OS-C501
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
Offshore Standard DNV-<strong>OS</strong>-<strong>C501</strong>, November 2013<br />
Sec.4 Materials - laminates – Page 55<br />
3.11.3 This section is applicable for estimating the characteristic (and subsequently the design) time to failure<br />
under a specified load for a laminate exposed to static or cyclic load, provided the plot of log stress vs. log time<br />
is linear.<br />
3.11.4 If the linear relationship cannot be documented, an equivalent approach shall be used, taking the nonlinearity<br />
into account.<br />
3.11.5 Values shall be based on data that are fairly evenly distributed over the plot of log time to failure vs. log<br />
load, or log number of cycles vs. log load. Load is usually expressed as stress or strain. At least 15 data points<br />
should be used.<br />
3.11.6 To obtain the characteristic curve the mean S-N curve of the form:<br />
log σ = log σ 0 fatigue - α log N<br />
or the mean stress rupture curve of the form:<br />
log σ = log σ 0 stress rupture - β log t<br />
shall be converted to the form:<br />
log(X) mean = log(X 0 ) – k⋅logσ<br />
where X represents the time (or number of cycles) to failure under a sustained stress σ (or stress range σ). X is<br />
a function of σ and exhibits a natural variability from point to point within the material.<br />
logσ<br />
0<br />
log X =<br />
0<br />
β<br />
stress rupture<br />
k =<br />
1<br />
β<br />
or<br />
or<br />
log<br />
log =<br />
X<br />
0<br />
1<br />
k =<br />
α<br />
σ 0 fatigue<br />
α<br />
Guidance note:<br />
Usually, estimates of k and logσ 0stressrupture (or logσ 0fatigue ) can be obtained from linear regression analysis of log X<br />
on log σ.<br />
---e-n-d---of---G-u-i-d-a-n-c-e---n-o-t-e---<br />
3.11.7 When the standard deviation σ ε of the variations in log(X) about the mean is constant, i.e. when σ ε does<br />
not depend on the sustained load or stress range σ, then the characteristic value of log(X) c can be taken as:<br />
log(X) c = log(X 0 ) – k⋅logσ - x⋅σ ε<br />
in which σ ε is estimated from available tests, and x is taken from Table 4-7 depending on the number n of<br />
available data pairs (logs, log X) from tests.<br />
Table 4-7 Values of coefficient x<br />
x<br />
n (# of tests) Case 1 Case 2<br />
10<br />
15<br />
20<br />
50<br />
100<br />
Infinite<br />
3.9<br />
3.4<br />
3.1<br />
2.6<br />
2.4<br />
2.0<br />
4.7<br />
4.0<br />
3.7<br />
3.0<br />
2.6<br />
2.0<br />
3.11.8 The coefficient values marked as Case 1 are valid and can be used for sustained loads or cyclic stresses<br />
within the range of σ-values covered by available tests, i.e. whenever the available tests cover a wide enough<br />
range of σ-values. These coefficient values will be non-conservative if applied for sustained loads or stresses<br />
σ outside the range of log σ-values covered by available tests. When values for x are needed for σ-values<br />
outside this range, coefficient values marked as Case 2 can be used for extrapolation within a concentric range<br />
of logσ twice the length of the range covered by tests.<br />
3.11.9 The mean curve can be transformed back into the standard formulation of an S-N curve, or stress rupture<br />
or fatigue curve using the same equations as given above.<br />
log<br />
— Stress rupture curve: σ<br />
X<br />
1<br />
0<br />
log 0stress<br />
rupture = and β = with X 0 as time.<br />
β<br />
k<br />
log<br />
X<br />
1<br />
0<br />
— Fatigue curve: log σ 0 fatigue = and α = with X 0 as number of cycles.<br />
α<br />
k<br />
3.11.10 The characteristic mean curve can be transformed back into the standard formulation of an S-N curve<br />
or stress rupture curve using the same equations as given above.<br />
DET NORSKE VERITAS AS