Professional Documents
Culture Documents
Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct
1. Introduction flood light towers (see Fig. 1) are among the most representative.
Generally, these towers are made up of several stretches or
This paper presents a methodology for the structural integrity modules (cylinder or cone trunk shaped) which are individually
assessment of tubular towers containing lack of penetration carried to the final location (i.e., wind farm, stadium . . . ) and then
defects on the butt welds of their circumferential sections. joined.
In most cases, when the towers have large dimensions, these Starting from the steel sheet reception in the production centre,
structures are made up of several stretches (manufactured in plant) the whole manufacturing and construction process of the towers
which are, simultaneously, composed of different rings. The joints consists (in many cases) of the following steps:
between the rings are generally made using butt welds, while the
joints between the stretches composing the tower are performed (1) Shaping: the sheets (usually from 20 to 40 mm thick) are
through butt welds or, alternatively, through a system of flanges inserted in a machine that shapes the rings using a system
and bolts. In any case, the welding process requires a strict and of rollers. In this case, only one sheet per section was used,
exhaustive quality control in order to avoid any kind of defects leading to just one longitudinal joint per ring.
threatening the structural integrity of the tower, particularly in (2) Welding: The longitudinal joints are welded (submerged arc
those welds performed in situ when the stretches are joined welding, SAW) through double V type butt welds, obtaining
through welding processes. the corresponding rings. Then, the rings are joined performing
Unfortunately there are situations where the defects are circumferential welds (also using SAW techniques and double
not avoided. In such cases, the structural consequences of the V type butt welds). As a result, primary cylinder/cone trunk
defects are not straightforward, making it necessary to perform shaped stretches of different lengths are obtained. The number
a structural integrity assessment of the towers considering the of rings joined depends on the length of the stretches
presence of such defects. (generally varying from 10 to 30 m in the case of wind towers).
There are some characteristic examples of structural tubular Fig. 2 shows a scheme of a double V type butt weld, before
towers. Perhaps the towers of wind turbines (wind towers) and and after the welding process. No defects were found on the
longitudinal welds, so the analysis performed here is only
referred to the circumferential ones.
(3) Surface treatments (i.e., shoot peening, painting . . . ), drying and
∗ Corresponding address: Departamento de Ciencia e Ingeniería de Materiales,
assembly of the auxiliary equipment (flanges, ladders . . . ).
Universidad de Cantabria, Av/Los Castros s/n, ETS Ingenieros de Caminos, 39005,
Santander, Spain. Fax: +34 942201818. (4) Transport from the centre of production to the final location
E-mail address: ciceros@unican.es (S. Cicero). (i.e., wind farm).
0141-0296/$ – see front matter © 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.engstruct.2009.03.013
2124 S. Cicero et al. / Engineering Structures 31 (2009) 2123–2131
Table 1
Dimensions of the analysed circumferential sections.
Section Diameter (mm) Thickness (mm)
S1 4000 36
S2 3865 35
S3 2630 21
3. Case study
3.1. Geometry of the tower and material properties (Steps (a) and (b))
Once the allowable crack sizes are determined, two main The loads acting on the towers have an evident random
possibilities may occur: component which mainly arises from the random nature of wind
- The allowable crack sizes are larger than those detected in the loads. Wind effects were determined following [16] and then
structure. In such cases, there is no need to take any remedial the corresponding loading conditions for the tower, designed
action. according to [17], can be inferred. The resulting acting loads
- There are some allowable crack sizes that are smaller than (caused by wind and the rotation of the rotor blades) were obtained
those detected in the structure. In such cases, remedial actions as a Markov matrix (supplied by the owner), distinguishing
(i.e., removal of the damaged welded zone and subsequent bending moment range, mean bending moment and number of
rewelding) are required. The number and the size of the defects cycles for each type of cycle. The compression arising from their
own weight is not considered. This constitutes a conservative
determine the feasibility of the repair, both in economical and
assumption, given that it provides higher tensile stresses than
technological terms.
the actual ones. As an example, Table 3 gathers part of the loads
This methodology (summarised in the flowchart shown in in Section S1. It should be noted that these loads are defined as
Fig. 3) is applied below to a case study in order to illustrate more operational stability loads, which represent operation of the wind
clearly the different steps explained above. farm over a service life of 20 years.
2126 S. Cicero et al. / Engineering Structures 31 (2009) 2123–2131
Table 2
Mechanical properties of the different materials.
Section Material [15] Yield stress (σy , MPa) Ultimate tensile stress (σu , MPa) Charpy T27J (◦ C) ∆Kth (MPa m1/2 ) C (da/dN in m/cycle) m
Table 3
Markov matrix gathering the load spectrum in Section S1.
Range (N m) Mean (N m) Cycles
Table 4
Maximum tensile stresses in the different sections analysed.
Section Maximum tensile stress (MPa)
S1 175
S2 177
S3 161
Fig. 6. Scheme of lack of penetration defects found on the butt welds of the towers.
