111al process capability strategies

1、1-11a-p1 week 1: introduction to process capability general electric proprietarydeveloped by w. scott lasater - geics process capability analysis overview brief 1-11a-p2 week 1: introduction to process capability general electric proprietarydeveloped by w. scott lasater - geics collect timely data p

2、rocess stable/consistent? (spc) process acceptable? (capability, z) develop measures with respect to standards identify standards (specs) that address customer requirements (ctqs) yes no* yes analyze phase use “special cause strategy” (how to improve an unstable system) act quickly to search for the

3、 cause: what was different? which x just exerted its influence? “message is in a single point” use “common cause strategy” (how to improve a stable system) investigate relationships among inputs (xs) and output (y) via: multi-vari study(stratification) experimentation (doe) disaggregation (break int

4、o components) “message is in all the points” measure phase focus, define, process map, c&e, fmea, etc. evaluate measurement system (msa) improve phase plan & implement short-term remedy (immediate containment) plan & implement long-term remedy verify impact/success (before vs after) plan & implement

5、 long-term remedy verify impact/success (before vs after) control phase modify control plans: hold gains complete hand-off activities ensure new procedures are used and are effective monitor for innovation & improvement opportunities monitor ever-changing customer needs and requirements no * =this d

6、ata from an unstable, unpredictable process can still be used for a “pseudo-capability study” - a rough “before vs after” guess at how much progress may be possible if we succeed in improving the “approximate z(lt)” (using “overall stddev”) to z(st). developed by w. scott lasater - geics define 1-11

7、a-p3 week 1: introduction to process capability general electric proprietarydeveloped by w. scott lasater - geics what sorts of data do you see being collected around your area? (list them below) _ _ _ _ _ _ _ _ _ _ _ attributes datavariables data 1-11a-p4 week 1: introduction to process capability

8、general electric proprietarydeveloped by w. scott lasater - geics how to get capability indices via minitab (1) variables data (2) attributes data .a quick look 1-11a-p5 week 1: introduction to process capability general electric proprietarydeveloped by w. scott lasater - geics open file “capab_v.mt

9、w”. characterize process capability. lsl= 40 and usl= 100 what type of data is this (attribute or variables)? variables zshort term= 4.67 zlong term= 3.74 your capability study on “percent late shipments” over the past four weeks has revealed at total of 400 “late” out of a total of 4000 shipments.

10、characterize process capability. what type of data is this (attribute or variables)? attribute zshort term= 2.782 zlong term= 1.282 important note: x-bar chart shows unstable process - may even be 2 different processes. 1-11a-p6 week 1: introduction to process capability general electric proprietary

11、developed by w. scott lasater - geics minitab command: six sigma/process report/single column output/subgroup size subgroup (see next pages for details) capability for variables data open file “capab_v.mtw”. characterize process capability. lsl= 40 and usl= 100 what type of data? variables data zsho

12、rt term= 4.67 zlong term= 3.74 important note: x-bar chart shows unstable process - may even be 2 different processes. 1-11a-p7 week 1: introduction to process capability general electric proprietarydeveloped by w. scott lasater - geics .then click ok. 1-11a-p8 week 1: introduction to process capabi

13、lity general electric proprietarydeveloped by w. scott lasater - geics a c tual (lt) p otential (s t) 100908070605040 process performance usllsl a c tual (lt) p otential (s t) 1,000,000 100,000 10,000 1000 100 10 1 3020100 potential (st)actual (lt) sigma ppm (z.b enc h) process benchmarks 1.51584 3.

