Select Page

## 20+ FAQ’s about the calculation of OEE

### 1. Isn't the OEE Calculation just a fruit salad of several numbers?

Q: OEE, Isn’t that just a fruit salad of several numbers, a complete bogus?

Arno Koch •     Is OEE fruit salad? No, it is a simple yet powerful balanced time study that helps shop floor people to visualize and eliminate equipment losses. Why is it not fruit salad?

Let me explain. First of all we have to make a clear distinction between EFFECTIVENESS and EFFICIENCY. In many discussions this- and other things unfortunately are being mixed up; I hope to get it clear with some examples.

## Effectiveness

Effectiveness is the relation between what theoretically could come OUT of a process and what really came out.

If your machine is capable of making 100 good units an hour, and it makes 60, it is 60% EFFECTIVE, but we do not know how EFFICIENT it was, because nothing is said about what we had to put in (how many operators, energy, raw material).

## Efficiency

So if a machine runs 50% effective with 1 operator and becomes 60% effective with 2 operators, the EFFECTIVENESS goes up 20% (yes, 60 is 20% more than 50…)  but it’s EFFICIENCY dropped down to 50% !

The same goes for YIELD. If you are making French Fries, all fries can be perfect, so the quality is 100%, but if you throw away halve the weight of the potato because you make the potato square first, your yield is only 50%; in this example you would be 100% effective but only 50% efficient.

Basically OEE is (as the name says) about EFFECTIVENESS; it is the rate between what a machine theoretically could produce and what it actually did. So the fastest way to calculate it is simple: If you take the theoretical maximum speed (i.e 10 units per minute) you know that at the end of a 480 minutes shift there should be 4800 units. Then count what’s on the pallet; if there are 2400 good units your EFFECTIVENESS was 50%, right?

## Where are the losses?

No rocket science so far… Now why bother to have availability, performance and quality in there? Well; every loss-buster wants to know where those OTHER 2400 units are gone, that are NOT on the pallet. Why are thy not on the pallet? (assuming we would have the customer waiting for the product, otherwise we could have stopped producing after half a shift).

Now OEE is going to track down where we lost it. And now we have to take care; every day where we run 50% OEE, we can lose units in a different way, and every loss pattern has its own cost structure. If we lose 2400 part because the machine ran beautiful smooth with no quality loss but at halve the maximum speed, that’s completely different as producing 4800 units at full speed, and then dumping 2400 out of specs parts.

Aha, so now it starts to make sense why OEE is not calculated the fast way, because you would have mist that tiny detail… (at least on paper, sure you would have seen this, but mostly it is more complex!).

So how do we find out what we lose when and how do we prevent to make a fruit salad of numbers?

Lets track down a normal day in our potato factory;

A shift takes 480 minutes.

## Availability

First of al we are going to find out whether the machine runs or not. One thing is for sure; if it does not run although it should, is simply can not be effective right?

Our operators take 10+10+30 minutes breaks, and do 2 setups of 25 minutes each; the rest of the time the machine runs. This means we lost  100 minutes and there is only 380 minutes left to be effective. Even if we run the rest of the time at full speed with no quality losses, we can never be more than  380/480= 79.2% effective at this shift. This ratio we call ‘availability’.

## Performance

Lets see how we spent those 79.2% of our time…

Let’s assume our potato cutter has an ideal cycle time of 1 second per cut; so it can make 60 cuts per minute.

This means in the resting 380 minutes the machine can make 380×60 cuts= 22800.

So if at the end of this shift the machine would have performed 22800 cuts, that would mean that during the time it was running, it performed at 100% speed. If it would be slower, let’s say the cycle time would become 2 seconds, it would slow down to halve the maximum speed, thus its performance would become 50%. The actual output now at 50% performance is 11400.

Running at 50% performance in this case would mean that we lose another (380 * 50% = 190 minutes.

If at this point all output would be in spec, what would be the effectiveness?

From the 480 minutes we lost 100 minutes in ‘not running’ and 190 minutes on ‘to slow cycle time’; so (480-(100+190))/480 = 39.6% so far.

## Quality

Whether this is the actual effectiveness depends on how many cuts where in spec.

