Downloaded from https://www.everyspec.com AMIP 70&109 BF aw~ls No. M(11 This example usesv Data Sft D, Table 2-4. Dt SetD. DK-4 ]b"Oda Method (Tabl sh , F8 , 2-E).l 1. PrWp 'oCol. 3. Choose an s-Mo- 1. N , 21; 1 - I/N - 0.952. idence level, and find KS, from the Choose 95% e-oonfidence. reverse side of FMS. 2-6. KS, - 0.29 (from the formula). 2. Plot the data in Col. 3 and the 2. See Fig. 2-4. bounds, as indLoastd on the reverse 3. Draw the two parallel lines for en- 3. See lines #I and #2. I aide of Fig. 2-6. timating the lower and upper limits ofr, a, andau. acandsuarethe intersections of lines 01 and #2 au - 280 hr. with the heavy dashoэ line. 4. Draw the two intersecting lines for 4.
For line 03, choose the deuoads from estimating the lower and upper Irn[- 30 to 300, St - 2. 05, B 2 "3. 10; its of P, . and Ou. P& and Pu are thuI AL "3..10 - 2 . 05 a 1. 05. Fox calculated fron. the elopes of lines line #4, choose the half-decade from 03 and $4. 100 tO 31.6. B, - 1,1 0, Bt - 3. W; thus Ou - (S. 90 - i. 10)/0. 5 - 5. 0, 5. Use the "best fit' line to make point. 5. From line #5: estimate ofa and 0f, & aned#. -220 hr Thome data could riky have come from a Weibull distribution, although the data A do not lie on a very straigIt line. The parent population aotually was WOiWll with a - 200 hr and p - 2.5. ehis line (not shown in Fig.
2-6) is well within the 9-8 ou- vtulop, bt is not a good fit to the data points. The data themmelvA..: do not lie well on their beet.-fKt line (05). !Se very skeptical when tm data do oit, on u nice utraight linel someone may have given thi data a prolimtnary mkwnэ-.)ulasion. ...................... .. ... . . .. .4 ., k ..