14、74 93.8238 4.67 process demographics 40 100 opportunity: nominal: lower spec: upper spec: units: characteristic: process: department: project: reported by: date: report 1: executive summary report #1 1-11a-p9 week 1: introduction to process capability general electric proprietarydeveloped by w. scot

15、t lasater - geics 15 10 5 0 s=5.876 3.0sl=12.27 -3.0sl=0.000 3020100 82 72 62 xbar and s chart subgroup x=72.81 3.0sl=81.19 -3.0sl=64.42 100 40 88.752451.2476 potential (st) capability process tolerance specifications iii iii 100 40 94.634350.9790 actual (lt) capability process tolerance specificati

16、ons iii iii mean stdev z.usl z.lsl z.bench z.shift p.usl p.lsl p.total yield ppm cp cpk pp ppk ltst capability indices data source: time span: data trace: 1.25 1.37 93.8238 99.9906 0.000094 0.000003 0.000091 0.9336 3.7351 4.5165 3.7437 7.2637 72.8067 1.45 1.60 1.51584 100.000 0.000002 0.000001 0.000

17、001 0.9336 4.6687 4.8094 4.8094 6.2378 70.0000 report 2: process capability for output note: x-bar chart shows unstable process, may be 2 different processes report #2 1-11a-p10 week 1: introduction to process capability general electric proprietarydeveloped by w. scott lasater - geics your capabili

18、ty study on “percent late shipments” over the past four weeks has revealed at total of 400 “late” out of a total of 4000 shipments. characterize process capability. what type of data? attribute data zshort term= 2.782 zlong term= 1.282 example 1 defect = 1 late shipment400 1 unit = 1 shipment4000 1

19、opportunity for defect per unit1 capability for attributes data 1-11a-p11 week 1: introduction to process capability general electric proprietarydeveloped by w. scott lasater - geics .then click ok. 1-11a-p12 week 1: introduction to process capability general electric proprietarydeveloped by w. scot

20、t lasater - geics zone of average technology zone of typical control 3.0 2.5 2.0 1.5 1.0 0.5 0.0 6543210 z.shift z.bench (short-term) world-class performance report 8b: product benchmarks note: zbench = zst triangle shows that zst is somewhere slightly less than 3. 1-11a-p13 week 1: introduction to

21、process capability general electric proprietarydeveloped by w. scott lasater - geics total 1 charac teris tic 2.782 2.782 zb enc h 1.500 1.500 zs hift 100000 100000 p p m 0.100000 0.100000 dp o 0.100 dp u 4000 4000 totopps 1 opps 4000 units 400 400 defs report 7: product performance note: zst = “zbe

22、nch” = 2.782 zshift is assumed 1.5 for attribute data zlt will be 1.5 worse than zst. thus, zlt = 2.782 -1.5 = 1.282 z(st)=2.78 and, thus, z(lt)=1.28 1-11a-p14 week 1: introduction to process capability general electric proprietarydeveloped by w. scott lasater - geics % r e je c te dz (l t )z (s t )

23、 9 9 %-2 .3 3-0 .8 3 9 8 %-2 .0 5-0 .5 5 9 7 %-1 .8 8-0 .3 8 9 6 %-1 .7 5-0 .2 5 9 5 %-1 .6 4-0 .1 4 9 4 %-1 .5 5-0 .0 5 9 3 %-1 .4 80 .0 2 9 2 %-1 .4 10 .0 9 9 1 %-1 .3 40 .1 6 9 0 %-1 .2 80 .2 2 8 9 %-1 .2 30 .2 7 8 8 %-1 .1 70 .3 3 8 7 %-1 .1 30 .3 7 8 6 %-1 .0 80 .4 2 8 5 %-1 .0 40 .4 6 8 4 %-0

24、.9 90 .5 1 8 3 %-0 .9 50 .5 5 8 2 %-0 .9 20 .5 8 8 1 %-0 .8 80 .6 2 8 0 %-0 .8 40 .6 6 7 9 %-0 .8 10 .6 9 7 8 %-0 .7 70 .7 3 7 7 %-0 .7 40 .7 6 7 6 %-0 .7 10 .7 9 7 5 %-0 .6 70 .8 3 7 4 %-0 .6 40 .8 6 7 3 %-0 .6 10 .8 9 7 2 %-0 .5 80 .9 2 7 1 %-0 .5 50 .9 5 7 0 %-0 .5 20 .9 8 6 9 %-0 .51 6 8 %-0 .4

25、71 .0 3 6 7 %-0 .4 41 .0 6 6 6 %-0 .4 11 .0 9 6 5 %-0 .3 91 .1 1 6 4 %-0 .3 61 .1 4 6 3 %-0 .3 31 .1 7 6 2 %-0 .3 11 .1 9 6 1 %-0 .2 81 .2 2 6 0 %-0 .2 51 .2 5 5 9 %-0 .2 31 .2 7 5 8 %-0 .21 .3 5 7 %-0 .1 81 .3 2 5 6 %-0 .1 51 .3 5 5 5 %-0 .1 31 .3 7 5 4 %-0 .11 .4 5 3 %-0 .0 81 .4 2 5 2 %-0 .0 51 .

