Growth of Cracks in Metal Due to Fatigue

SUMMARY:

The Fatigue data frame has 262 rows and 3 columns. These represent growth of cracks in metal for 21 notched test units where the initial crack length was set to 0.90 inches. Each unit was subject to several thousand test cycles. After each 10,000 test cycles the crack length was measured. Testing was stopped if the crack length exceeded 1.60 inches, defined as a failure, or at 120,000 cycles. Test units, or paths, are labelled in increasing failure time or decreasing terminal crack length.

DATA DESCRIPTION:

This data frame contains the following columns:

VALUE:

Path
an ordered factor giving the test path (or test unit) number. The order is in terms of increasing failure time or decreasing terminal crack length. Thus Path == 1 denotes the test unit that failed first and Path == 21 denotes the unit with the shortest terminal crack length amongst those that did not fail.
cycles
number of test cycles at which the measurement is made (millions of cycles).
length
crack length (inches).

SOURCE:

These data are given in Lu and Meeker (1993), "Using degradation measures to estimate a time-to-failure distribution", Technometrics, 35, no. 2, 161-174. The authors note that "We obtained the data in Table 1 visually from figure 4.5.2 on page 242 of Bogdanoff and Kozin (1985).".

SEE ALSO:

, , , .

EXAMPLES:

Fatigue.lis <- nlsList(log(length/0.90) ~ Fatigue.func(cycles, m1, m2), 
          data = Fatigue, cluster = ~ Path) 
Fatigue.nlme <- nlme(Fatigue.lis)