Machine Downtime improvements
Six Sigma – iSixSigma › Forums › Old Forums › General › Machine Downtime improvements
- This topic has 24 replies, 15 voices, and was last updated 17 years, 5 months ago by
Abdelaziz Moustafa.
-
AuthorPosts
-
November 9, 2004 at 1:22 pm #37495
I am currently putting together a green belt project charter for a maintenance manager. Both the maintenance manager and myself are in agreement that we need to reduce down time due to machine break downs through a series of improvement activities / projects. Unfortunately we are in disagreement over the level of improvement that we should be aiming for.
Does anyone have a example of world class benchmark for machine down time reduction due to break downs.
0November 9, 2004 at 2:31 pm #110496Machine break down, world class manufacturer ? these are pretty broad categories. What industry are you in?
RAC0November 9, 2004 at 5:38 pm #110503Here is some data for you, (US and UK)
Percentile 25 50 75 90 95
Equipment Availability % 78 84 92 94 98
Equipment Availability = actual uptime/required uptime
Also suggest you look at OEE as a measure.
Please bear in mind this is general industry….there are indexes to adjust for industry type.
Rich0November 9, 2004 at 5:40 pm #110504Plastic injection moulding.
Injection moulding range of 250 tonnes to 1500 tonnes.
Home storage solutions.0November 9, 2004 at 6:43 pm #110509As an engineer, I love when maintenance managers get involved in 6s!
Personnally though, I don’t like leading projects that “Reduce Mechanical downtime” as a metric, mainly because this is hard to measure, and the causes cannot always be controlled. A strong preventative maintenance program is the answer for that.
In our food processing facility, we run 24/7 and our mechanical downtime is at 1.5%. However, in a paper mill for example, downtime could be as high as 20% (what industry are you in?). Instead, maybe you could look at mean time between a specific type of failure, whose causes are known, as opposed to a general downtime %.
If you are looking for a good starting project for a maintenance manager, I recommend looking at time on tools for your tradespeople (reduce O/T by being more efficient) or “right-sizing” your spare parts inventories, or even looking at your shutdown schedules to see if you can decrease the duration and get the same amount of work completed.
Good luck,
Joe BB
0November 9, 2004 at 7:42 pm #110515
Mike CarnellParticipant@Mike-CarnellInclude @Mike-Carnell in your post and this person will
be notified via email.Pete,
If you take the metric began with at Motorola we had to get a 10X improvement every 2 years (approx 68% reduction per year). It worked well but as any glittering generalization it is easy in some situations (low capability) and difficult in others (high capability).
First you are doing your managers job. he should be writing the charter. Aside from that don’t worry about it until you get some data. Find some catagories lower than “reduce down time.” If you are at a lower level you will have a little more visibility of what is possible.
Just my opinion.
Good luck0November 10, 2004 at 6:26 am #110527
Chris SeiderParticipant@cseiderInclude @cseider in your post and this person will
be notified via email.Pete,
Do not let the others shy you away from addressing machine availability. For an extrustion facility, one of our machines has a lot of unplanned downtime. This was always tolerated until a little analysis showed a common failure for 40% of those unplanned occurrences. A team was launched and that 40% of the unplanned occurrences was reduced drastically–over 50%. However, we are not running into a proven raw material supplier causing unplanned downtime and are having to work strongly with the supplier to shape up or ship out.
A simple metric for machine downtime is the # of occurences of short stops or breakdown for a certain time frame. If you have some data collected already, look for commonality and tackle that first level pareto issue.
Good luck!0November 10, 2004 at 7:01 am #110530
GourishankarParticipant@GourishankarInclude @Gourishankar in your post and this person will
be notified via email.Pete
First you should freeze the metrics you would use to measure machine downtime . OEE ( overall equipment effectiveness ) is a good but broad indicator of the machine utilization % .
Two significant metrics you can use are MTBF ( mean time between failures ) and MTTR ( mean time to repair). While the first is an indication of the Effectiveness of the maintenance process ( preventive , predictive , breakdown processes ) , the second is a good measure of the Efficiency of the maintenance process . You should be targetting improvement in both areas.
