forces caused by production technology or assembly
(welding, pre-straining from assembly insufficiently annealed cast =
shrinkage stress etc.) - typical example is additional force (stress) on vessel
necks from a bit shorter piping when are flanges drawn together with higher
force of screws.
• Effect of overloading caused during operation by lack of technological
discipline, problems with control, troubles etc.
• Effect of internal stress caused by micro- and macro-roughness of surface.
• Suitability or unsuitability of a hypothesis used for a case.

Example of utilisation of various hypothesis used for dimensioning of thick-


walled cylindical vessel loaded with internal overpressure

Given:
ri – internal radius; re – external radius; pi - overpressure
k = re / ri – dimensionless wall thickness

Task usually is to specify following 3 internal pressures:

58852921.doc 29 / 43 Print date: 22.5.2011


P. Hoffman
• Internal pressure (pi)i=K, at what starts to plasticize only some internal string
but in the rest of the profile are only elastic deformations. A condition for
this state is that the equivalent stress in the internal string just reaches the

yield point. If we transform pressure to a dimensionless quantity (pi / σ )


K

we can obtain, for individual hypothesis, dependence of this ratio (pi / σ )


K

on the dimensionless wall thickness k (see fig.13).


• Internal pressure (pi)e=K at what is material just plasticized in all profile
(thickness). The condition is used for creep.
• Internal pressure (pi)i=P at what comes to a material rupture etc. In some
cases a material hardening is taken into account.

Derivation of dependence (pi / σ K) – k for membrane theory of vessels

Tangential (circumferential) stress is in a cylindrical vessel 2 x higher than


an axial (see p. 34 – for sphere they are the same). Therefore we will take into
account this stress. According a membrane theory in a membrane wall (thin-
walled shell) are only tensile stresses (for internal pressure). We will take into
account the 1st case it is reaching of the yield point only in internal string.

σ t = pi * ri / s ≤ σ K s = re - ri k = re / r i re = k * ri

After substitution and modification we obtain a membrane formula:

σ t = pi * ri / (re – ri) = pi * ri / (k*ri – ri) = pi / (k – 1) ≤ σ K

(pi / σ K) = k – 1

Similar formulas can be derived for other hypotheses.

58852921.doc 30 / 43 Print date: 22.5.2011


P. Hoffman
Lamée (pi / σ K) = (k2 – 1) / (k2 + 1) (ASME for k > 1,5)

Guest (pi / σ K) = (k2 – 1) / 2k2

St. Venant (pi / σ K ) = (k2 – 1) / (1,3k2 + 0,4)

H–B (pi / σ K) = 2(k2 – 1) / √(6 + 10k4)

H-M-H (pi / σ K) = (k2 – 1) / (√3*k2)

ASME ( pi / σ K ) = (k – 1) / (0,6k + 0,4) for k ≤ 1,5

For fully plasticized state (profile):

Guest (pi / σ K) = ln(k)

H-M-H (pi / σ K) = 2/√3*ln(k)

Results calculated from these dependences are in the following table and
on fig.13. Boundary between thin-walled and thick-walled cylinders is
at value of k = 1,17. In practice more safety value k = 1,1 is used.

Theory for pi/σ K k = 1,1 k = 1,17 k = 1,4


Membrane 0,100 0,170 0,400
Guest 0,087 0,157 0,245
Lameé 0,095 0,156 0,324
St. Venant 0,106 0,169 0,326
H-B 0,092 0,148 0,288
H-M-H 0,100 0,156 0,283
ASME 0,094 0,154 0,323

In the region (k < 1,17 – thin-walled cylinders) is difference between


results according Guest and actual values c. 19 % ( for k ≈ 1,17) on safety side,

58852921.doc 31 / 43 Print date: 22.5.2011


P. Hoffman
it means that a vessel is oversized about 19 %. Results according H-M-H are
slightly oversized too. Membrane theory gives slightly oversized values.
For thick-walled cylinders gives membrane theory undersized results (it
allows higher pressure pi), Guest gives oversized results (it allows lower pressure
pi).

