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.
This data frame contains the following columns:
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).".
Fatigue.lis <- nlsList(log(length/0.90) ~ Fatigue.func(cycles, m1, m2), data = Fatigue, cluster = ~ Path) Fatigue.nlme <- nlme(Fatigue.lis)