Downloaded from https://www.everyspec.com AMCP 709.138 TABLE 2-30* TABLE 2-31 RANK-SUM TEST RUN-TEST "-IGNIFICANCEI CRITERIA" *a11IGNIFICANCE CRITERIA" ar-Ul~pileneas Level a5% R A D N 0.06 0.02 0.01 n 06100e No If leaempt He If (if E WD -n" - R) ;w A hos uIs In dott) 616 27 24 23 6 3 4 7 37 '34 32 6 3 4 8 40 40 43 7 4 5 0 63 59 so S 6 a in 79 74 71 9 a 7 *11 97 91 57 10 a 8 7 IN* disfavored) *12 116 110 105 11 7 0 8 (No favored) 13 137 130 125 12 9 10 0 (N# favored) 14 160 152 147 13 9 11 10 IH 0 favored) Is lee 176 170 14 9 11 10 INOdisfavored) Is1 212 202 1"6 16 10 12 11(No diufavored) 17 241 230 223 16 11 13 12 (No sliIhtiv favored) 1s 271 259 252 20 14 16 1S (No disfavored) 119 303 291 262 20 338 324 315 Me - mull hypoomoll "weeIs no CiHerenos in "M~ Cdlu of th~e 2 aglebled, 'Reprinted from, MA60hma~o~etAwmho*.
In Reliailit~y Irvnwfsong with permvlsolon of McGraw-*Hill look Campmny, Inc. N *tfluflto of it@" in eac WOIIfloo.e from subsystem to subsystem, depending on tern) and all types of teets into meaningful their availability for testing and the amount component, equipment, and subsystem fail- of previous experience with similar Items. ure rates and reliability predictions. The Techniques have been developed for estimat- failure rates and reliability estimates can be Ing system reliability based on subsystem, updated continuously as new test data be- equipment, and component reliability data.
come available, Of course, estimating system reliability TBE23 from lower level data can never be as accurate TIS22 ns direct testing of the assembled system. MAXIMUM-OEVIATION. However, these estimates are very useful for TEST &SIGNIFICANCE decision makin before completed systems CRITERIA" are availlable for testing. 5% *41ignifleenna Level 2-12.1 ADVANTAGES OF MODEL Sample sine:
N w 10 This paragraph describes a statistical model P, mr Pr that can be used hy contractors for reliability - - - estimation and by the Army for weapon- 1 4 0.32 system reiiabiilty-monitorlng, The model 3 5 0.17 permits the combination of test data from a 0.96" all levels (from component to weapon sys- 10 a 0.95(1D 2-114