Quality control Final exam: 2012/1/12 (Thur), 9:00-12:00 Q1 Q2 Q3 Q4 Q5 YOUR NAME______________________ NOTE1: Please write down the derivation of your answer very clearly for all questions. The score will be reduced when you only write answer. Also, the score will be reduced if the derivation is not clear. The score will be added even when your answer is incorrect but the derivation is correct. 1 Q1 Basic calculations / Over-dispersion. Chips of an engineering plastic have average size 45mm and standard deviation 3mm. Upper and lower specification limits for the chips are 48mm and 41mm, respectively. 1) Suppose that X 1 ,..., X 9 are the size of the chips. Then, what is the probability of X being outside the specification limit? 2) We obtained sizes of chips (39, 44, 39, 41, 47). Find an estimate of the standard deviation by “range method” (use table). 3) Let {Y1 , Y2 , Y3 } be a independent sequence of 0-1 variables (Bernoulli trials) with 1 with probability 0.5 Pr(Yi 1) , (i 1,2,3) , Compute 0 with probability 0.5 Var ( X ), where X Y1 Y2 Y3 . 4) Let {Y1 , Y2 , Y3 } be a sequence of 0-1 variables with 1 with probability 0.5 Pr(Yi 1) , (i 1,2,3) , 0 with probability 0.5 and let Pr(Y j 1 1 | Y j 1) 0.8, ( j 1,2) . Derive Pr(Y j 1 0 | Y j 1) Pr(Y j 1 1 | Y j 0) Pr(Y j 1 0 | Y j 0) 5) Compute Var ( X * ), where X * Y1 Y2 Y3 . 6) Show over-dispersion in the above example. 2 Q2. Shewhart chart for continuous variable We conduct the testing hypothesis: H 0 : B 0 versus H1 : B 0 for the following data of k=5 groups and n=4 samples. Group i=1 i=2 i=3 i=4 i=5 Mean 10 22 13 12 14 16 20 15 13 14 12 22 12 15 9 10 16 12 12 11 1) Conduct a test for H 0 : B 0 (5% level). 2) Draw a X -Chart (draw both action limits and warning limits) 3) Draw a Range-chart (draw both action limits and warning limits) 3 si2 Range Q3. Average run length for X -chart ~ Let R be the run length for X -Chart with both warning and action limits. ~ ~ 1) E R | X1 [LWL, UWL] = , (in terms of E[R ] ) ~ ER | X ~ ER | X ~ ER | X ~ ER | X ~ ER | X 1 [LAL, UAL] = 1 ~ [UWL, UAL], X 2 [LWL, UWL] =_____, (in terms of E[R ] ). 1 [UWL, UAL], X 2 [LAL, UAL] =_____. 1 [UWL, UAL], X 2 [UWL, UAL] =_____. . [UWL, UAL], X [LAL, LWL]= ~ ( in terms of ER | X [LAL, LWL] ) 1 2 , 1 2) Let p0 Pr(LWL X1 UWL) , p1 Pr(UWL X 1 UAL ) , ~ p2 Pr(LAL X 1 LWL) . Derive E[R ] in terms of p0 , p1 and p2 . ~ 3) Derive E R | X 1 [UWL, UAL] . 4) Let X ij ~ N ( , 2 ) with 2 / n , and UAL= 3 / n , UWL= 2 / n , ~ ~ LWL= 2 / n , LAL= 3 / n . Calculate p0 , p1 , p2 and E[R ] . Compare E[R ] with the value in Table below. 