The level of improvement can be decided once you establish a baseline sigma for the above metrics. Of course , the ideal target should be 6s!!
You could also consider “industry” benchmark data – it may not be wise to use data general downtime data as it varies from process to process.
0November 10, 2004 at 2:02 pm #110540Hi Gour….,
After giving more thought to Pete’s issue and considering your point. It would seem that MTTR and MTBF data is actually difficult to collect accurately. Easy to game, tremendous variability in data entry practices.
Would it make more sense to look at a proactive to reactive maintenace ratio, as a bellwether for maintenace. But of course there is another problem….minor stop off’s handled by operators often un- documented as to cause but usually recorded as downtime.
Therefore another Y could be a ratio of minor stoppage to downtime from all sources. My thinking is if MTBF and MTTR are improving then these two metrics would show it.
Food for thought.
Rich
0November 10, 2004 at 6:42 pm #110565
Mike CarnellParticipant@Mike-CarnellInclude @Mike-Carnell in your post and this person will
be notified via email.Gourishankar,
Tracking MTBF and MTTR without tracking Availability will lead you to missed opportunities. When you understand demand and Availability you can trade off Availability to work on MTBF.
Regardless of that these metrics are 30,000 foot metrics and not projects. These are Y’s and they are a function of specific issues around the machines. It may be a good spot to begin looking but these are not projects.
Just my opinion.
Good luck.0November 12, 2004 at 5:39 am #110657
GourishankarParticipant@GourishankarInclude @Gourishankar in your post and this person will
be notified via email.Rich
I agree with you. Data accuracy is an issue . I don’t think secondary data can be used for meaningful analysis.Maybe a data collection exercise needs to be carried out for sepcifically for the project.
Mike has made a good point on MTBF and MTTR , although I am not clear how Ys could be generated out of MTBF /MTTR data. I think at best you may be able to drill down the same to machine- specific or process – specific data .
0November 12, 2004 at 6:16 am #110665
GourishankarParticipant@GourishankarInclude @Gourishankar in your post and this person will
be notified via email.Mike
You’ve made a good point on the metrics. Can you suggest a typical Y ? I think a Pareto of MTBFs/MTTRs will yield machine – specific or process specific improvement opportunities that can be converted into Ys.
I have tried this in one situation and come up with significanlty high MTBFs and MTTRs for bearing – related maintenance. Two opportunities identified were ” x % reduction in bearing consumption” and “a% reduction in downtime due to bearing failures”. A second order Pareto will further indicate specific parts of the machine or process that will require improvement.
However , as mentioned in an earlier post , data accuracy for MTTRs and establishing baseline sigma are major issues.
0November 12, 2004 at 12:27 pm #110678
Mike CarnellParticipant@Mike-CarnellInclude @Mike-Carnell in your post and this person will
be notified via email.Gourishankar,
If you will reread your postagain you just provided your own example. MTBF was not he project. The projects were bearing consumption and bearing failures – the independent x’s (they may even become Y’s.
MTBF and MTTR will always be a Y. You don’t want to convert things to Y’s you want to convert them to x’s
Good luck
0November 18, 2004 at 3:45 pm #110926
PharmBBParticipant@PharmBBInclude @PharmBB in your post and this person will
be notified via email.I am just wondering if someone has statistics for Packaging changeover (set-up ) time for pharmaceutical industry , Also the packaging line downtime as industry world class performance.
Thanks
PharmBB0November 18, 2004 at 4:15 pm #110929
Vik SharmaMember@Vik-SharmaInclude @Vik-Sharma in your post and this person will
be notified via email.We have an excellent reporting software that was customized to capture key metrics for all the equipment at our facility. This reporting package records machine uptime, downtime, faults and key prodution counts. From recording all these measure it reports the following:
Machine Availability
Mean Time To Repair
Mean Time Between Failure
Overall Equipment Effectiveness
The vendor is Sidel, the application is called Efficiency Improvement Tool. It is developed by their group in Montreal, Canada. If you need to capture machine data they have the an excellent package. The only downside in that they do not currently have SPC capability built-in.