If from the 11400 cuts, there were 2000 out of spec cuts, the quality rate of those cuts was (11400-2000)/11400 = 82.4%

## Combining Availability, Performance and Quality

In other words:

• we lost 100 minutes by not running, (=21%)
• from the resting 380 minutes we lost 190 minutes by slow running (=50%)
• from the remaining 190 minutes we lost 33 minutes making scrap (=17.4%)

Theoretically we could make 480 * 60 = 28800 cuts. At the end there where 9400 cuts, so the Overall Equipment Effectiveness was 9400 / 28800 = 32.6 %

As we did it here;

availability 79.2% * Performance 50% * Quality 82.4% = 32.6%

## Conclusion:

OEE is purely time based, but since 1 tact time equals 1 unit, OEE can be calculated partially in units for ease of use.

At many machines operators will not say ‘Today I ran at  0.027 seconds’ but ‘today I ran 2222 units per minute’ which is the same!

‘I stopped for 5 minutes’ is the same as ‘I lost 11111 potential units’

‘We changed the tact to 0,031’ means 2222 units à 1935 unit per minute. We are losing 287 units every minute, potentially 480 * 287 = 137760 units a day.

## Awareness

OEE helps to create this kind of awareness; with operators, with engineers, with logistic departments, and with anybody else involved in the value adding process. It gives a common language to everybody involved in manufacturing.

## Detailed Information

Now for those who like formulas instead of my Mickey Mouse explanation, please read

• TPM development program by Seiichi Nakajima, Productivity press (originally published by the Japanese Institute of Plant Maintenance)
• TPM for the Lean factory; Ken’ichi Sekine and Keisuke Arai, Productivity press.

I sincerely hope that this gives some basic understanding of OEE. I was by no means complete and we can have long discussions about the role of OEE in lean transformation and OEE versus cost price.

I studied the subject intensively and I can assure you: OEE is not the only- but the most beautiful and powerful KPI’s I have ever seen. Specially if you define it a little wider and tie the 100% time limited together with the actual time limit which is 24 hrs * 365 days a year. If you want to now how to do so, please read the OEE Industry Standard:

Https://oee.coach/oee-standard/

The relationship between efficiency, effectiveness and productivity

### 2. Is the average OEE of an average machine between 35-45%?

Q: You mention within your site the OEE of an average machine is between 35 and 45%. What data is this figure based upon?

Arno Koch •  It is based on my analyses of more than 1000 machines in all kind of processes and environments. I tried to set the definitions as close as possible to the OEE Industry Standard in order to eliminate large deviations based on differences in definitions. However; please ONLY use this as a ‘Rule of Thumb’. As I have to keep saying: it does not makes sense to compare OEE numbers. The only reason I mention those numbers is to make you aware that there might be a huge improvement potential in your factories that you might not have discovered yet!

### 3. What is the trend of OEE over time, say last 10 years?

Q: What is the trend of OEE over time, say last 10 years?

A: I have no hard evidence, but my impression is that the majority remains in the same old pattern. In some environments I see sometimes quite impressive improvements since TPM and Lean become more widely spread. To me it is clear: where TPM is seriously being implemented, OEE increases also seriously…!

On the other hand the real low performers seem to disappear slowly; this might be a ‘natural selection’ or due to the awareness of productivity in general.

### 4. Does the OEE for a machine vary from global region to region?

Q: Does the OEE for a machine vary from global region to region?

A: Again no hard prove, but countries that have embraced TPM-like improvement strategies seem to show more nice improvement examples than those who don’t. I think of Brazil and Germany. But be carefull… also in those countries such factories are still a minority…

### 5. Does the OEE figure vary by sector e.g food packaging to Automotive?

Q: Does the OEE figure vary by sector e.g food packaging to Automotive?

A: Definitely. The dynamics and specifics of a sector can be seen in the numbers, although I would not like to say this to encourage benchmarking! Typically in Pharmacy we see very low OEE’s and in the better Automotive or FMC’s plants substantially higher numbers can be found (of course there are reasons to that!)

### 6. How should I choose OEE software that does all the calculations for me?

Q: I have to compare different OEE Softwares. There is a wide range of suppliers. The software should be very simple. It should be able to collect the data manual, maybe with a bar-code scanner and later with a sensor. A significant data evaluation is important. Could you recommend some suppliers?

Arno Koch •    This question was asked more often (see also ‘LinkedIn OEE’ forum) however there is not just one answer possible. May I make a proposal and give you some guidelines?