4 Q4. Control chart from specification limit 1. Draw X -Chart and derive sample size for the following set of parameters under action limit only and a single specification limit at USL=0. NOTE: Denote the position of USL, UAL, and center. No need to write theoretical derivation. Producer’s risk point Consumer’s risk point pa 0.01 La 300 pr 0.05 Lr 5 X ij ~ N ( , 2 ) 1 2 2. In practice, USL is not 0. For instance, the upper specification limit for the size of chips is 48mm. Why one can always set USL=0 without loss of generality? (this problem has some erroneous point about the position of “center”) 5 Q5. Control chart for discrete data A manufacture conducts a waterproof testing for 5 electric boards in a PC. The number of electric circuits and the number of defective circuits on the board is recorded as follows: Board ID The number of defectives circuits The number of circuits 1 10 200 2 20 200 3 20 200 4 35 200 5 15 200 1. This data type is called (attribute data / countable data / continuous data) since the number of defective circuits follows (Binomial distribution / Poisson distribution / Normal distribution). A suitable chart is (c-Chart / np-Chart / X -Chart). 2. The dispersion test is to test the hypothesis: H 0 : . (write a formula) 3. Compute test statistic ( D ) and the cut-off point, and then test H 0 . 4. Draw a control chart (both action and warning limits). Is the data in-control or out-ofcontrol? 6 Tables x2 1 exp dx 2 2 z p Table: the normal distribution z P 0.5 0.6915 0.84 0.80 1 0.8413 1.5 0.9332 1.64 0.95 2 0.9772 2.33 0.99 2.5 0.9938 2.71 0.9967 3 0.9986 Table: Conversion of range to standard deviation n 2 1.128 dn 3 1.683 4 2.059 5 2.326 6 2.534 Table: Square, square-root table number square square root 5 25 2.236 6 36 2.449 7 49 2.646 8 64 2.828 9 81 3.000 10 100 3.162 11 121 3.317 12 144 3.464 Table: Factors for constructing range charts from an average range Action limit Warning limit n 4 5 D1 D3 D3 D4 0.16 0.21 2.36 2.22 0.37 0.42 1.81 1.72 Table: Average run length for various sizes Deviation: /( / n ) from the target value Average ran length (ARL) 0 278 0.5 100.06 1 25.61 1.5 8.78 2 4.07 2.5 2.41 Table: Critical point of F-distribution: F0.05(df1, df2) df1=1 2 3 4 5 6 7 8 9 df2=10 4.96 4.10 3.71 3.48 3.33 3.22 3.14 3.07 3.02 11 4.84 3.98 3.59 3.36 3.20 3.09 3.01 2.95 2.90 12 4.75 3.89 3.49 3.26 3.11 3.00 2.91 2.85 2.80 13 4.67 3.81 3.41 3.18 3.03 2.92 2.83 2.77 2.71 14 4.60 3.74 3.34 3.11 2.96 2.85 2.76 2.70 2.65 15 4.54 3.68 3.29 3.06 2.90 2.79 2.71 2.64 2.59 16 4.49 3.63 3.24 3.01 2.85 2.74 2.66 2.59 2.54 17 4.45 3.59 3.20 2.96 2.81 2.70 2.61 2.55 2.49 7 3.5 0.99977 4 0.99997 Degrees of freedom (df) χ2 value 1 0.004 0.02 0.06 0.15 0.46 1.07 1.64 2.71 3.84 6.64 10.83 2 0.10 0.21 0.45 0.71 1.39 2.41 3.22 4.60 5.99 9.21 13.82 3 0.35 0.58 1.01 1.42 2.37 3.66 4.64 6.25 7.82 11.34 16.27 4 0.71 1.06 1.65 2.20 3.36 4.88 5.99 7.78 9.49 13.28 18.47 5 1.14 1.61 2.34 3.00 4.35 6.06 7.29 9.24 11.07 15.09 20.52 6 1.63 2.20 3.07 3.83 5.35 7.23 8.56 10.64 12.59 16.81 22.46 7 2.17 2.83 3.82 4.67 6.35 8.38 9.80 12.02 14.07 18.48 24.32 8 2.73 3.49 4.59 5.53 7.34 9.52 11.03 13.36 15.51 20.09 26.12 9 3.32 4.17 5.38 6.39 8.34 10.66 12.24 14.68 16.92 21.67 27.88 10 3.94 4.86 6.18 7.27 9.34 11.78 13.44 15.99 18.31 23.21 29.59 P value (Probability) 0.95 0.90 0.80 0.70 0.50 0.30 0.20 0.10 0.05 0.01 0.001 8 IMPORTANT NOTE: This is a very simplified answer. In the exam, you need to write down more detailed calculations. Answer1. 