PSI - Issue 13

Konstantinos Kouzoumis et al. / Procedia Structural Integrity 13 (2018) 868–876

873

6

K. Kouzoumis et al. / Structural Integrity Procedia 00 (2018) 000–000

(a) r i / B = 1, surface flaw

(b) r i / B = 60, surface flaw

(c) r i / B = 1, through thickness flaw

(d) r i / B = 60, through thickness flaw

Fig. 1: Comparison of procedures in assessing axial flaws in cylinders

for r i / B = 1 and 20% for r i / B = 60. However, the modified BS 7910 equation designated “mod BS”, lies between the API / ASME solutions for both geometries ( r i / B = 1 & r i / B = 60), above the R6 local solution for thick cylinders ( r i / B = 1) and slightly below the later for thin cylinders ( r i / B = 60). As for through thickness flaws it is apparent that, for all geometries considered, the BS 7910 original solution produces the highest values of L r . Specifically, for thin cylinders ( r i / B = 60), the original BS 7910 equation lies further out from the solutions of all other procedures, which are grouped at lower values of L r . The modified BS 7910 equation agrees with the solutions from other procedures in assessing thin cylinders ( r i / B = 60), while for thick cylinders ( r i / B = 1) it approximates the API / ASME global solution at low values of 2 a / W and starts deviating from it at higher. Overall, the BS 7910 original solution gives much higher L r values than do other procedures, suggesting that it is reasonable to propose that the factor of 1.2 is not needed. The next step of this work is to validate the proposed modified BS 7910 formula by assessing it against experimental data, where the conditions during failure were known.

4. Validation of modified BS equation

In the process of an ongoing BS 7910 validation exercise, various experimental data has been gathered from di ff erent sources in the literature. The data presented here relate to cylinders (pipes / vessels) which contain axially oriented flaws. The number of specimens studied from each source and the type of flaw they include are presented in Table 2. All calculations were carried out with the use of CrackWISE R . The data has been analysed with the original and modified BS 7910 reference stress solutions, while stress intensity factors were calculated as advised in the procedure. In most cases the yielding behaviour of the materials assessed, i.e. continuous or discontinuous, was not known. Hence, all the assessments used FADs both with and without yield