PSI - Issue 19

Jennifer Hrabowski et al. / Procedia Structural Integrity 19 (2019) 267–274 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

273

7

for very large eccentricities with e/h 0 = 0.35. The influence of  is the opposite for the chord and the braces. For SCFs of the chord, small wall thickness ratios  are favorable, but larger ones for the braces.

Table 3. Recommendation for geometric parameters of K-joints with gap. RHS

CHS

Axial Load

In-plane bending Out-of-plane bending

Brace loading Parameter

Axial Load

In-plane bending

small

small

small

small

Wall thickness ratio  = t i / t 0

(large)

(small)

(small or) large

(small or) large

Width ratio

 = b i / b 0

small small

small small

small small

small small

Chord slenderness 2  = b 0 / t 0

Brace opening angle 

For the structural design of the CHS K-joints under fatigue loading, it is recommended to choose the wall thickness of the brace thin (small  -value) in relation to the chord wall thickness, and the chord profile should be thick-walled (smaller 2  value). The smaller the parameters  and 2  are, the lower are the SCFs and thus more favorable for the fatigue design for the three load cases AX, IPB and OPB. Regarding the width ratio  , which has rather small influence and is therefore in parenthesis, it is advisable to select large  -values for the load case AX and small  -values for the load cases IPB and OPB, so for IPB and OPB the same recommendations apply. For RHS as well as for CHS K-joints smaller brace angles  are preferable. In addition to the recommendations in Table 3, it is also recommended to keep the joints’ eccentricities as small as possible . Generally, it is also recommended to perform butt welds with appropriate profile preparation instead of fillet welds to ensure a full connection of the brace. On the other side, the fillet weld provides a smoother weld seam transition. Therefore, it may also be useful to weld a fillet weld onto the butt weld (Kuhlmann et al. 2015). After the assessment of the state of strain of the framework and selecting the most favorable dimensions, the decision about the verification method must follow. This is expected to be largely determined by the scope. The nominal stress approach applies only for wall thicknesses up to 8 mm and is therefore very limited. The detail categories of EN 1993-1-9 (2010) or FAT of IIW (Hobbacher 2016) and ISO 14347 (2008) represent a lower limit as they include multiple influences such as weld execution, load introduction and parameter dependency, as the evaluation of the test results in Fig. 2 shows. The nominal stress approach often cannot meet the complex behavior of hollow section joints and the available detail catalogues do not provide sufficient categories for hollow sections. This is why CIDECT (Zhao et al. 2001) and IIW (Hobbacher 2016) as well the offshore industry prefers the structural stress approach for the fatigue design. It accounts for the uneven stress distribution in hollow section joints. The structural stress at the decisive position, the so-called hot spot, includes effects from joint geometry, stiffness distribution and the load cases. In contrast to that the influence of the local weld geometry and imperfections are considered in the relevant  ,hs -N f diagram. Zhao et al. (2001) offer formulae and diagrams for the determination of stress concentration factors (SCF). The scope of application is considerably larger and depending on the influencing parameters, see Table 4. The offshore standard DNV GL-PR-C203 (2016) e.g. offers an even wider range of applications. However, in the calculated SCFs some discrepancy to Zhao et a. (2001) occur, which makes it clear that the scope should be strictly respected.

Table 4. Application range for SCFs of K-joints with gap acc. Zhao et al. (2001). Parameter RHS

CHS

0.25 – 1.0 0.3 – 0.6

0.25 – 1.0 0.35 – 1.0

Thickness ratio

 = t 1 / t 0  = b 1 / b 0

Width ratio

24 - 60

10 - 35

Chord slenderness Brace opening angle

2  = b 0 / t 0

30° - 60°

30° - 60°



0

-0.55 – 0.25

Non-dimensional eccentricity e / h 0