Downloaded from https://www.everyspec.com AMCP 705.110 Examele No, 13 There are serious reliability difficulties with the gun/turret drive-sydtem on a heavy tank. A full factorial experiment with no repliation is to be run. Factor t is temperature and there are 2 levels; factor v is vibration and there are 3 levelui fac- tor b is humidity and there are 2 levels. There are 12, 1. e., (2 X 3 x 2 - 12), ex- pertments. Table 2-21(B) shows the data for this case. Each datum is the number of 10-hr missions-to-failure, Procedure Example 1. Record the data. Convert to 1. Table 2-21(B) is the original data. Subtract "overall mean ts zero." 5.28 from each to get Table 2-21(C).
Begin Begin to fill out a tabl o. Table 2-21(A) (pptterned after Table 2-20WA). Cole. I and 3 can bo completed. 2, Calculate overall averages 2. x.. * (4.75 +3.75 +2.75 +3.75 for the temperature factor. - 0, 25 - 0.25)/6 - 2.4167 !I. . - -2. 4167. 8. Cr.loulate overall aveages 3. . - (4.75 -1.25 + 3.75 - 0.25)/4 for the other factors = 1.7500, R.I. - 0.5000, R~.B. -- 2.25001 , -(4.75 - 1.25 + 2.75 - 2.25)/4 - 0.0833, .2 0. 00833. 4. As a check, the sum of the 4, OK, by inspection. factor average$, for each factor, must be zero. 5. Calculate the first 3 081s: t, v, 5. 88t - 3 x 2 x J(2.4167)1 + (- 2.4167)11 h. The suma shown in Table - 6 X 11. 0809 -a 70.
0853 2-20(A) can be simplified be- Y - 2 x x x 8.3750 - 3.1,5000 cause, in each case, 2 sums are trivial: they can be re- 88h - 3 X 2 x 0. 0139 - 0. 0833 placed with a multiplicatior.rd W Table 2-21(A), Col. 2. That multiplication is also ex- actly what is needed to convert the variance of the means to a variance of the population. 6. Calculate 1' - 8/P, for each 6. 4 - 70.0853/1 - 70. 0803 of the 8 factors. 16 0 7500 * - 0. 0833; Record in Table 2-21(A) Col. 4. ............ .........