3.3. Weld inspection and definition of defects (Steps (d) and (e))
Failure C
Assesment
1.0 Line B
A = Acceptable Condition
B = Limiting Condition
C = Unacceptable Condition
Kr
A
0 1.0
Lr=F/Fy Lrmax
Fig. 8. Scheme showing the FAD methodology [1]. F : applied load; Fy : plastic
collapse load; Lmax
r : maximum permitted value of Lr ; Kr : fracture ratio of applied
Fig. 7. Geometry of cracks analysed [2]. 2c = 5 mm, π R/2, π R y 2π R. elastic K value to Kmat (material toughness).
The definition of the geometry of the detected defects was not is defined from the proof stress (here Rp ), the ultimate tensile
straightforward and did not present any general characteristics strength (Rm ) and Young’s modulus (E):
in terms of length or depth. Also, interactions between adjacent −1/2
f (Lr ) = 1 + 0.5L2r S Lr ≤ 1
defects were not ruled out. For these reasons, the defects were
idealised as embedded and circumferential (see Fig. 7) with f (Lr ) = f (1)Lr(N −1)/2N 1 ≤ Lr ≤ Lmax
r
(1)
2c values of 5 mm, quarter, half and whole circumference, 2a f (Lr ) = 0 Lr ≥ Lmax
r
Table 5 1.2
Estimation of fracture toughness from Charpy values.
Section Material [15] Charpy T 27J (◦ C) Thickness, B (mm) Kmat (MPa m1/2 ) [1] 1
S1 S355J2 −20 36 39.3
S2 S235J0 0 35 34.3
0.8
S3 S235JR 20 21 32.3
2c=5 mm
2c = 1/4 circumference
Kr
1.2 0.6 2c = 1/2 circumference
2c = whole circumference
0.4
1
2c=5 mm
0.2
0.8 2c = 1/4 circumference
2c = 1/2 circumference
2c = whole circumference
0
Kr
0.2
Table 6
Critical crack sizes (2ac ) in Sections S1, S2 and S3 for the four crack extension (2c)
0 hypotheses (mm).
0 0.2 0.4 0.6 0.8 1 1.2 1.4
Section 2c = 2c = 2c = 2c =
Lr 5 mm 1/4 circ. 1/2 circ. whole circ.
S1
Fig. 9. FAD analysis in section S1. 17.5 17.5 17.5 13.2
Thickness: 36 mm
S2
7.4 7.4 7.4 7.4
1.2 Thickness: 35 mm
S3
5.9 5.9 5.9 5.9
Thickness: 21 mm
1
0.8 Table 7
Crack sizes (2ath ), below which there is no fatigue propagation (mm).
2c=5 mm Section 2c = 2c = 2c = 2c =
Kr
Table 8 Table 9
Final crack sizes (2af ) at the end of the lifespan of the towers (mm). Bold characters Maximum tolerable lack of penetration defects (2amax ) in the wind towers (mm).
correspond to those situations where 2ath causes the structural failure before the Bold characters correspond to situations conditioned by the fatigue threshold; those
lifespan finishes. in italics are limited by crack propagation; the underlined value is conditioned by
fracture–plastic collapse.
Section 2c = 2c = 2c = 2c =
5 mm 1/4 circ. 1/2 circ. whole circ. Section 2c = 2c = 2c = 2c =
5 mm 1/4 circ. 1/2 circ. whole circ.
S1
5.9 13.3 15.5 >2ac
Thickness: 36 mm S1
4.9 3.1 3.1 2.9
S2 Thickness: 36 mm
5.4 >2ac >2ac >2ac
Thickness: 35 mm S2
4.8 2.9 2.9 2.8
S3 Thickness: 35 mm
>2ac >2ac >2ac >2ac
Thickness: 21 mm S3
5.9 4.2 4.1 3.8
Thickness: 21 mm
S3: 2c = 5mm
Design loads
S2: 2c = 5mm
S1: 2c = 5mm
2a max(mm)
Fig. 12. Maximum tolerable lack of penetration defects (2amax ) in the wind towers for the different hypotheses of applied loads (mm).
Table 11 [4] Schwalbe K-H, Koçak M, editors. Mis-matching of welds. In: Proceedings
Maximum tolerable lack of penetration defects (2amax ) in the wind towers (mm). of the international conference Mis-Match’93, ESIS 17. London: Mechanical
Loads 10% reduced. Engineering Publications; 1994.
[5] Schwalbe K-H, Koçak M, editors. Mis-matching of interfaces and welds.
Section 2c = 2c = 2c = 2c =
In: Proceedings of the 2nd international conference Mis-Match’96. Germany:
5 mm 1/4 circ. 1/2 circ. whole circ.
GKSS Research Centre Publications; 1997.
S1 [6] Seib E, Koçak M. Fracture analysis of strength undermatched welds of thin-
6.1 3.8 3.8 3.5 walled aluminium structures using FITNET procedure. IIW Doc. X-1577-2005.
Thickness: 36 mm
S2 [7] Cicero S, Yeni Ç, Koçak M. Fracture analysis of strength undermatched Al-Alloy
6.0 3.7 3.7 3.4 welds in edge cracked tensile panels using FITNET procedure. Fatigue Fract Eng
Thickness: 35 mm
Mater Struct 2008;31(9):738–53.
