PSI - Issue 57

Inge Lotsberg et al. / Procedia Structural Integrity 57 (2024) 569–580 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

570 2

the fillet weld roots. Post weld improvement can be performed of the fillet weld toes; however, as the fillet weld roots cannot be improved it is questioned if fatigue cracking from these areas will occur if higher stress ranges are accepted after improvement of the weld toes. Due to this uncertainty weld improvement of fillet welded doubling plates is not recommended by IIW in Hobbacher (2009). At a late fabrication stage of a floating platform, the calculated fatigue life at a doubling plate supporting a ladder was too short, and it was proposed to perform grinding of the weld toes to improve the fatigue life. As doubling plates is listed under examples of details that are not suitable for weld improvement in IIW (Hobbacher, 2009) it was decided to perform fatigue testing of ground doubling plates with the same geometry as was used in the floating platform. Furthermore, it was decided to perform fatigue testing of the doubling plates in as-welded condition to get a reference for the effect of the grinding improvement. Some of the results from the fatigue tests are further described in section 2. The hot spot stress analysis methodology using finite element analysis is based on calibration with fatigue test data, and it goes back to the beginning of the 90s for details in plated structures, see e.g. Niemi (1995). This methodology was described in the first recommendations by IIW in 1996 edited by Hobbacher (1996). Before this time the hot spot stress methodology had been used in fatigue design of tubular joints in offshore structures in the North Sea since the end of the 70s; see e.g. the Fatigue Handbook (1985) and some historical development of the hot spot stress methodology in Lotsberg (2016). More welded details with different plate geometries were analysed in a study on hot spot stress presented by Fricke (2001) which improved the confidence in these methods for fatigue design of plated structures. A comparison of different analysis procedures was also presented by Poutiainen et al. (2004). A guideline on hot spot stress methodology was presented by Niemi et al. (2006). Similar guidelines on finite element modelling and hot spot stress derivation are also included in DNV-RP-C203 (2019). In some cases, it may be difficult to perform finite element model according to recommendations in fatigue design standards such as DNV-RP-C203 (2019) as basis for hot spot stress derivation. Analysts may also believe that a refined finite element model may provide more accurate hot spot stresses. However, here it is important to remember that the methodology has been derived from calibration with test data using predefined finite element meshes where convergence in results have not necessarily been achieved. An example with a fillet welded detail where a different finite element model is used than that recommended in the design guidance in DNV-RP-C203 (2019) is shown in Fig. 1. A plate is fillet welded to a vertical section as indicated in the sketch. The plate is also fillet welded to the edges of vertical section on the back side of the plate along the dotted lines shown in this figure. A significant dynamic vertical load is transferred through the vertical section to the bottom plate through the fillet welds. Due to the finite element modelling made it was assessed that some further analyses were recommended to improve the confidence in the calculated fatigue lives at the fillet weld toe and root. For this purpose, the referred large-scale test specimen with fillet welded doubling plates was selected for analysis using a similar element modelling. This work is further described in Section 3. First mainly a relative assessment of the two details in Fig. 1 and Fig. 2 was performed. Later the work was extended to learn more about the accuracy of the different hot spot stress analysis methods that are presented in literature and in fatigue design standards. On this basis the main purpose of the performed analyses described in this paper was to assess the sensitivity in results from finite element models that are somewhat different from recommendations in analysis guidelines by comparison of calculated values with fatigue test data.