## Proposal comparing OEE software:

After I wrote the first commercial OEE software in the 90’s, many different solutions are being developed, each with different approaches, different purposes, different solutions to all kind of issues. And of course different ‘tastes’.
The discussions about what software to use, mostly compare to ‘religious discussions’; few arguments and lots of ‘believes’.
So I feel users ánd suppliers would benefit from an overview of the different features and approaches in the different OEE solutions.

## Format to compare OEE Software

At this moment I am preparing a format to list and compare relevant issues (based on returning questions users ask) in the different software. It is not my intention to rank the systems; I truly believe there is room for different approaches to different situations.

So software providers interested in participating to setup a ‘listing format’ please contact me. The list will be published at www.oeeFoundation.com

## Guidelines choosing OEE software:

Ok, now some key points from my personal checklist when I need to implement software;

• How long does it take to get it technically running?
• How long does it take to configure it?
• Can it cope with the different situations/needs for definitions that I am facing?
• Does it support the OEE Industry Standard?
• How friendly is it to the user/operator? (check need for training)
• Who’s  solution will it be (is it IT owned, management owned or shopfloor owned)
• How deep are the possibilities to analyze losses? (beware of overkill or even non-information)
• How are losses visualized, will the user easily understand this?
• How easy is it to make quick reports/presentations?
• Does it generate the information I need when running a SGA/Kaizen event?
• What possibilities are there to visualize line/supply-chain issues?
• Does it also allow also efficiency parameters to be monitored?
• Does it give sufficient feedback to operators (ie in case of errors or when interaction is needed?)
• How difficult is it to tap into machine signals?
• Can it be used for manual data-collection?

There are more points to consider, however those point hopefully lead you to the solutions that would fit your need.
Good Luck, Arno

### 7. Can OEE be measured in renewable energy industry?

Q: Is the OEE used in the hydro electric industry? Or other renewable energy industry: Wind, Hydro, Solar, Biogaz, etc.? Or can it only be used in manufacturing industry?

## Yes, OEE can be measured in Renewable Energy Industry

Arno Koch •  OEE can be used on any type of equipment. It sure would be interesting to see it being implemented on equipment with variable performance!

## Examples requested!

I have not seen it before, any examples are welcome, since this is a crucial type of equipment!

### 8. Can OEE calculations be used in Oil and Gas industry?

Q: Can OEE be used or adapted for use in the oil and gas industry? Are there any concrete examples of how the metric can be used successfully, either off or on-shore?

Arno Koch • OEE can be used on ANY type of equipment for (industrial) conversion. So yes, also on drilling well’s, refining crude oil, cracking Nafta, extracting tar sands and other processes where OEE can be measured in Oil and Gas industry.

## What to do:

• Install OEE there where you want to put your focus for improvement activities. There is no use for OEE if you do not use its information to improve it…
• Do not think your equipment is so special it will need special treatment. It is as special as any other machine… If you go through the categories of losses to define as suggested in the OEE Industry Standard, you will notice that all the losses in your processes are mentioned.
• Use the data-collection (no matter how you do it) as a method to involve the people that can influence the effectiveness of the equipment. Make sure they feel ownership, otherwise OEE remains a managements toy and no improvement is to be expected.
• Do not try to be exactly precise. 80% correct will allow you to see what needs to be done, specially in the beginning. So all the sales-talk about how you really need to connect the whole world of computers together… do not take it too serious. Be pragmatic and get your crew involved, ask them what is a best plausible way of getting information about ´running´ and output.
• Just start, even if it is just on a piece of paper. You will be amazed what information you will get, assuming you defined OEE for optimal loss visualization.

A New Approach to Operational Excellence for Oil and Gas

### 9. Can I measure OEE for a machine producing batches?.

Q: Is it possible to use OEE in an intermittent production, when producing batches?

Arno Koch •     There are two problems in this question;

1. How do I measure an accurate OEE when running batches
2. Can I measure an OEE for each batch

I will answer the second question here;

## Can I measure an OEE for each batch

The major goal of OEE is to identify whenever the machine is not running smoothly causing losses. In the situation, when a machine has to be stopped regularly (even when there seem to be ‘plausible reasons’), it makes sense to identify and quantify such reasons. This might reveal a new perspective to the ‘accepted’ equipment losses.

Losses do not only occur while running batches. A major part of the losses might occur between the batches. So we want to see those losses.