S3
7.4 5.1 5.0 4.6 [8] Anderson TL. Fracture mechanics: Fundamentals and applications. 2nd edition
Thickness: 21 mm
Boca Raton: CRC Press; 1995.
[9] Cicero S. Evaluación de la integridad estructural de componentes sometidos a
condiciones de bajo confinamiento. Ph.D. thesis. Santander (Spain): University
tolerances have been determined for the different defect geome- of Cantabria; 2007.
tries considered. These values can be assumed to be conservative [10] Cicero S, Gutiérrez-Solana F, Álvarez JA. Structural integrity assessment of
and ensure the lifespan considered in the initial design. components subjected to low constraint conditions. Eng Fract Mech 2008;75:
3038–59.
Moreover, a sensitivity analysis has been performed, as the final [11] SINTAP: Structural integrity assessment procedures for European industry.
step in the structural integrity assessment, in order to determine Brite-Euram Project no. BE95-1426, Contract no. BRPR-CT95-0024. Final
how variations in the different inputs can affect the final results. report; 1999.
[12] Schwalbe K-H, Kim Y-J, Hao S, Cornec A, Koçak M. EFAM ETM-MM 96: The
All this analysis, together with precise measurements of ETM method for assessing the significance of crack-like defects in joints with
the existing defects, will allow the owner-designer to take the mechanical heterogeneity (strength mismatch). GKSS Report 97/E/9. GKSS
corresponding decisions (basically, to repair or not to repair). Forschungszentrum; 1997.
[13] Paris PC, Gomez MP, Anderson WP. A rational analytic theory of fatigue. Trend
Given the accumulated conservatism caused by the different Eng 1961;13:9–14.
assumptions considered in the analysis, and also considering that [14] Forman RG, Mettu SR. Behaviour of surface and corner cracks subjected
the design loads are defined based on the site conditions (obtained to tensile and bending loads in a Ti-6Al-4V alloy. Ernst HA, Saxena A and
McDowell DL (editors). Philadelphia: Fracture mechanics 22th symposium 1
in situ using instrumentation devices and including extreme
American society for testing and materials. ASTP STP1131; 1992. pp. 519–646.
events), the final recommendation to the owner, and for the case [15] EN 10025: Hot rolled products of structural steels. 2004.
study being analysed, is to compare the detected defects with those [16] IEC 61400-SER, Wind turbine generator systems - ALL PARTS. 2008.
equivalent values (in terms of crack extension hypothesis, (2c)) [17] Eurocode 3: Design of steel structures. 1997.
[18] Creager M, Paris C. Elastic field equations for blunt cracks with reference to
gathered in Table 9 (obtained prior to the sensitivity analysis). The stress corrosion cracking. International Journal of Fracture 1967;3:247–52.
rest of the values gathered in Tables 10 and 11 (including those [19] Niu LS, Chehimi C, Pluvinage G. Stress field near a large blunted V notch and
obtained in the sensitivity analysis) can be taken as a reference, application of the concept of notch stress intensity factor to the fracture of very
brittle materials. Eng Fract Mech 1994;49:325–35.
also showing the coherence between the loads considered and the [20] Pluvinage G. Fatigue and fracture emanating from notch; the use of the notch
allowable crack sizes obtained. If the detected defects are larger stress intensity factor. Nucl Eng Des 1998;185:173–84.
than those gathered in Table 9, repairs are required if the structural [21] Fenghui W. Prediction of intrinsic fracture toughness for brittle materials
from the apparent toughness of notched-crack specimen. J Mater Sci 2000;
integrity of the towers has to be ensured. 35:2543–6.
[22] Kim JH, Kim DH, Moon SI. Evaluation of static and dynamic fracture toughness
References using apparent fracture toughness of notched specimens. Mater Sci Eng A
2004;387–389:381–4.
[1] Kocak M, Webster S, Janosch JJ, Ainsworth RA, Koers R, editors. FITNET fitness- [23] Taylor D, Cornetti P, Pugno N. The fracture mechanics of finite crack extension.
for-service (FFS) procedure- Volume 1. Geesthacht, Germany: GKSS Research Eng Fract Mech 2005;72:1021–38.
Centre; 2008. [24] Dugdale DS. Yielding in steel sheets containing slits. J Mech Phys Solids 1960;
[2] Kocak M, Hadley I, Szavai S, Tkach Y, Taylor N, editors. FITNET fitness-for- 8:100–4.
service (FFS) annex - Volume 2. Geesthacht, Germany: GKSS Research Centre; [25] Burdekin FM, Stone DEW. The crack opening displacement approach to
2008. fracture mechanics in yielding materials. J Strain Anal 1 1966;144–53.
[3] Kocak M, Laukkanen A, Gutiérrez-Solana F, Cicero S, Hadley I, editors. FITNET [26] Harrison RP, Loosemore K, Milne I. R6: Assessment of the integrity of structures
fitness-for-service (FFS) case studies and tutorials – Volume 3. Geesthacht, containing defects, CEGB report R/H/R6; 1976.
Germany: GKSS Research Centre; 2009. [27] Wallin K. The scatter in KIC results. Eng Fract Mech 1984;19:1085–93.