Downloaded from https://www.everyspec.com AMCP 706.198 ,Eample No. 10 A group of 50 relays is life tested (Ref. 4). The numbers of cycles to failure are given in Table 2-10.
TABLE 2.10 CYCLES TO FAILURE 00 10 20 30.o 1283 4865 8185 13167 28946 1887 5147 8559 14833 29264 1888 5350 8843 14840 30822 2357 5353 9305 14988 38319 3437 5410 9460 16306 41554 3606 5536 9595 17621 42870 3752 6499 10247 17807 62690 3914 6820 11492 20747 03910 4394 7733 12937 21990 68888 4398 8025 12956 23449 73473 Because it often ts assumed that relay life data have a Weibull distribution, we will estimate the parameters of the Weibull distribution which fit the data; then we will test the hypothesis that the data came from that exact distribution. Procedure Examnle 1. Estimate the parameters by a graph- 1.
Sfjc} = sveifc (Cc(p) W eXpE-(c/a)01 tcal method (details not gtven here) c - cyoles-to-failure. (Eq. 2-38) = 1.2 S= 16.6 X 10э cycles N =50. 2. State the number of points. Choose 2. k 5 50/5 - 10. the number of cells k. There are 50 points, and it is nice to have an s- expected number of points in each cell of 5 or more. 3. Calculate the cell boundaries. Use 3. The Sf cell boundaries are Ч x the "equal probability" method be- , 1, 2,..., 10: 0.0, 0. 1, 0.2, cause it is handy, and does not de- 0.3,..., 0.9, 1.0. pond on the data. The range of 0-1 for Sf is divided into k equal parts. -..... 2i!