Per J. Haagensen

Fatigue design of welded aluminum rectangular hollow section joints

Abstract Fatigue design methods for welded aluminum joints are reviewed, including various approaches to fatigue life estimation currently adopted in design codes across a range of industrial applications. The applicability of these established methodologies to the fatigue design of automotive space frame structures is critically assessed. The hot spot stress method is identified as the most promising in terms of providing a coherent and comprehensive approach to design. Particular problems related to implementation are considered such as failure sites and determination of appropriate stress concentration factors from physical models, finite element calculations or parametric equations. Preliminary results from finite element stress analyses and fatigue tests are also presented for rectangular hollow sections welded in a T-joint configuration. Recommendations are made for a design methodology for welded rectangular hollow-section joints in aluminum space frames, including use of a single hot spot S–N curve.

Fatigue Testing and S-N Data for Fatigue Analysis of Piles

Design S-N curves in design codes are based on fatigue test data, where the stress cycle is under external tension load. It is observed that during pile driving most of the stress cycle is compressive and the design procedure used for fatigue analysis of piles might therefore be conservative. In order to investigate this further, it was proposed to perform laboratory fatigue testing of specimens that are representative for butt welds in piles under relevant loading conditions. In the present project 30 test specimens made from welded plates were fatigue tested at different loading conditions to assess effect of compressive stress cycles as compared with tensile stress cycles. In 2006, the Edda tripod in block 2/7 was taken ashore. This platform has been in service since 1976 and the piles are considered to be representative for the piles installed in the North Sea jacket structures during the 1970s. Therefore it was suggested to investigate the pile weld at the sea bed in detail to assess the stress due to fabrication and 30 years of in-service life and the residual fatigue life of the pile. Six test specimens made from the Edda pile were fatigue tested. The results from the assessment and the fatigue testing are presented in this paper.

Fatigue strength improvement methods

Abstract: Methods for improving the fatigue performance of welded structures are reviewed, and the suitability of individual methods for industrial applications are evaluated. Methods involving geometry modification and residual stress obtained by peening are dealt with in some detail. Methods involving weld shape modification are introduced during the initial welding process. The influence of various factors such as base material strength, residual stress, variable amplitude service loading and specimen size are examined. In this chapter recommendations are given for a unified approach for the derivation of S–N curves for use in design with regard to the influence of main influencing parameters such as base material strength, size, weld and joint geometry, and stress concentration factor of the joint.

Experimental verification of HFMI treatment of large structures

Design recommendations for high frequency mechanical impact (HFMI)-treated welds have been proposed based on available experimental fatigue data of axially-loaded high strength steel specimens which include longitudinal, cruciform and butt welds. Test specimens were of a size appropriate for laboratory study. However, in reality, structures in civil, offshore and ship industries generally include large-scale and more complicated components, such as bridges, cranes, platforms, excavators etc. This paper presents a further validation of the design proposals by considering fatigue data sets which are obtained from large-scale components. The extracted fatigue data from the available literature includes bridge, crane and beam like components. In total, 65 published test results of weld details with various yield strengths (250 ≤ fy ≤ 725 MPa) and stress ratios (-1 ≤ R ≤ 0.56) are presented. All the data are found to be in good agreement with the previously-shown design curves.

Relative Importance of Defects in Girth Welds of Linepipes

The Ormen Lange deepwater gas field is located at water depth down to 1100 meters. The irregular seabed gives severe challenges to pipeline design and verification program was launched to demonstrate adequate fatigue capacity. The research included: modern welding techniques (5G and 2G welding positions), mapping of actual welding defects, misalignment (high/low) and lack of penetration. The thick walled pipe (35mm) showed low or even compressive residual stresses at the inside. This will to some extent be “protective” to the root of the weld. The exceptions to this pattern were the repair welds and the two-sided welds. The small scale test results fell close to the full scale pipe tests when taking into account the geometrical weld distortions, loading mode, and the distribution of weld defects. The importance of the parameters influencing the fatigue capacity could be ranged as follows, most detrimental first: large crack-like defects (LOP, undercut > 1mm), hi/lo, and V-shape (radial shrinkage at the girth weld).Copyright

Fatigue Damage Repair and Life Extension of a Floating Production Unit: The VFB Platform Revisited

The Veslefrikk B platform was built in 1985 as a drilling exploration unit but was converted to a production platform in 1989. After only two years in service fatigue cracks were discovered and several repairs were made. However, extensive fatigue cracking continued and a retrofitting program was planned. In addition, increased payload was necessitated by more topside equipment required for a tie-in to the Huldra field which was scheduled to start production in 2001. In 1999 the platform was temporarily decommissioned and dry-docked for a comprehensive repair and upgrading program, this was completed in approximately two months. The life extension program was described in the OMAE 2000 conference paper 2954. However, after only one more year of service new cracks were found and subsequent fatigue damage necessitated new repairs. It is noteworthy that cracking this time occurred only in areas of the structure that were left untreated in the 1999 retrofitting program due to assumed low levels of stress in those areas. The paper describes the original repair and strengthening program, and the types of subsequent fatigue damage that required new repairs. Most of the cracks occurred in the hull skin plates and caused water leakage. The objective of the recent life extension program is to ensure safe operation of the platform for a period of another 20 years.Copyright