• If you do not run between batches (ie because the line is not balanced) register “waiting for process X”.
• If buffers are full register “No output due to buffer full” etc.
• If demand is irregular register “waiting for production order”.

This will indicate where to focus your improvement efforts, being the main goal of OEE.

## What is the problem?

The problem would be: Are you going to assign such losses to a/the batch? Probably not since there is no correlation to that batch. So you would miss this loss…

So measuring OEE of batches only, could make you blind for a substantial part of the losses on the equipment. I would rather widen your timeframe where you measure OEE (in cases of long batch time). Instead of OEE per shift, consider an OEE per 24 Hr or per week. In this case you will keep all losses in the picture.

The question you should consider is this: If you measure OEE only during batch-time, what are you NOT measuring?

## What are we NOT measuring?

If the process is intermittent, I would like to know this! How often is it interrupted? How long? How much variance in the intervals? Here are many indicators; something strange is going on in the value stream.
Many batch processes run on capital equipment. Why is it not ‘batching’ all the time? Blinding out setups, preparation time, loading and unloading etc. etc. might bring you to wrong conclusions about what to improve in the equipment, the line or the value stream.

The point is: Beware of sub-optimization. While analyzing many hundreds pieces of equipment, I found out: Losses can be in complete other areas as usually searched for, or where you might expect them. The only way to find them is not to blind out any shelter where a loss might hide.

### 11. How to measure OEE in long-run milling process?

Q: We are tool & die maker for  the automotive industry. Overall each die will take 20 to 35 days (based on size) to complete. Depending on the shape (contour) of the die, the process is done manually or auto mode. The mode depend on our CAD/CAM capability to do the programming.

How do we create a correct OEE calculation in our milling process?

### Our situation

There is no handling or movement once the die is placed on the machine. Only tool change is done few times depending on the shape of the surface.

The ‘speed’ of the processing is determined by several factors namely material, shape of the die and tool bit used

Arno Koch •  In the ideal situation, the mill, where the die is mounted, would be milling 24 hrs each day non-stop, at maximum milling speed, given the material and the tool mounted that moment, and there would be no quality defects.

### Steps to consider

1. If the machine is not scheduled for a 24/7 operation, make sure to register the non-scheduled time and take it either in the OEE or at least visualize it in OOE or TEEP. (Your customer is waiting!)
2. If the process goes in a multitude of milling stages, the registration could be setup per shift of 8 hrs. If the milling-runs are very extensive, so the machine runs multiple shifts without a stop, it could be necessary to define a (very) long ‘shift’.
3. Now, when the machine is scheduled to run, register each instance when it is not milling. Since the throughput time is up to 35 days, there will be a substantial part of time been lost by idling of the machine. Handling like tool-change, measuring etc is all waiting time to the machine.
4. Whenever the machine is milling, determine the theoretical maximum milling speed for the tool and the material that is being milled at that point. (ie X grams/minute)
5. When a stage of the milling is done, determine the actual speed of the milling. This can be done by weighing the die, weighing the milled chips or calculating the milled space through the CAD software.  If this is not possible; try to make a rough estimation; that will give enough clues later on.
6. Whenever a milling-run resulted in an out-of-spec situation, the whole milling-run has to be seen as non-quality.

### What to look for

The ultimate number of your OEE/TEEP calculation is no so relevant; it will be very low anyway. The value of the measurement here would be to get an insight why it takes so long to make this die; where the losses are and thus where the potential points of improvement are.

Please let us know what your findings are?

### 12. How to measure OEE for Installations with parallel transformation stations?

Q: Our machine has multiple parallel transformation stations… (See picture). how do I proceed?

Arno Koch •   Some machines have one (or more) input-line(s). The Installation has more than one parallel transformation stations. The finished products leave the system through one (or more) output-line(s).

Examples:

• Barrel Filler (Beer)
• Foaming Installation (Refrigerators)

There are several possibilities for defining the OEE measurement:

1. The multiple workstations will be measured as a “Black-box”. Idling time and Failures of stations will be shown in Performance Rate (Minor Stoppages).
2. Each station will be measured separately.
3. Collecting running time of each workstation separately.

The first option is preferred, because of several benefits:

• Takes fewer registration time than the other options.
• Registration is easy to start.
• Shows sufficient information for improvement teams.