Dependence pi/σk on k
1,200

1,000

0,800 thick-walled Membrane


Lamée
pi/σ k

Guest
0,600
St.Venant
H -B
0,400
H-M-H

0,200

0,000
1,00 1,10 1,20 1,30 1,40 1,50 1,60 1,70 1,80 1,90 2,00

k = De /Di

Fig. 13 a, b: Dependence of maximal dimensionless pressure on dimensionless


thickness of cylindrical vessel.
Dependence pi/σk on k
0,200
0,180
0,160
0,140

 Membrane

0,120  Lamée
pi/σ k

Guest
0,100
St.Venant
0,080 H -B
0,060 H-M-H
0,040
0,020
0,000


1,000 1,025 1,050 1,075 1,100 1,125 1,150 1,175 
1,200

k = De /Di
58852921.doc 32 / 43 Print date: 22.5.2011
P. Hoffman
In the region for k > 1,17 (thick-walled cylinders) results according hypothesis
H-B and H-M-H are in the best conformity with actual measured values.
Membrane formula allows much higher pressure and such cylinder would be
very undersized!! Results according Lamée and St. Venant undersized a
cylinder too, but not so much like the membrane theory. On the contrary results
according Guest are somewhat oversized. (see fig. 13).

9. Example:

Given:
Cylindrical vessel loaded with internal overpressure pi = 0,6 MPa, external
diameter De = 1800 mm, material is steel with yield point σ K = 230 MPa.

Task:
What is a needed wall thickness of the vessel s = ?.
Because of simplification we will study a long cylinder without effects of
transient stress near covers and footings.

A)Specification of external loading

External loading (= internal overpressure) acts upright to the inner cylinder


wall and causes internal stress in the cylinder wall in tangential and axial
directions. No other forces act to the vessel.

B) Calculation of internal stress

58852921.doc 33 / 43 Print date: 22.5.2011


P. Hoffman
It is a typical statically determined structure with the membrane stress. Only
balance of powers acting on the vessel is sufficient for calculation of internal
stress. We can calculate only primary stresses as secondary stress and stress
peaks are not in the vessel.

Balance of powers in axial direction


(fictitious section upright on the cylinder axis)

σ
φ
a external force Fea ≈ π *D2/4 * pi
D (for Di = D - s ≈ D)
pi
internal force Fia = π *D*s*σ a

s σ a

Balance of forces in the section Fea = Fia → π *D2/4 * pi ≈


π *D*s*σ a

Axial stress in the cylinder wall σ a = pi * D / 4s = pi * r / 2s

Balance of powers in tangential direction


(fictitious section in the cylinder axis)

σ σ
t t Cylinder length is L
φ
D
external force Fet ≈ pi*D*L
pi s (for Di = D - s ≈ D)

internal force Fit = 2*L*s*σ t

58852921.doc 34 / 43 Print date: 22.5.2011


P. Hoffman
Balance of forces in the section Fet = Fit → pi*D*L ≈ 2*L*s*σ t

Tangential stress in the cylinder wall σ t = pi * D / 2s = pi * r / s


From it follows that the σ t = 2*σ a →
tubes splits longitudinally.
C) Definition of equivalent stress according Lamée and
Guest hypothesis

τ For bi-axial stress is according:

τ Lamée σ e=σ t≤ σ K
max
τ
3
ma . and according Guest
x2
σ σ τ max2 = (σ t – 0) / 2 ≤ τ K =
0 σ
σ K / 2-p p a t

Fig.14: Mohr´s circles for our example


(long cylinder with internal For both cases it is valid that
overpressure)
σ e=σ t

Condition of dimensioning is σ e≤ σ D

For tri-axial stress (we take into account compression stress in cylinder wall too)
following formulas are valid for definition of equivalent stress:

according Lamée σ e=σ t≤ σ K σ e=σ t≤ σ D

according Guest τ max3 = (σ t – (-p)) / 2 ≤ τ K = σ K /2 σ e =σ t

+p≤ σ D

58852921.doc 35 / 43 Print date: 22.5.2011


P. Hoffman
D)Allowed stress

For the bi-axial stress are in the wall only tensile stresses (in the case of tri-
axial stress is the compression stress negligible – see next page) and there are no
bending moments. Therefore we must choose the safety factor x = 1,5 (fig.7).
Allowed stress is thus

