Downloaded from https://www.everyspec.com AMCP 76-11 TABLE 2.22 REGRESSION TEST RESULTS FOR TURBINE BLADES Sleds No. Reloanse Blade No. ROnaMM Frequency, rpm Frequency, rpm Hz Hz 7 9M0 -125.1 9400 +121.5 27 1062 - 23.1 1050 79.9 18 969 -116,1 9400 + 18.3 3 1069 - 16.1 10700 148.1 2 986 - 9911 9550 - 32.3 11 1078 - 7.1 10650 -107.1 28 ON6 - 97.1 9750 144.4 17 1086 - 0,1 10600 61.2 1 909 - 87.1 9650 - 72.4 5 1090 4.9 10G60 -147.2 16 998 - 87.1 9S60 +127.6 22 1130 44.9 11000 -264.5 21 1011 - 74.1 9800 - 74.3 26 1149 63,9 11400 - 86.5 9 1012 - 73,1 10100 214.0 29 1169 83,9 11900 1798 8 1025 - 00,1 10000 - 37.9 30 1180 94.
11750 -96.7 12 1036 - 50.1 10300 145,3 4 1181 , 11600 -26004 23 1042 - 43,1 10000 -236,5 13 1190 104,0 11900 - 66,5 25 1043 - 42,1 10200 - 48,2 19 1215 129.9 11950 -307.6 10 104V - 38,1 10300 + 6.1 24 1217 131.9 12200 - 80.9 20 1055 - 30,1 10600 111,6 14 1240 154.9 12350 -199.6 15 1058 -. 27.1 10300 -123.1 6 1271 185.9 13800 888.2 Units are omitted In the aglcgultions, Mean of x' a,- 10•.10 tamned by means of a statistical sampling detect a difference is determined by a,13 procedure, there is a chance of making an (for a given 8), and N the sample size.
The incorrect decision, The probability of quantity I - P (6) is called the Power of making an incorrect decision usually can be the test to detect a difference 8 with a reduced by Increasing the number of sam- sample of size N when the test is performed pies tested. Two types of wrong decisions at an a level of s-significance. The relation- are possible: ship between these parameters can be de- scribed graphically by an 'Operating Chur- I. It is concluded that (here is a differ- acteristic (OC) curve. These curves describe ence but, in fact, there is none (Error of the the discriminatory power of a test. There is First Kind).
The probability of making this a unique OC curve that corresponds to error is denoted by a. specified values of N and a. These and other kinds of OC curves are discussed in greater 2. It is concluded that there is no differ- detail in Chapter 3. once but, in fact, there is one (Error of the Second Kind). The probability of' making Two basic types of tests can be con- this error is denoted by /. sidered: The probability of an Error of the Second I. Does the mean differ from a specified Kind Is related to the size of the difference requirement? 4 6 being measured. The value of 1 associated with a particular 8 decreases as 6 increases. 2.
Does the mean of one design differ ', " For a specific statistical test, the ability to from the mean of another design? -- ': ',. W . , . - .. .3 .3 ,,, 3 .3 * t3 ' , , . ,I o -