The consequence of this is that several losses of the separate stations will be reflected in a lower performance rate. If that is the case then you will have to go into that with a focused improvement team or set up a detailed registration of the separate sations. Be aware that the amount of effort stays in relation to the expected insight.

## When Minor stoppages are the biggest loss..

If the Minor Stoppages are the largest OEE loss, it will require a temporary detailed time registration of the time losses of each workstation. An improvement team can measure the detailed information during 1 or 2 shifts. The downtime reasons will be recorded very detailed. The improvement team will look for root causes, based on their knowledge and experience. During the improvement the OEE (and especially the Minor Stoppages) will be followed very accurate. After the improvement the Minor Stoppages shouldn’t be the largest loss anymore.

An alternative approach would be to measure óne transformation station and see this as a representative for the others. If this one has a lot of failures, it would be very strange if the others haven’t…

The rule of thumb is: OEE should give focus and direction for improvement

### 13. Can I calculate OEE for a Line or Cell?

Q: Can I calculate OEE for a Line or Cell?

Daniel Högfeldt •    A line or a cell with many operations has to be looked at as one unit. Ideally all machines should run at the same rate to get a balanced one-piece flow through the line. Usually that is not the case though. Usually one piece of equipment is slower than the other equipments making it the bottleneck of the line/cell. Let us say the line consist of two machines. The first machine has an ideal cycle time of 60 seconds and the second machine has an ideal cycle time of 30 seconds. If that is the case ideally the bottleneck would run with an OEE of 100% but the faster machine could then only run at 50% OEE, because it had to wait for pieces to be completed by the previous machine.

If they both were measured and an average was made for the line the OEE would be 75%. If the line is looked at as one piece of equipment and is only measured on the bottleneck then the line OEE would be 100%. If on the other hand the ideal cycle time of the fastest machine was used and the measurement still was made on the bottleneck the line OEE would be 50%.

To really expose the bottleneck in the line the best way would then be to measure the OEE on the bottleneck using the ideal cycle time of the fastest machine. When improvements have been made there could be a different machine that is the bottleneck and then the measurement should be made on that machine instead. That way the bottleneck always will be in focus for improvement.

The other way would be to use the ideal cycle time of the bottleneck when measuring the line OEE. That would show more accurately how the line is performing based on what it is capable of performing. If data is collected correctly at the bottleneck it should show if the machine is spending much time waiting for pieces or waiting to send pieces. If it spends a lot of time waiting for pieces some machines upstream is the current bottleneck. Then focus has to be to fixing that problem. If the machine spends a lot of time waiting to send pieces the current bottleneck would be somewhere downstream. Then focus has to be there for improvements.

Master’s Thesis: PLANT EFFICIENCY A value stream mapping and overall equipment effectiveness study

### 14. Is it possible to use the OEE calculation for cellular manufacturing

Q: Is it possible to use the OEE calculation for cellular manufacturing or is it only applicable for an assembly line? The productivity of the cells depends a lot on the number of workers. But it is always underlined that OEE does not measure the persons, just the maschine.

Arno Koch • In principle OEE can be used to measure the effectiveness of ANY –machine related- conversion process. I am a little hesitant when I say ‘machine related’. It can also be used to find the losses in manual assembly processes or in surgery. When determining the theoretical maximum speed of a machine you will search for its physical limitations which leads to the theoretical maximum. In non-machine related processes this is more difficult and may lead to unwanted side effects. I.e. what is the maximum speed of a surgeon…?
When talking about a cell I assume there will be more than one machine, so the question is; which one- or how many you want to measure. There are several strategies possible to make the losses of the cell visible.

## Effectiveness – Efficiency – Productivity

The next point is that OEE measures EFFECTIVENESS (the amount of output realized with a certain input. It does NOT make any statements about the EFFICIENCY (the amount of input to realize a certain output. So if your effectiveness goes up by using more workers, your efficiency goes down. Now you have to determine what bottom line the effect will be (this is your PRODUCTIVITY)

## Measuring machines, not people.

Your last sentence is crucial; I am happy we agree on this. This perspective may give you many answers: if you put yourselves in the shoes of the machine and ask yourselves: What is happening to me (the machine)? What would be best for me and the process I am part of? You may see things you missed before! Just try it!

### 15. Is there something like 'Batch Effectiveness'?

Q: What is Overall Batch Effectiveness?