26、4 5 5 1 %-0 .0 31 .4 7 5 0 %01 .5 example: if a process has 10% reject rate, we would say it has a capability of z(lt)=1.28 and z(st)=2.78 z-table for attribute data: %rejected to z(lt), then z(lt)+1.5=z(st) 4 9 %0 .0 31 .5 3 4 8 %0 .0 51 .5 5 4 7 %0 .0 81 .5 8 4 6 %0 .11 .6 4 5 %0 .1 31 .6 3 4 4 %0

27、 .1 51 .6 5 4 3 %0 .1 81 .6 8 4 2 %0 .21 .7 4 1 %0 .2 31 .7 3 4 0 %0 .2 51 .7 5 3 9 %0 .2 81 .7 8 3 8 %0 .3 11 .8 1 3 7 %0 .3 31 .8 3 3 6 %0 .3 61 .8 6 3 5 %0 .3 91 .8 9 3 4 %0 .4 11 .9 1 3 3 %0 .4 41 .9 4 3 2 %0 .4 71 .9 7 3 1 %0 .52 3 0 %0 .5 22 .0 2 2 9 %0 .5 52 .0 5 2 8 %0 .5 82 .0 8 2 7 %0 .6 1

28、2 .1 1 2 6 %0 .6 42 .1 4 2 5 %0 .6 72 .1 7 2 4 %0 .7 12 .2 1 2 3 %0 .7 42 .2 4 2 2 %0 .7 72 .2 7 2 1 %0 .8 12 .3 1 2 0 %0 .8 42 .3 4 1 9 %0 .8 82 .3 8 1 8 %0 .9 22 .4 2 1 7 %0 .9 52 .4 5 1 6 %0 .9 92 .4 9 1 5 %1 .0 42 .5 4 1 4 %1 .0 82 .5 8 1 3 %1 .1 32 .6 3 1 2 %1 .1 72 .6 7 1 1 %1 .2 32 .7 3 1 0 %

29、1 .2 82 .7 8 9 %1 .3 42 .8 4 8 %1 .4 12 .9 1 7 %1 .4 82 .9 8 6 %1 .5 53 .0 5 5 %1 .6 43 .1 4 4 %1 .7 53 .2 5 3 %1 .8 83 .3 8 2 %2 .0 53 .5 5 1 .0 %2 .3 33 .8 3 0 .9 %2 .3 73 .8 7 0 .8 %2 .4 13 .9 1 0 .7 %2 .4 63 .9 6 0 .6 %2 .5 14 .0 1 0 .5 %2 .5 84 .0 8 0 .4 %2 .6 54 .1 5 0 .3 %2 .7 54 .2 5 0 .2 %2

30、 .8 84 .3 8 0 .1 %3 .0 94 .5 9 0 .0 1 %3 .7 25 .2 2 0 .0 0 1 %4 .2 65 .7 6 1-11a-p15 week 1: introduction to process capability general electric proprietarydeveloped by w. scott lasater - geics process capability: the concepts behind the calculations 1-11a-p16 week 1: introduction to process capabil

31、ity general electric proprietarydeveloped by w. scott lasater - geics y = weight (lbs)220 160 100 process = hose 1 drop = 1 unit of output histogram is . a pile of individual values dotplot : : : : . : . : : : : : : : : : : : : : . : . . :.: :.:.:.:.:.: : : : : . : . . -+-+-+-+-+-+-c1 100 125 150 17

32、5 200 225 1-11a-p17 week 1: introduction to process capability general electric proprietarydeveloped by w. scott lasater - geics 5 ft 4 ft 3 ft 2 ft 1 ft 0 fttime what has changed? centering? _ spread? _ 4.03.02.0 time 1-11a-p18 week 1: introduction to process capability general electric proprietary

