PSI - Issue 19

Edrissa Gassama et al. / Procedia Structural Integrity 19 (2019) 711–718 Gassama et al./ Structural Integrity Procedia 00 (2019) 000 – 000

716

6

Table 1. Predicted lives for various levels of equivalent structural stress commonly found in VHCF applications

    

    

Single Slope ( 3 m = )

Bilinear Slope ( 5 m = )

Nonlinear Fit Equation (14)

MPa

Equivalent Structural Stress

m

2

−

ss

m

2

mm

ss

114.45

1.00E7 1.50E7 3.55E7 1.20E8 9.59E8

1.00E7 1.96E7 8.28E7 6.28E8 2.0E10

1.35E7 2.20E7 6.47E7 3.46E8 1.83E10

100

75 50 25

Fig. 1. Predicted curve using the model in Equation (13) when including run-outs in the Master S-N dataset.

5. Concluding Remarks

In this study, much of the dataset used to fit the Master S-N curve was investigated and many of the original sources were reviewed. The points within the dataset were verified and the run-out data points from the same sources were added (but flagged as being run-outs). An iterative least-squares approach was presented that statistically accounts for the effects of censored data. The data (with run-outs included) was then fit, and the resulting curve was compared to the single-slope Master S-N curve and a frequently recommended bilinear curve, often used for evaluating variable amplitude loading (BS7608 (2014), DNV-RP-C203 (2011)), wherein the slope is set to a value of 5 m = beyond 10 7 cycles. The nonlinear fit shows good agreement with the bilinear curve, however, the continuous nature of the nonlinear fit eliminates the ambiguities present in the bilinear curve (e.g. the slope is not prescribed and there is no need for an arbitrary intersection point to be selected). From an implementation perspective, the continuous curve is preferred over the bilinear curve as it avoids any discontinuities in the standard deviation when applying probabilistic approaches. The present work technically supports use of the bilinear curve and directionally improves the popular Master S N method to closer align with vibration fatigue legacy methods. Many industry recognized piping vibration assessment techniques either implement endurance limits or use piping vibration screening criteria that inherently incorporate assumptions of infinite life (Wachel (1990), Energy Institute (2008)). Further experimental data at stresses below 100 ( ) ( ) 2 2 ss ss MPa m m − along with the methods presented herein to incorporate runout data will provide further needed improvements to the Master S-N method.