Dennis McCarthy •     The logic of batch OEE is to measure the flow of material through the end to end process stages rather than each asset individually.  Collate the data by asset and you can identify the losses at each asset.  You don’t need it where you have continuous flow but where there are assets that are not used all the time, you can quantify what impact that asset had on the value stream.  I hope that this helps

Arno Koch •    Ok, so I understand you ‘sit’ on the batch and see what is happening to it, right? Might be called “overall Batch effectiveness” or even better “Overall Valuestream Effectiveness”.
I see the charm in this approach; it gives a ‘number’ to the performance of a value stream (however it is quite far from the concept of OEE).
Now how do you determine the theoretical maximum speed for the batch, considering it moves over equipment with different maximum speeds?
And how do I see in this approach how a 60 Mio € machine has to wait all the time because it is in a non-balanced line?

Dennis McCarthy •      Effectiveness is a measure of how well you did what you planned to do.  So overall batch effectiveness (great idea) is about how well the batch went through the system as designed.  We apply this OBE approach to systems such as logistics or stores operation by considering Availability losses as things which stop you from doing what you plan to do.  Performance as things that take longer than expected and quality losses as non optimum output quality.  If you look at OEE the six hidden losses can be categorised as 1. planned or inbuilt losses such as (set up, reduced speed and start up losses) and unplanned losses (breakdowns, idling and minor stops and rework).  These are 6 different types of problems with 5 different focussed improvement tactics.  These tactics can be applied to any system.  The question of line balance is not something that OEE is designed to measure but because OBE records time the batch spends waiting for processing it also highlights the out of balance condition because there is no waiting time at assets with capacity and lots of waiting time at bottleneck assets.  The approach works fine.  Its not right for every cirumstance but I hope that you can see how Batch OEE or OBE doesn’t create blind spots, it was designed to shine a light on losses that would otherwise remain hidden.  Again I hope that this helps

Dennis McCarthy •      Sorry one more thing.  In terms of maximum throughput calculate the maximum output at each stage of the process then you can see how the batch performed against it.  An alternative is to treat the system as a black box which as with any line OEE will be determined by the pacemaker unit.

Arno Koch •     Dennis, it’s an interesting discussion that shows how different perspectives give different views! I fully agree that your ‘OBE’ approach is absolutely a valuable measure to identify losses in a value stream. I would encourage measuring this in ANY process, since it might reveal many obstacles that are taken granted for.
OEE however, as you also mentioned, has a different focus. It looks at one point, at one machine in the value stream. While looking there you usually get tons of hints about problems in the value stream or in the organization around the machine (assuming you defined the parameters smart), but still it looks at one fixed point. This has its advantages and disadvantages.
If we look at OEE in its purest form, it defines 100% OEE as what the machine could do at its THEORETICAL maximum. Of course there is a grey zone that might be discussed. However if the goal is to reveal ALL POTENTIAL losses, it has to be taken literally. So OEE does not measure against what you planned to do but against what you could have done in the theoretical/hypothetical maximum (even if at this point you see no way how to get there; that is the challenge for many SGA’s); this is a fundamental distinction from OEE and most other measures!
At this point one might call me an extremist… If I measure OEE of a machine in a chain of machines, most people take the maximum speed of the slowest machine as the maximum of the whole chain; thus “hiding” the capacity of all other -faster- machines. I would do the opposite: I take the maximum speed of the fastest machine in the chain as the maximum of the line. Now everybody is being challenged to find out what could be done to get there, knowing that maybe we will never achieve it… Hey, what the heck; it was and is never the goal to reach 100% OEE (and even 85% OEE might be complete nonsense in many situations, despite the persistent idea this would be World Class…)
Yet I know out there due to this approach now there are many lines that run faster than the original slowest machine ever did at its maximum! This is a phenomenon seen in every factory: look at the elder machines: even without TPM/LEAN or whatever; many of them already run far over the original design speed. Point is: Maximum speed is not written in stone, but you need to challenge it to stretch it!
So hip hip hurray for the new measure: OBE

### 16. How to calculate OEE with mixed output?

Q: Our machine can run mixed output, so several different products, each shift. Till now we just recorded the time the machine was running and the output we had. Our software now requires that we register activities and quantities to account for every minute in a shift. But we are not tracking the number of minutes it ran each product.