1) X ~ N (45,1) , Pr( X 41) Pr( X 48) Pr(Z 4) Pr(Z 3) 0.0014 2) sd=(47-39)/2.326 = 3.44 3) Var(X ) =3*(1/2)*(1/2)=3/4=0.75 4) Pr(Y j 1 0 | Y j 1) =1-0.8=0.2, Pr(Y j 1 1 | Y j 0) Pr(Y j 0, Y j 1 1) / Pr(Y j 0) {Pr(Y j 1 1) Pr(Y j 1, Y j 1 1)} / Pr(Y j 0) (1 / 2 0.8 / 2) /(1 / 2) 0.2 Pr(Y j 1 0 | Y j 0) =1-0.2=0.8 5) Pr( X * 0) Pr(Y1 0, Y2 0, Y3 0) (1 / 2) * 0.8 * 0.8 0.32 . Pr( X * 1) 0.18, Pr( X * 2) 0.18, Pr( X * 3) 0.32 . Var ( X * ) 0.32 * (1.5) 2 0.18 * (0.5) 2 0.18 * (0.5) 2 0.32 * (1.5) 2 1.53 6)Over-dispersion since Var ( X * ) 1.53 0.75 Var ( X ) . Answer2 1) Overall mean=14, s B2 =46/4=11.5, sW2 =26/5=5.2, F nsB2 / sW2 =46*5/26=8.85 Critical point=F(k-1,(n-1)*k)=F(4, 15)=3.06 (5%). Reject H 0 : B 0 . 2) LAL=14-3*3.4=3.8, UAL=14+3*3.4=24.2, LWL=14-2*3.4=7.2, UWL=14+2*3.4=20.8. (Add these line and group means in the chart). 3) R =23/5=4.6 LAL=0.16*4.6=0.736, UAL=2.36*4.6=10.856, LWL=0.37*4.6=1.702, UWL=1.81*4.6=8.326. (Add these line and ranges in the chart) Answer3 ~ ~ 1) E R | X1 [LWL, UWL] = 1+ E[R ] ~ ER | X ~ ER | X ~ ER | X ~ ER | X ~ E R | X1 [LAL, UAL] = 1 . 1 ~ [UWL, UAL], X 2 [LWL, UWL] =2+ E[R ] 1 [UWL, UAL], X 2 [LAL, UAL] =2. 1 [UWL, UAL], X 2 [UWL, UAL] =2. 1 [UWL, UAL], X 2 ~ [LAL, LWL]= 1+ ER | X 9 1 [LAL, LWL] ~ 2) E[ R ] (1 p1 )(1 p2 ) (see your notes for derivation) 1 p0 p0 p1 p0 p2 p1 p2 p0 p1 p2 ~ E[ R ] 1 p1 ~ E R | X 1 [ UWL, UAL] 1 p1 (1 p1 )(1 p2 ) 1 p1 1 p0 p0 p1 p0 p2 p1 p2 p0 p1 p2 1 p1 3) (1 p2 ) 1 1 p0 p0 p1 p0 p2 p1 p2 p0 p1 p2 4) Since X1 ~ N (2 / n , 2 / n) , Z ( X1 2 / n ) /( / n ) ~ N (0,1) . p0 Pr(2 / n X 1 2 / n ) Pr(4 Z 0) 0.5 , p1 Pr(0 Z 1) 0.8413 0.5 0.3413, p2 Pr(5 Z 4) 0 . ~ E[R ] =(1+0.3413)/(1-0.5-0.5*0.3413)=1.3413/0.33=4.06. This is close to 4.07 in the Table. Answer4 1. pa =0.01 Z pa =2.33, pr =0.05 Z pr =1.64 La =300 qa 1 / La =0.0033 Z qa =2.71, Lr =5 qr 1 / Lr =0.2 Z qr =0.84. n ( Z qa Z qr ) 2 /(Z pa Z pr ) 2 =(2.71-0.84)^2/(2.33-1.64)^2=(1.87/0.69)^2=7.34. Set n =8. k A Z pa Z qa / n =2.33-2.71/sqrt(8)=1.37. UAL=- k A =-1.37*2=-2.74, center=- =-2 (usually, center is below UAL). 2. If USL=48mm for a variable X, set a new variable Y=X-48. Then, USL for Y is 0. Answer 5 1) Attribute data; Binomial distribution; np-Chart. 2) H 0 : Var ( X i ) np(1 p) , where X i is the number of defectives with ID = i , and p is unknown probability of being defective in the circuit. 3) p̂ =0.1, X =20, s 2 =350/4=87.5, s 5 7 2 / 2 =5*2.646*1.414/2= 9.354. D=350/(200*0.1*0.9)=19.44 > df2 4 (0.95) =9.49. Reject H 0 : Var ( X i ) np(1 p) . 4)UAL=20+3*9.354=48.062, UWL=20+2*9.354=38.708, LWL=20-2*9.354=1.292, LAL=20-3*9.354=-8.062 (set 0). In-control (Figure omitted). 10