Professional Documents
Culture Documents
Date
Feb-03
Summary
The use of FEA in the prediction of pipeline girth weld residual stresses and a comparison with
experimental measurements is described. The effects of hydrotesting on the weld residual
stresses are also considered. A summary of some possible future workshop activities for the
Durability and Life Extension technology areas of FENet are presented.
Advantica
A premier provider of advanced technology and systems solutions that help high
performance energy and water delivery companies world-wide improve their
operating performance.
Proven track record of over 30 years experience servicing more than 550 clients in
over 50 countries.
1.2
E/B < 50 J/mm2
1
0.8
0.6
0.4
0.2
0
-0.2
-0.4
-0.6
0.2
0.4
0.6
0.8
(z/B)
Weld elements at
tensile yield (475MPa) in
hoop direction.
NONE !!
Conduct detailed experimental and finite element analysis to determine the residual
stress fields in the vicinity of gas transmission pipeline girth welds.
Determine the effects of hydrotesting on pipeline girth weld residual stress fields.
500
400
300
200
100
Parent M01-04
Parent M01-05
Parent M01-06
0
0
0.005
0.01
0.015
True Strain
0.02
0.025
Welding Details
AWS E6061 electrode for root and second pass, and E8010 for other 4 passes.
Pass
No.
1
2
3
4
5
6 (cap)
Travel
Speed
(cm/min)
26
35
24
12
13
13
Interpass
Temp.
(C)
50
60
80
50
70
90
Electrode
Size
(mm)
4
4
5
5
5
5
Electrode
Type
AWS
E6010
E6010
E8010
E8010
E8010
E8010
Heat
Input
(kJ/mm)
0.84
0.76
1.26
2.16
1.86
1.59
o+5
30 0
+1
1.5 0
Typical Welds
6 meters
Air abrasive
centre hole
drilling
method.
Weld Heat
Flow Model
Thermal
History
Mechanical
Model
Microstructure &
Property Model
Weldment Microstructure
& Mechanical Properties
Experiment
Validation
Applicable for girth weld with the exception of the weld start/stop positions
Half model for perfectly aligned pipes, Full model for misaligned pipes
Four noded linear isoparametric quadrilateral element
Very fine mesh (element length about 0.1mm) in the weld and HAZ region for
microstructure analysis
700
Elastic Modulus
500
Specific Heat
Conductivity
Stress (MPa)
Poissons Ratio
600
HAZ
400
Base
Metal
300
200
100
0
200
400
600
800
1000
Temp (K)
FENET THEMATIC NETWORK
COMPETITIVE AND SUSTAINABLE GROWTH
(GROWTH) PROGRAMME
1200
1400
1600
Ferrite Fraction
Pearlite Fraction
Hardness (VHZ)
200
150
X
100
50
O
0
0
FENET THEMATIC NETWORK
COMPETITIVE AND SUSTAINABLE GROWTH
(GROWTH) PROGRAMME
10
X (mm)
15
20
Axial Stresses
Max value 392MPa
Hoop Stresses
Axial Stresses
Max value 196MPa
Hoop Stresses
As Welded Results
0.8
Max Axial
Stress Point
Outer Surface
Location 2
Location 1
Weld Centerline
ace
0.6
0.4
0.2
0
-0.2
-0.4
-0.6
-0.8
-1
0
0.1
0.2
0.3
0.4
0.5
z/B
FENET THEMATIC NETWORK
COMPETITIVE AND SUSTAINABLE GROWTH
(GROWTH) PROGRAMME
0.6
0.7
0.8
0.9
1
0.8
0.6
0.4
0.2
0
-0.2
-0.4
-0.6
-0.8
-1
0
0.1
0.2
0.3
0.4
0.5
z/B
FENET THEMATIC NETWORK
COMPETITIVE AND SUSTAINABLE GROWTH
(GROWTH) PROGRAMME
0.6
0.7
0.8
0.9
Some Observations
The objectives of the experimental programme of work were achieved in that it did
demonstrate that hydrotesting does significantly reduce girth weld residual stresses.
The through wall axial residual stress distribution recommended by BS7910 was
shown to be conservative at and near to the inner surface of the weld. At the outer
surface of the weld this is not the case, for both as welded and after hydrotesting.
The supporting analytical work using a sequentially coupled Thermal-MechanicalMetallurgical Weld Stress Model was able to predict reasonably well the as-welded
microstructure and the mechanical properties (Hardness) observed in actual welds.
Hence the predicted residual streses from this Model would appear to have greater
credibility than the simplified modelling approach described at the outset.
Observations Continued
The scatter in the as welded measured axial stresses at the inner surface ranged
from high tensile (244MPa) to compressive (-136MPa). The high tensile value was
measured at the weld start/stop positions. The axisymmetric FE simulation is unable
to predict stresses at the weld start/stop positions. If the weld start/stop locations
are not considered then the comparison of the measured and FE stresses is much
better.
Therefore, an upper bound to the experimentally measured axial residual stress was
used instead in the Fitness-For-Purpose assessments,
What Next ?
It would be nice to have: Simulation model extended to 3D and attempt to reproduce weld start/stops.
Confidence limits associated with the through wall residual stress profiles.
Hence a stress distribution to an appropriate confidence level could be used in
defect assessments. Possibly by using Monte Carlo Simulation techniques with
the Thermal-Mechanical-Metallurgical Weld Stress model.