Arno Koch •    NOT your software requires this! It is required to get an accurate OEE calculation! And thus that is how it has to be done in the software…

## Running Mixed Output

If you run mixed output, so different products each shift, you might face up with products having different maximum speeds within that shift. So to know how much product you could have made (the target for a production run), you need to know how many minutes the machine ran at what maximum speed. By calculating the target output for a run, and comparing it to the actual output for that run, you are calculating a performance rate for this particular run. If you do this for every run in a shift, the performance rate is a weighted average of multiple runs (no matter the different maximum speeds per product!).

If you would just add all output and all production time, you lose the weight-factor! It LOOKS as if you get a correct OEE calculation, but it isn’t. Many of the home build systems I saw had this algorithm error.

### Example

Let’s take an example: a moulding-press. It is being operated with two different dies: one die for a large part, making one part each stroke and one die for a small part making 4 similar parts each stroke.

The press can make two strokes a minute. So the standard for the large part is 2 parts per minute and for the small part it is 8 parts per minute (2 strokes x 4 parts) Just for the example we assume changeover is timeless and everything else is perfect.

Let me show you two different days;

First in the way as you describe:
‘… we just recorded the time it was running and the output we had.’

 Day 1 (480 min) Day 2 (480 min) Time Running 480 min 480 min Output 2340 1260

Obviously day 1 was your best day.
Now, can you tell me what your OEE was on day 1 and on day 2?

Lets see how you can solve this (and how good software should do…)

 Day 1 Day 2 Minutes Part Large 120 min 360 min Max. Output Part Large (120×2=) 240 parts (360×2=) 720 parts Actual Output Part Large 180 parts 540 parts Minutes Part Small 360 min 120 min Max. Output part Small (360×8=) 2880 parts (120×8=) 960 parts Actual Output Part Small 2160 parts 720 parts

What was your best day?

 Day 1 Day 2 Availability 100% 100% Performance 75% 75% Quality 100% 100% OEE 75% 75%

The point is: To calculate the performance rate, it is necessary to know what the EXPECTED output was versus the ACTUAL output. In a mixed production this means you’ll have to know how long you ran what product since that is what determines the expected output.

Let’s make it a little more difficult:

### Parallel Output

Some machines, like presses, can produce different products in one stroke.
Imagine a press that can have two different dies in one stroke (so you could have a changeover of half a press…)
So in one shift the press can first produce 1 A and 4 B parts, after that 2 C and 3 D parts
and next run 2 C and 8 E parts each stroke. Dazzling???

Your software should be able to handle this situation seamlessly due to the mechanism as described above!

Q: We have a similar machining environment and what we do is the following:
If we have 4 variants a-d running at ideal cycle times x1-x4, planned for quantities a1-a4, and in reality what got produced was b1-b4, then performance is calculated in terms of total time as

(b1x1 + b2x2 ….) / (a1x1 + a2x2 …)

Thus it kind of gives a weighted indication of what the performance is, without skewing it like an average.
Awaiting your opinion on this, and whether there is a more accurate method of doing it.

Arno Koch •    Indeed, the way to calculate a correct performance rate in a mixed output situation is to calculate a weighted average expressed in time.

### 17. How to calculate OEE over multiple shifts?

Q: My boss wants me to calculate and present a graphical OEE representation per week or month. So we need to calculate an average OEE over multiple shifts. I present this as I do with the daily data. I calculate the monthly OEE by taking the average of the daily OEE’s.

Is it alright to present a monthly OEE of the machines by calculating the average of several daily OEE’s in a month?

Arno Koch •  No. Taking 20 OEE numbers, adding them and dividing by 20 does NOT give a correct OEE number. To calculate an OEE over multiple shifts you proceed as follows:

## OEE is a Weighted Average

Since OEE is a weighted average, it is not possible to just average multiple OEE numbers. To calculate OEE over a longer span of time you will need to recalculate the whole equation as if it was one OEE.

## How to calculate OEE over several shifts?

So calculate the availability over the whole time-span, the performance, the quality and (re)calculate the OEE.

Unfortunately not all OEE software does this correct, leading to wrong conclusions.

### 18. Can I make an hourly OEE Calculation?

Q: Suppose I want to calculate OEE continuously during the running shift, in order to present the team an ongoing status, how should I calculate this, specially when the machine is not running? Does the OEE drop to zero?

What if the Availability has been 50% for last hour but the actual production has been zero? What do you do with the Performance number? Should it be equal to zero?