33、developed by w. scott lasater - geics dot plots 1st observation2nd observation 1.11.0 1.15 1.2 1.25 1.3 1.35 1.4 1.05 suppose we have a manufacturing line that is producing shafts. diameters range from 1.0 to 1.4 inches. as we make a measurement of a shaft, we record the value with a dot on the abov

34、e scale ex: 1st observation = 1.4 inches 2nd observation = 1.1 inches 1-11a-p19 week 1: introduction to process capability general electric proprietarydeveloped by w. scott lasater - geics and suppose we continue sampling until 150 shafts have been measured what statements can you make about our pro

35、cess ? : :. . . :. : : : . :. .: : :. :.: .:.:.:.: . dot plots 1.11.0 1.15 1.2 1.25 1.3 1.35 1.4 1.05 1-11a-p20 week 1: introduction to process capability general electric proprietarydeveloped by w. scott lasater - geics : :. . . :. : : : . :. .: : :. :.: .:.:.:.: . dot plots 1.11.0 1.15 1.2 1.25 1.

36、3 1.35 1.4 1.05 now imagine the same data, grouped into intervals with bars used to represent how the data “looks”. 1-11a-p21 week 1: introduction to process capability general electric proprietarydeveloped by w. scott lasater - geics data represented just with the dots is called a dot plot using da

37、ta represented in the above “bar format” is called a histogram histogram distribution 1.11.051.0 35 30 25 20 15 10 5 0 frequency 1.151.21.251.3 1.351.4 1-11a-p22 week 1: introduction to process capability general electric proprietarydeveloped by w. scott lasater - geics so how are we doing? are we c

38、apable of doing what the customer requires? 7.7%92% what % will fall out here (outside the boundary)? 1-11a-p23 week 1: introduction to process capability general electric proprietarydeveloped by w. scott lasater - geics histogram now weve combined the histogram with our lower and upper specificatio

39、ns. question #1 : what are specifications ? where do they come from ? question #2: what can you say about our process now ? 1.11.0 1.15 1.2 1.25 1.3 1.35 1.4 1.05 upper specificationlower specification .0012.0 1-11a-p24 week 1: introduction to process capability general electric proprietarydeveloped

40、 by w. scott lasater - geics histogram suppose the customer has given us new specifications ! question: what can you say about our process now ? 1.11.0 1.15 1.2 1.25 1.3 1.35 1.4 1.05 lower specification 1.1 upper specification 1.3 1-11a-p25 week 1: introduction to process capability general electri

41、c proprietarydeveloped by w. scott lasater - geics ucl and lcl versus usl and lsl upper control limits = ucl lower control limits = lcl upper specification limits = usl lower specification limits = lsl if a point falls beyond the ucl or lcl does this mean we are making a defective item for the custo

42、mer ? ucl lcl time y-axis (response) 1-11a-p26 week 1: introduction to process capability general electric proprietarydeveloped by w. scott lasater - geics ucl and lcl versus usl and lsl upper control limits = ucl lower control limits = lcl upper specification limits = usl lower specification limits

43、 = lsl is the process below making defective items ? ucl lcl time y-axis (response) usl lsl 1-11a-p27 week 1: introduction to process capability general electric proprietarydeveloped by w. scott lasater - geics ucl and lcl versus usl and lsl upper control limits = ucl lower control limits = lcl uppe

44、r specification limits = usl lower specification limits = lsl is the process below making defective items ? ucl lcl time y-axis (response) usl lsl 1-11a-p28 week 1: introduction to process capability general electric proprietarydeveloped by w. scott lasater - geics the goal . . . 1-11a-p29 week 1: i

45、ntroduction to process capability general electric proprietarydeveloped by w. scott lasater - geics phase one - unpredictable performance - variation (special / natural causes) - unpredictable (hourly, daily) - detect and eliminate special causes phase two - stability - in control - natural variatio

46、n only time how we progress toward the goal not “capable” of getting all the water output into the clowns mouth? which approach to use? special cause strategy? common cause strategy? 1-11a-p30 week 1: introduction to process capability general electric proprietarydeveloped by w. scott lasater - geic

