Downtime Calculation? 2

[SteveMeisel]

We have a palletized assy line with 13 Stations on it. Only 3 of the stations require operators. If one of the stations goes down they (production dept) count it as line downtime. In reality it is only 1 out of 13 stations that has stopped the others have not quit working. Most often it takes less than 2-3 minutes to get it back running and only one operator has to get up to fix the problem. Is there a formula for calculating the downtime if only one station has stopped?

 

[SteveMeisel]

Let me be more specific about what I am wanting:

I am not concerned with quality/repairs/changeovers only downtime.

Example:
The assembly line runs 8 hours.
A part is assembled every 30 seconds(for our example actually the line assembles a part every 26.2 seconds)

The assembly line has 13 stations on it.
Station 11 stops for 2 minutes.
During that time every other station is still working. No station has run out work. There are still parts comming of the end of the line with out any delay.

At the end of the day we have 860 parts

(8 hours - 30 min for lunch and 20 min for breaks = 8hrs*60sec=480min-50min=430min*2parts/min=860parts)

Overall we have had no downtime, but Station 11 has had 2 minutes of downtime.

The assy line can only physically make 1 part every 30 sec and if you do the math there could not possibly have been any downtime that day. But the production dept is claiming that they had 2 minutes of downtime on the line. Which from the math is physically impossible.

Is there a way to figure the line downtime from this example. This has caused many heated arguements between us and the production depts.

Steve

 

[uppili11]

you have got a most efficient line and much worried about 2 minutes!!!.neverthless you will have some problem only if you move to single piece flow with perfect line balancing only then the two minutes becomes very important otherwise your overall line efficiency is the best figure to control.please convince your production dept.(another way could be to start thinking in terms of OEE since nowadays the maintenence has become every body's reponsibility and not only that of maintenance dept.

 

[tomwalz]

At a guess, the production line is staging parts. Is there a “cushion” of parts before each machine?

Goldratt has written some good stuff about this.

Tom

 

[PSE]

Would you still have produced 860 parts even if station 11 did not have any down time or would you have been capable of 864 parts? You could look at the following:

Loss of additional throughput (4 parts)* part value (cost) + 2 min of "lost labor" * labor & burden rate.

This should give you a rough idea as to the expense of the station being offline.

You did mention that there was no delay in parts coming off the line. Assuming that there is a queue of parts for downstream operations, are the other stations capable of keeping the flow? It almost appears that station 11 is redundant and unnecessary to run (full time). If so, you may have an opportunity to reduce cost by purposely shutting it down at times.

Regards

 

[SteveMeisel]

Each station usually has a queue of 2 to 3 parts. The problem is not with the 2 minutes at station 11 but it is the 2 minutes at stn 11 + 2 minutes at stn 9 + 5 min at stn 3.
Our slowest is stn 11 but if stn 3 was down for 5 min there still would never be a delay at the end of the line where they are waiting for completed parts.

At the end of the shift they add up all the accumulated station downtime and that is the total overall downtime of the line.

uppili11 - please explain more about OEE.

PSE - When a stn has completed all of it's work it just sits and waits for another pallet to arrive on the coveyor then it starts in operation.
Often we take advantage of this time to do a minor adjustment. But as soon as the stn is taken offline that is counted as more downtime. When in reallity we haven't interupted the flow of parts.

 

[automatic2]

Make sure the downtime is documented. Don't worry about the descriptors used, as long as everyone understands and there is no misrepensentation of facts. Should the problem grow and affect output, then it would be nice to have a history. A meaningful stat that you could incorporate would be actual output over target, with a breakdown hinderance of 3 min. Starts to sound better as your hinderance time increases and your output ratio remains. On the other hand;

If a breakdown time of 2-3 minutes does not affect output rate, then your line efficiencies are less than maximum. If you where asked to increase the lines throughput, documented knowledge of these breakdowns would contribute to realistic values.

 

[uppili11]

OEE:

Actually this is overall equipment effectiveness.

This is a measure of equipments performance(A) ,the down time(B) and the quality(C)
1.0 PERFORMANCE(A)
It is easy to understand this with an example:
Equipment “A” rated output is 1000 parts per hour. The available time is say 8hrs per shift. This equipment is suppose to produce 8000 parts in the shift.
But due to downtime then available time could be 7 hrs and in 7 hrs if that equipment produces the rated output as 7000 parts the performance part of the equipment is 100%.

Let us fix some value: rated output:1000 per hour
Available time :8hrs(NO DOWNTIME)
Suppose to produce:8000
Actual production:7000

Performance of the equipment = 87.5%(A)

2.0 DOWN TIMEB)

The available hours per shift is 8 hrs.
10minutes machine breakdown and 20 minutes for change over-total down time is 30minutes

Down time in % 6.25% and available time is 93.75%(B)


3.0 QUALITY(C)

REJECTED PARTS PERCENTAGE :SAY 1%
OK PARTS =99% (C)

THE OEE IS A*B*C = 87.5% * 93.75%* 99%=81.2%


Normally for continuous process industries the value of OEE should be above 95% and in batch processes since there will be some inventory the recommended /BENCHMARK will be above 85%.

Hope it is clear to some extent.

It is not enough if you conider only downtime.That is the reason why people are moving towards TPM and a measure of that is OEE.(MANDATORY FOR QS9000 COMPANIES TO CONTINOUSLY IMPROVE ON THIS)

 

[BML]

Steve,

I'm curious to know what the run times of each station are. You may find that you can reduce overall costs and still get the same output by combining operations so that each station takes more time, but run two parallel lines. From what you describe, if station 3 is 5 min and stations 9 and 11 are 2 min each, then you are losing 3 mintues at both of those stations for every single part you make. That adds up to quite a bit more lost profit that 2 minutes of "downtime" in the middle of a run.