σ D =σ K / 1,5 = 230 / 1,5 = 153,3 MPa

Needed calculated wall thickness (without effects of a weld weakening v,


allowance for corrosion c etc.) will be for various hypotheses:

- Biaxial stress according Lamée and Guest and tri-axial stress according Lamée
(D = De –s ≈ De)

σ e = σ t = pi * r / s = pi * D / 2s ≤ σ D

sv1it ≥ pi * D / (2 * σ D) = 0,6*1800 / 2*153,3 = 3,52 mm (1.iter. D = De)

sv2it ≥ (0,6*(1800-3,52)) / 2*153,3 = 3,52 mm (2.iter. D = De –s)

Note:
Calculation with an average diameter D or external diameter De has not effect
on the resulting value of calculated wall thickness.

- Triaxial stress according Guest (D = De –s ≈ De)

σ e = σ t + pi = pi * r / s + pi = pi * D / 2s + pi ≤ σ D

sv1it ≥ pi * D / (2 * (σ D - pi)) = 0,6*1800 / 2*(153,3-0,6) = 3,53 mm (1.iter.)


sv2it ≥ (0,6*(1800-3,53)) / 2*(153,3-0,6) = 3,52 mm (2.iter.)

58852921.doc 36 / 43 Print date: 22.5.2011


P. Hoffman
- According ČSN 690010 (Czech standards)

sv ≥ pi * De / (2 * σ D*v - pi) + c

For c = 0 and v = 1 (see presumption above) is

sv ≥ pi * De / (2 * σ D - pi) = 0,6*1800 / (2*153,3-0,6) = 3,53 mm

These individual results are from the point of view of technical praxis practically
the same → for such thin-walled cylinder results do not depend on used theory.
Specification of single stresses in the cylindrical vessel wall
(with the calculated thickness s = 3,53 mm)

Tangential stress
σ t = pi * D / (2*s) = 0,6*(1800-3,53)/(2*3,53) = 152,7 MPa

Axial stress
σ o = σ t / 2 = 152,7 / 2 = 76,3 MPa

Radial stress
σ r = - pi = -0,6 MPa

From this comparison follows that Mohr´s circles (fig.14) for τ max are for the
case of biaxial stress as well as for triaxial stress practically the same (pi <<< σ o

and σ t). The circle between points 0 and –pi is practically reduced to a point.

If we utilize relations derived before for these theories for equivalent stress
specification (the diagram on fig.13 is based on them) we can specify the

58852921.doc 37 / 43 Print date: 22.5.2011


P. Hoffman
minimal calculated wall thickness too. Instead σ K we use σ D. Results according
various hypothesis are on the following pages.

Calculations are performed for 3 pressures so that we can see an effect of thin.
or thick-walled cylinders.

Specification of cylinder wall thickness for given pressure according


various hypothesis (see sooner derived formulas pi /σ K = f(k) and fig. 13)

k = re / ri = De / Di = De / (De – 2s) → s = De * (k – 1) / 2s

Membrane
(pi / σ K) = k – 1 → k = pi / σ D +1
De σ D pi k s
(mm) (MPa) (MPa) (-) (mm)
1800 153,3 0,6 1,00391 3,51
1800 153,3 20,0 1,13046 103,9
1800 153,3 50,0 1,33 221,4 too undersized !!!

