Reliability Analytics Corporation
Previous | Next      Search Home

First 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 Last

First 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 Last

Previous | Next      Search Home
MIL-HDBK-781 The eoodness-of-fit of the AMSAA model to the Darticular test data beini! Generated musi he p - | w V -tested by use of the Cramer-von Mises goodness-of-fit test. First, the level of significance (a of the test must be chosen and the critical value of the test statistic (C2M) determined from TABLE V. The (C2m) calculated from the observations (equations 6 and 10 in TABLE III) must then be tUic /^fiti/"<" 1 irolitA If tKa etatietii* ic IACC tVion tViA taKiilatfo/4 pritip^l traltiA tUa wuiiipoivu UIIJ vnuwai Voiuv. ii uiv atatiaiiw 10 IVJO UIUII uiv louuioivu viuivai vaiuw, MW AMSAA model cannot be rejected and the calculation procedure in steps a through g below can be used. If the statistic is greater than the tabulated critical value, then the AMSAA model is rejected. If the model is rejected, follow the procedures given in step h below. a. If the AMSAA model is appropriate, the system intensity function may be estimated as a function of time by: A A (5 I . , p(t) = ' (for large samples) p(t) = (for small samples) The intensity function is equal to the derivative, at time (t), of the expected number of failures in the interval (0,t). b. Then calculate (p(t)) or (p(t)) at the end of the test (or at the point in the test at which the calculation is being made). c. From TABLE VI for failure-terminated test and from TABLE VII for time-terminated test, obtain the two-sided lower confidence bounds (LN, Y) and two-sided upper confidence bounds (UN,y) for N failures and (y) percent confidence coefficient. d. Compute the interval estimate of MTBF from: LNy u^y "T S MTBF X -r PW IiU) e. If the number of failures is 20 or more, the same percentiles may be used to construct approximate confidence bounds on the future MTBF. f. The MTBF is: A M(t) = 1 /p(t) (for large samples) M{t) = 1 /p(t) (for small samples) g. From the confidence limits on M(t) previously calculated, the corresponding limits for (t) may be found from: = i'Mrt

Toolkit Home

Reliability Analytics Corporation
reliabilityanalytics.com