BML

 

[uppili11]

Yes .but it is very difficult to have absolute line balancing.each machine will not produce at the same rate.how effectively you design the line determines the thruput as well as eficiency.(synchronous line balancing)taking a cue from TOC the output of the line is always determined by the bottleneck capacity or the weak link.so one has to continuously work on strengthening the weak link and the link extends beyond manufacturing once the manufacturing is perfect and not a bottleneck.since nowadays capacity is not a constraint many people are working towards TAKT time and forming cells .when there is good demand the cell will be fully manned otherwise one person will handle many processes.

 

[griffengm]

Steve,
You might look at it this way.
Each time someone from maintenance has to come to the line, it increased the cost of the product being made there. It may not stop the line but it adds cost to the overall operating cost of the line. Thus higher cost product.
If several lines opperate this way then it means you have an extra maintenance person on staff to care for these machines.
If you must keep extra product in the pipeline to cover for the eventual breakdown, that is also a cost. At the end of the run, what does an operator at the beginning of the line do while the queue is being finished? Are they able to start changeover immediately or must they wait until the last part is run? If they must wait, then there is a non-value-added cost associated with that wait. It adds nothing to the value of the product but does increase the cost.
Also by maintaining a queue, you are using space simply for storage - another non-value-added cost.
One way or another, you are paying a cost for that breakdown. It may not be as obvious as one less part off the line for the day but it is there nonetheless.

 

[SteveMeisel]

Sorry for not replying sooner. Our line runs 2 products Part A which is our primary product and Part B which we only run 2 or 3 times a year for 1 or 2 shifts. So change overs are few and far between.

As for cycle times we have:

Stn 1: 22.1 sec
Stn 2: 8.2
Stn 3: 20.2
Stn 4: Not Used For Part A
Stn 5: 16.3
Stn 5a: 12.3
Stn 6: 17.1
Stn 7: 15.5
Stn 8: 24.4
Stn 9a: 16.9
Stn 9b: 20.8
Stn 10: Not Used For Part A
Stn 11: 51.9 per part/2 parts at a time/26.6 avg/10 pcs
Stn 12: Not Used For Part A
Stn 13: 23.9

 

[Rentapen]

SteveMeisel,
On your pallet race track line, do you have the old style bolt to the floor type stations or the newer on rollers bolt together station pallet race track (management loves the new lines, ask Flint Mich., you can pull a whole line in one weekend and move to a different location). Sometimes with the newer line if you have a alternate assembly profile you can Y or split an unfinished part to a new straight line for final completion. as for the calc's, your not going to get The big Um's or the little's "delphi, etc" to change the program, just learn it and adapt to your advantage.
Just wandering
rentapen

 

[SteveMeisel]

Good idea on the moveable stns. All of ours are bolted to the floor. Sure makes a lot of sence if you do a lot of rearanging. Thanks.

 

[uppili11]

If anybody has gone thru the PPAP requirement of QS9000 the line or stations on wheels ha become an accepted practice and does not even warrant a PPAP SUBMISSION.It is very common in automotive industries which practices LEAN majority of the equipments will be on wheels and it does not require a week end but within 2 hrs a new line could be assembled.now in our industry we started talking of SMEL !!!(SINGLE MINUTE EXCHANGE OF LINE)even the most complicated epoxy coating machine itself is on wheels.

 

[Rentapen]

uppili11,
I just reviewed a tack-spot weld company in Blissfield Mich. do does work for all the Big Ums, they just patented a quick change weld process. You buy one large stationary machine (beer cooler size) then you buy a bunch of smaller forklift loaded pallets (large tv box size) with your amp , air, etc plugs sticking out the back. They can change over as fast as the forklift operator can unhook two de-sta-co's pull the pallet out, insert new pallet, latch de-sta-co's, walk to back of machine and punch program #2 button, insert small metal piece that needs tacked or weldnut and bang, he's done. took less then a minute! This lean changeover can also be applied to a straight or race track palleted line. Way of the future. Buy one press like machine and lots of smaller pallets, no more turnkey machines sitting in the backyard after a run.
just wandering
rentapen

 

[uppili11]

yes it is interesting.If you have time go thru the DENSO technology 2000/2001 WEB PAGES.The robots are made in a way for a quick chageover for jobs say from making car starter to truck starter which is amazing!!!against the conventional robots fixed onto specific welding or painting

 

[conduit]

Your problem: 13 stations palletizing line
3 stations are manned
1 station goes down

Do you consider line downtime or internal time downtime?

Downtime for station: 2 or 3 minutes. One operator addresses the problem.

Magic formula I don't have. However, if this is a progressive line. In other words it takes all thirteen station to come up with one finished product. Then production is right in recording it as line downtime. You should have a line rate based on the bottleneck station. The only time when the downtime is internal to the established rate is if the station that goes down is not the bottleneck station or in other words the station with the longest dwell time or build time.

 

[griffengm]

If the operator did not need to stay close to attend to stations that fail, could his time be used in a more productive way somewhere else? If the answer is "yes", then there is a cost to the down time and in turn, the product. It does not matter whether it is for a critical station or not. If it requires an input, be it maintenance or operator, it adds cost.

 

[SteveMeisel]

Thankyou all for your views on this posting. It has shed a lot of light on the situation. I have started collection data with the plc on why each station stops. This will shed more light on the situation. And will allow me to make a permament fix on some the nusiance breakdowns. Hopefully eliminate the short breakdowns that have caused us to bump heads. Once again thanks for your input.

Steve