Lamée - σ max

(pi / σ K) = (k2 – 1) / (k2 + 1) → k = ((σ D + pi) / (σ D –pi))0,5


De σ D pi k s
(mm) (MPa) (MPa) (-) (mm)
1800 153,3 0,6 1,00392 3,52 thin-walled
1800 153,3 20,0 1,14021 110,7 ≈ boundary

58852921.doc 38 / 43 Print date: 22.5.2011


P. Hoffman
1800 153,3 50,0 1,40 258,5 !! thick-walled

Guest - τ max

(pi / σ K) = (k2 – 1) / 2k2 → k = (σ D / (σ D – 2pi))0,5


De σ D pi k s
(mm) (MPa) (MPa) (-) (mm)
1800 153,3 0,6 1,00394 3,53
1800 153,3 20,0 1,1632 126,3
1800 153,3 50,0 1,70 369 too oversized !!

St.Venant – ε max

(pi / σ K) = (k2 – 1) / (1,3k2 + 0,4) → k = ((σ D+0,4pi) / (σ D-1,3pi))0,5


De σ D pi k s
(mm) (MPa) (MPa) (-) (mm)
1800 153,3 0,6 1,00334 2,99 undersized
1800 153,3 20,0 1,1632 100,5
1800 153,3 50,0 1,70 258 too undersized !!

H-M-H - Emax
(pi / σ K) = (k2 – 1) / (√3*k2) → k = (σ D / (σ D – 30,5 * pi))0,5
De σ D pi k s
(mm) (MPa) (MPa) (-) (mm)
1800 153,3 0,6 1,00394 3,06 undersized
1800 153,3 20,0 1,1632 108,2
1800 153,3 50,0 1,70 306 OK

58852921.doc 39 / 43 Print date: 22.5.2011


P. Hoffman
Some definitions:

Membranes: There are only tensile or compression forces (stresses).

Most common type of membranes = rotationally symmetrical


membranes.
Thin-walled, with constant stress in all profile section.
Usually have continuous curvature, thickness and loading.

Shells: They carry bending moments, torsion, shear local forces etc.
too.
In wall profile section are not only tensile or compression
stresses but bending stress too (and the bending stress can
vary (see above – the elliptical tube).

Loading: Fluid pressure (inner, outer, hydrostatical).


Self-weight, weight of content.
Local forces and moments (supports, attachment, ...).
Wind, snow, seismicity.

58852921.doc 40 / 43 Print date: 22.5.2011


P. Hoffman
Examples: Heat exchangers, evaporators.
Vessels, tanks, silos.
Reactors, columns, absorbers.
Piping systems.

measuring

Laplace formula for membranes

N N
R
+
α

β
R
+ p =0 β

α
z Nα = σ α *
s
Nα [N/m] ...
N = σ β * between axis of
normal force in section for α =β const. (angle
rotation
and section plane that cross the axis of rotation); the force is related
to 1 m of membrane length (in the section) – it does not depend on
membrane thickness
Nβ [N/m] ... normal force in section for β = const. (angle between chosen
base
plane and section plane that passes through axis of rotation)
Rα [m] .... radius of curvature of membrane in section for α = const.
Rβ [m] .... radius of curvature of membrane in section for β = const.

Ex.1: Sphere, hemisphere – internal overpressure p

Nα Nβ 2* N
+ = =p
R R R
58852921.doc 41 / 43 Print date: 22.5.2011
P. Hoffman
Rα = Rβ = R; pz = - p

Nα = Nβ = N ..... for sphere (axis can be chosen at will, forces must be the
same)

p*R
N= (N/m)
2
or for membrane with wall thickness s it is

N p*R
σ =σ =
α β
= (Pa = N/m2)
s 2*s

Ex.2: Cylinder – inner overpressure p, wall thickness s

Rα = ∞ ; Rβ = R; pz = - p

N N N
+ =p=
α β α

R R β
R α

Nα p* R p* D
N = p*R
α or σα = σt = = =
s s 2* s

Nβ is not possible to specify from the Laplace formula → we can specify it for
example from the forces balance in a cross section upright to the cylinder axis
(see above in chapter 9. Example ad C)

Section with plane β Section with plane α


that passes through axis that is upright axis

58852921.doc 42 / 43 Print date: 22.5.2011


P. Hoffman
58852921.doc 43 / 43 Print date: 22.5.2011
P. Hoffman

You might also like