SPC to the Overall Equipment Effectiveness
Six Sigma – iSixSigma › Forums › Old Forums › General › SPC to the Overall Equipment Effectiveness
 This topic has 8 replies, 6 voices, and was last updated 18 years, 11 months ago by Twan Kersten.

AuthorPosts

May 23, 2001 at 4:00 am #66720
Martinus,
OEE as you illustrated is a composite indicator. Calculating an average or range using all of these components together is not the correct way of handling these data. If you use the average and range OEE in the way suggested, what would be your action plan if it were to become unstable? With some thought you would recognize that you would still have to decompose the trends to identify which component is causing the problem. In an effort to simplify the reporting, you complicate the process of correction and/or improvement. If you would like to use this trending scheme for management fine. But, I would suggest you break each component of OEE out, and separately trend them. This way if a component drifts you will be able to catch it before the system becomes unstable.
Ken
0May 23, 2001 at 4:00 am #66721
Grant BlairParticipant@GrantBlair Include @GrantBlair in your post and this person will
be notified via email.From a practical standpoint, the only thing you need to plot is the per cent good product on the first pass. (Plot seconds and scrap seperately). This can be plotted either as a PChart, or as an IMR chart.
Get this number as near 100% as you can and the other measures will take care of themselves.
Technically, you are subgrouping related variables..won’t work…you will get too many special cause points to ever succeed in improving the process.0May 23, 2001 at 4:00 am #66722
Grant BlairParticipant@GrantBlair Include @GrantBlair in your post and this person will
be notified via email.My mistake…message should have read…Plot rework and
scrap separately.0May 23, 2001 at 4:00 am #27314
martinus todingParticipant@martinustoding Include @martinustoding in your post and this person will
be notified via email.OEE is a measurement of equipment effectiveness that creates capabilities in the firm which lead to a competitive advantage. It combines equipment availability, performance efficiency, and rate of quality products [Patterson et al.,1997]
Example: OEE = Availability (95%)* performance efficiency (95%)*rate of quality product (95%). So, OEE = 86% (0.95*0.95*0.95)
I need to control the OEE using the SPC. So,I use equipment availability as X1, performance efficiency as X2 and rate of quality product as X3 (n=3) to calculate RBar, XBar and process capability indices.Is this right or wrong?
0May 24, 2001 at 4:00 am #66728
martinus todingParticipant@martinustoding Include @martinustoding in your post and this person will
be notified via email.Ken,
This is an example:
(subgroup #1)Availability (.95), Efficiency (.92),and Quality (.90). The average and range would be 0,923 and 0,05.
(subgroup #2)Availability (.98), Efficiency (.92),and Quality (.88). The average and range would be 0,793 and 0,1.The average of subgroup #1 is greater than the average of subgroup #2 and range of subgroup #1 is smaller than subgroup #2. Of course Subgroup #1 is better than Subgroup #2.
We can predict using the XBar and Rbar Chart (n=3)that subgroup #2 may be out of control (unstable).
Now, we can break out each of the OEE component (subgroup#2) which make the OEE out of control (assignable causes must be there!).I think this would be effective.
Martinus
0May 24, 2001 at 4:00 am #66730
Joe PeritoParticipant@JoePerito Include @JoePerito in your post and this person will
be notified via email.Martinus, you can not use SPC to control the total OOE. The SPC may not let you know when the normal process OOE has changed to something other than what the normal OOE composite number looks like. One problem is that you would have to do further analyses to determine which one of the three factors changed the composite number. Also, it is unlikely that each one of the factors are exactly equal to the others. Therefore, they can not simply multiplied times each other. You would have to proportionately weight each one of the factors and then multiply them. Another problem is that one fact could go down and and a second factor could go up and the composite OOE would not change on the SPC chart. Using intergers, there could be 27 combinations of ratios that could end up with the same plot on the SPC chart. The number of combinations would expand exponentially with each decimal point in plotting percentages. Also, knowing how to weight the factors would be confusing. If the productivity fell off would this be do to maintenance problems, operator output, scheduling problems, lack of orders, hold ups from other departments, scrap rates, different set up times, and so on? I suggest using a line chart with coded or colored lines for each one of the factors. Then you would have to add explanations why any one of the factors improved or fell off.
0May 28, 2001 at 4:00 am #66759
Grant BlairParticipant@GrantBlair Include @GrantBlair in your post and this person will
be notified via email.Let’s suppose your equipment was operating at 100% efficiency and was available 100% of the time, but only produced 90% first quality product on the first pass through the system.
Now, explain to me how you plan to justify squandering 10% of your company’s time and resources by making offquality product. While you’re at it, please explain to me how you plan to either operate at 110% efficiency or make the equipment available 110% of the time to rework the 10% offquality product.
Now, if you’ll notice, being able to produce 100% first quality product solves all your problems. If this only requires 90% efficiency, I can idle the equipment 10% of the time. Or,I can schedule the eqipment for preventive maintenance 10% of the time. It’s your choice.0May 29, 2001 at 4:00 am #66769
Kim NilesParticipant@Kniles Include @Kniles in your post and this person will
be notified via email.Interesting subject but have a couple questions:
1 How are availability, efficiency, and quality determined? How often the machine is ready for use? How much throughput the machine had? the machine yield?
2 Patterson et al.,1997 was sited without a title. How can I learn more?KN
0January 2, 2003 at 4:07 pm #81811
Twan KerstenMember@TwanKersten Include @TwanKersten in your post and this person will
be notified via email.Dear KN,
To learn more about how to use, calculate and define the OEE visit http://www.oeetoolkit.com and download the OEE Industry Standard
Twan0 
AuthorPosts
The forum ‘General’ is closed to new topics and replies.