0November 18, 2004 at 4:45 pm #110930Pete:
I also work in a plastic injection molding plant (50 machines from 60 to 350 tons, PP water pipe fittings)
Here you have some benchmarks. source: http://www.immnet.com/articles?article=945
Best Regards
0November 18, 2004 at 7:13 pm #110939I would recommend to do a Kizen exercise and take note of activities/attitudes on more common activities where yuo have break down. Then modify this activitie involving opertors and maintenece people. This is faster.
adg.0November 18, 2004 at 8:23 pm #110944
Jonathon AndellParticipant@Jonathon-AndellInclude @Jonathon-Andell in your post and this person will
be notified via email.I did some work in injection molding too. Our shops tenbded to be lower tonnage, making multi-cavity small items. Here are some things to bear in mind as you move forward.It sounds like your management may be trying to “end world hunger” with a single project. Break out your down time categories. There is down time for change-over, for preventative maintenance, and for breakdowns. You can be maintaining either the press or the mold, so you want the breakdowns to reflect that.Once you start assigning down time to the respective categories, you can build a Pareto chart to see where your effort is spent best.If you need to speed changeovers, Shingo’s work on “Single Minute Exchange of Die” could provide some good insights.If you want to reduce breakdowns, you probably will need still deeper breakdowns of repair cause categories. Eventually, you will get specific: X hours of preventive mainetnance on such-and-such category will return Y hours of additional up time. Then you can see whether prevnetive maintenance can be built into other times when a unit is down – for instance, maybe you can maintain a press while a mold is being changed out for repair, and so on.Until you know what the “as-is” condition looks like, however, it is inappropriate to establish a goal. Each category, and each kind of break down, has its own range of as-is and could-be performance.Hope this helps!
0November 18, 2004 at 9:19 pm #110946Maybe the place to start your thinking is that the 6 sigma level with respect to a 24 X 7 X 365 operating machine schedule is about 107 seconds per year. Compute the six sigma level on your machines (99.99966%) using the best ideal operating schedule & compare with where you are. Then set the goal level at any future state sigma level you want to be. Evaluate your improvement investments accordingly.
0November 20, 2004 at 3:17 am #111024
Abdelaziz MoustafaParticipant@Abdelaziz-MoustafaInclude @Abdelaziz-Moustafa in your post and this person will
be notified via email.Hello
I would like to know how do you measure overall equipment availability…
OEE (i need to know the equation)
Thank you0November 20, 2004 at 3:21 am #111025
Abdelaziz MoustafaParticipant@Abdelaziz-MoustafaInclude @Abdelaziz-Moustafa in your post and this person will
be notified via email.Hello all
I have a project of machining process lost hours…can anyone recommends what data i can collect to identity the critical cause not satisfying my Y’s
Thank you
Abdelaziz0November 20, 2004 at 1:46 pm #111030
Vik SharmaMember@Vik-SharmaInclude @Vik-Sharma in your post and this person will
be notified via email.Check out the link below for a good explaination
http://www.bin95.com/Overall_Equipment_Effectiveness_OEE.htm
0November 24, 2004 at 2:48 pm #111249
SubramanianMember@SubramanianInclude @Subramanian in your post and this person will
be notified via email.machine downtime calculation method,
the calculation method and how to make graph.0November 24, 2004 at 2:48 pm #111251
SubramanianMember@SubramanianInclude @Subramanian in your post and this person will
be notified via email.machine downtime calculation method,
the calculation method and how to make graph.0December 18, 2004 at 6:29 am #112627
Abdelaziz MoustafaParticipant@Abdelaziz-MoustafaInclude @Abdelaziz-Moustafa in your post and this person will
be notified via email.Hello
I found that we lose 15% of the machine time every week in setup including the qulaity approval process and first piece approval….does anyone can give me an idea on how to collect data for Hypothesis, SPC, Multi var., and could help me in my DOE.
Thank you ALL0 -
AuthorPosts
The forum ‘General’ is closed to new topics and replies.