47、s in control, but not capable (variation from common causes excessive) in control and capable (variation from common causes reduced) time how we progress toward the goal size lower specification limitupper specification limit now it is “capable” of getting all the water output into the clowns mouth

48、- with extra elbow room for unexpected shifts & drifts (i.e., someone bumps your arm). which approach to use? special cause strategy? common cause strategy? 1-11a-p31 week 1: introduction to process capability general electric proprietarydeveloped by w. scott lasater - geics variables data 0% reject

49、ed target attributes data the goal is . 1-11a-p32 week 1: introduction to process capability general electric proprietarydeveloped by w. scott lasater - geics (1) pretending that i stay perfectly centered, is it even possible for this van to fit in there? (.a spread width issue) (2) now, taking into

50、 consideration my past steering (centering) performance, am i able to make it without bumping the walls/boundaries (the customer specifications)? (.both spread and centering issues) its sort of like trying to drive a u-haul moving van/truck through a tunnel. z(short term) z(long term) “short term” c

51、apability vs “long term” performance later, well discuss the importance and implications of consistency (stability over time) and the lack thereof. spread: how would my estimating/predicting ability be affected if my trucks body had been unexpectedly expanding and contracting on occasion? centering:

52、 how would my estimating/predicting ability be affected if my steering had been shifting significantly and unexpectedly every once in a while? 1-11a-p33 week 1: introduction to process capability general electric proprietarydeveloped by w. scott lasater - geics lslusl 1.11.3 capable? yesno meeting s

53、pecifications? capable? yesno meeting specifications? capable? yesno meeting specifications? capable? yesno meeting specifications? z(st)= z(lt)= z(st)= z(lt)= z(st)= z(lt)= z(st)= z(lt)= “visual estimates” of sigma levels: 1-11a-p34 week 1: introduction to process capability general electric propri

54、etarydeveloped by w. scott lasater - geics notes: 1-11a-p35 week 1: introduction to process capability general electric proprietarydeveloped by w. scott lasater - geics capability analysis roadmap variables data (i.e., continuous data) part 1 start evidence of inadequate measurement units? spread st

55、able? r or mr chart show reasonable degree of statistical control? calculate z(st) or cp centering stable? x or x-bar chart show reasonable degree of statistical control? histogram, nml plot, and nml test indicate normality? calculate z(lt) or cpk data is too discrete. control charts and capability

56、indices unreliable. need to improve meas. process. yes take necessary steps to address special causes take steps to address special causes & calculate rough guess of “pseudo-capability” = z(lt) (see comments in figure 1.1) no no yes yes no yes no consider another distribution to describe the process

57、. need transformation? see next page. call bb,mbb has activity of sources of variation been thoroughly considered relative to subgrouping structure? yes 1-11a-p36 week 1: introduction to process capability general electric proprietarydeveloped by w. scott lasater - geics appendix #1 probability and

58、area under a “normal distribution” curve (1) via minitab (2) normal tables (i.e., z-tables) 1-11a-p37 week 1: introduction to process capability general electric proprietarydeveloped by w. scott lasater - geics normal distribution (lower tail areas) i have the z value = -1.42, and need “area” (lower

59、 tail area) area =0.0778 z= -1.420 z0.000.010.020.030.040.050.060.070.080.09 -3.60.00020.00020.00010.00010.00010.00010.00010.00010.00010.0001 -3.50.00020.00020.00020.00020.00020.00020.00020.00020.00020.0002 -3.40.00030.00030.00030.00030.00030.00030.00030.00030.00030.0002 -3.30.00050.00050.00050.0004

60、0.00040.00040.00040.00040.00040.0003 -3.20.00070.00070.00060.00060.00060.00060.00060.00050.00050.0005 -3.10.00100.00090.00090.00090.00080.00080.00080.00080.00070.0007 -3.00.00130.00130.00130.00120.00120.00110.00110.00110.00100.0010 -2.90.00190.00180.00180.00170.00160.00160.00150.00150.00140.0014 -2.