Fatigue of High Quality Girth Welds

Several studies on fatigue strength of high quality girth welds are summarized and discussed. The fatigue performance of such welds is consistently above the common design classes, as long as key set of influencing parameters are controlled. Fatigue life and crack initiation depend on loading mode, and weld defects, weld geometry, residual stresses, and degree of weld distortions. The welding method, especially for the root, has also proven to be important, e.g. TIG and high quality STT are often superior to Cu-backed roots. In practice the most important factor will be the surface breaking flaws as root LOP, etc. Such flaws are rarely found at the cap side. Thick pipe walls will also reduce the fatigue capacity, e.g. a 45mm wall thickness with 25mm reference, will reduce the with the same amount as above. High fatigue performance requires absence of any weld discontinuity above certain critical sizes which may be a challenge for the accuracy and resolution of NDT systems. In the present study, some important factors that influence the fatigue strength are examined. Based on results and theoretical calculations, the effects of the various crack-like discontinuities are described and compared to current design standards. The geometrical misalignment of the joint (hi/lo) will also influence the fatigue capacity. In bending, the weld cap toe and the weld root are the critical locations. The residual stress distribution in the welded region may, however, alter this. With pipe wall thickness larger than ∼25mm residual stresses can be beneficial to the weld root area. However, due to the scatter in the measurements it is difficult to assess the influence on fatigue life exactly. FE modeling of the welding process is therefore used to supplement this discussion. Improvement methods are available to suppress the critical influence of surface cracks/discontinuities. Grinding of the cap weld toe has in many cases shown significant improvements of girth welds. Also, TIG-dressing, high quality STT, or fillers with high nickel content, have potential for improving the root performance.Copyright

Scale Effects: Correlation of Fatigue Capacity for Full-Scale Pipes and Samll-Scale Specimens

Increasing a component’s size is known to have detrimental effect on the fatigue performance. To account for this, reduction factors are given in several design codes. Away from plate products, there is no general consistency between the various correction procedures. The main causes of the scale effects can be divided into three phenomenon: Statistical effects (probability for defects that initiate cracks), technological effects (differences in material and production procedures for large and small components), and stress gradient effects (scaling differences for stress level for (small) cracks). In this paper the main factors influencing the scale effect is reviewed and correction rules are discussed. Fatigue data for 30” and ID6” pipes with girth welds are analysed and compared with medium sized sector specimens cut out from pipes. Also effects of weld quality and pipe alignment (hi/lo and angular distortion) in this context are discussed. On this basis, recommendations for applying the scale correction procedures for welded pipes are proposed.Copyright

Fatigue Design Data Derived From Full Scale Tests on 6”OD Pipe Girth Welds

A literature survey of high quality girth welds intended for pipelines risers was carried out and the results are compared with full scale resonance fatigue test data on 6” pipes. The samples were made from 168.3×9.9mm (OD×WT) seamless pipes, each having two welds. Axial misalignments (hi-lo’s) and lack of penetration (LOP) defects were introduced in the test pipes to study the effects on the mean minus 2 stand, deviation design S-N curve that was calculated. Post failure examination of the welds was performed to determine the type and size of defects in the failure initiation area. Fracture mechanics calculations were carried out to determine the effect of defects on fatigue life. The test results were compared with published data on 6” pipes with high quality welds. The scatter in the fatigue test data was reduced when comparisons were based on the local stress at the point of fracture initiation. The implications for design rules of the findings in this work are discussed.Copyright

Development of Fatigue Design Data for the Ormen Lange Offshore Project: An Overview

The Ormen Lange offshore pipelines from shore to the field go through very difficult terrain creating freespans in the range 40–80m for the 30” lines. For the 6” lines long freespans will be present prior to burial and vortex induced vibrations (VIV) will give a contribution during laying due to strong currents. Using existing codes for fatigue calculation was giving too conservative results compared to the welding technology used and experience from SCR work showed that better S-N data should be expected. A dedicated program was started as part of the Ormen Lange (OL) technology verification program overseen by Norwegian Authorities. An overview of the results is presented here. A full evaluation of the data is not yet complete. Papers will be published later presenting the full technical details and dataprocessing. Fatigue test results from the OL pipeline fatigue verification are presented focusing on the following topics: • Defect sizes in pipeline production welds; • Contractor-A: 5G welding position; • Contractor-B: 2G welding position; • 6” pipe full scale testing; • 30” pipe full scale testing; • Residual stresses; • Crack growth tests and sector specimen fatigue tests in production environments. The following are a summary of the main test variables in the program: • Mapping of actual welding defects compared to AUT results. • Welds with varying misalignment (high/low) and lack of penetration (LOP) from installation contractors tested in air. • Welds with natural welding defects in internal environment (Condensed water and formation water). • Welds with notches made by electrical discharge machining (EDM) (2×65mm and 2×15mm) in internal environment (condensed water and formation water). • Crack growth tests using large compact tension (CT) specimens in air, seawater and internal product environments (condensed water and formation water). • Full scale tests including worst case high/low, LOPs, and tests with normal welds including repair welds. The following main conclusions can be drawn from the work: • Small scale testing with representative worst case defects predicts well large scale testing results with the same features when the small scale specimen stresses are corrected for bending moments etc. arising from the cutout of the pipe. • Full scale testing of 30”×35.5mm wall thickness 2G pipes welded continuously (without start/stop) with worst case defects and high/low exceeds the D curve. • Full-scale tests of 30”×35.5mm wall thickness 5G non continuous welds with worst case defects and high/low exceeds the E curve. • Pipe welds showed low or even compressive residual stresses in the root. For continuously welded pipes the stress levels were low but more varying, also on the cap side. This partly explains the good results. • It is verified that the fatigue loads during operation are below the threshold of crack growth, and thus fatigue will not be a probable failure mechanism. This is under the condition that the measurements of vortex induced vibrations (VIV) during operation confirm the engineering calculations.Copyright