Arno Koch •  OEE can be used for several purposes; although steering improvement is the core, it can also come in handy at forecasting the potential capacity or to show the progress on the line, assuming you will use the number rather as a compass than a scientific precise guideline.

OEE, by definition, is always related to a certain time-frame; usually a shift (ie 510 min), but it also can be calculated for a day, a week or any other time-frame.

If you want to present an ongoing OEE during a shift, there are basically 3 ways to proceed:

1. Using a fixed time-frame of let’s say 1 hour: Calculating a ‘fresh’ OEE every hour; you would see 24 OEE’s during the day. So if there was no output that hour, there was no availability, so the OEE would be zero. This answers the second part of the question: You say the actual production was zero, so how can there be an availability of 50%? This, by definition, can not happen. (There is availability when there is output, regardless the speed and quality.)
2. Expanding the time-frame as time elapses: Whenever you (re)calculate the OEE, take the elapsed amount of time as 100% (meaning after 47 minutes you will calculate the theoretical output for 47 minutes versus the actual output in those 47 minutes and so on). So if you start with a running machine, the OEE will start immediately high. OEE will drop down when the machine idles and climb again when it restarts running.
3. Take the total shift duration as time-frame. At the start of the shift, the theoretical output for that shift is calculated. This can be done for every minute in the shift. You would see a growing ‘reference’-line. At the start there is no actual output so OEE is zero. Whenever there is output OEE will climb a bit, until the shift is done. The operator will see his OEE grow. He can now see his actual output against the reference line showing where he could have been when there would have been no losses.

## What could be wrong OEE goes over 100%

We came up with the following reasons why this could happen

1. Our time normes are not ok and our people can reach very quickly speeds over 100%
2. We have 0 ppm and 0 bad parts, but I don’t think this is true; I even believe this cannot be reached even in a fully automated process.
3. We didn’t stop for maintenance of the machine – but on the long term I don’t think that this is sustainable.

Are we correct?

Arno Koch •  Indeed, several possibilities can lead to an OEE going over 100%. if OEE goes over 100% the definitions should be carefully checked since with correct definitions (and without seriously re-engineered machine or product) this is not possible.

### What if we don’t stop for maintenance?

Not stopping for maintenance can not be the cause for this phenomenon since in that case availability would go up, maybe even until 100%, but it never can go over 100% since you can not put more time in a shift than 100%.

### Not reporting defects?

Not reporting defects does not influence the performance rate since this -by definition- contains good and rejected product. It would raise the quality rate, but you can not make more than 100% quality.

### Keep running during breaks?

Cheating on run time (you say you where running during the break but didn’t) does raise availability but equally decreases performance.

### “hiding time”?

“Hiding” time (you where running but did not report it) could bring up the phenomenon, but why would anyone do this regularly?

## How can OEE go over 100%?

The way to achieve this phenomenon is to produce more output than possible according to the definition; so in cases where the defined maximum speed is lower than the real maximum speed.

This is the reason why the THEORETICAL maximum speed should be defined as the 100% value of performance. Do not use some ‘practical or actual maximum’ that is just practical for the moment but can be improved in time. If the machine CAN go faster it WILL go faster one moment or the other.

## The golden rule for Maximum Speed in OEE:

The real maximum speed can only be improved by re-engineering the machine or product.

Also see: “Can performance go over 100%?”

### 20. How to calculate OEE in a 24x7 operation?

Q: How can I measure the OEE on machines that works in a  24×7 operation?  Or are there any other indicators that you recommend?

Arno Koch •   Ask your selves: Who is the ‘psychological’ owner of this machine? How do you want to focus this/those people on the losses of this machine? If you have an answer to those questions, you will be able to define the time-frames that OEE will be registered- and reported. Most probably you will talk to the people of each shift. At that point there is barely a difference between a 1, 2, 3 or 5 shift operation. Does this help to clarify your issue?

### 21.What is 'OEE Top'?

Q: The Name Plate Capacity (NPC) is the maximum speed of the machine to run its ideal product. And OEE is calculated based upon different theoretical maximum speeds for different products running on that machine. Now what is OEE Top?

Arno Koch •    You describe it right. There is only ONE NPC and there can be multiple Standards (Maximum Speeds) to accommodate the speed differences for different products.

The ideal product for a machine, from perspective of effectiveness, would be a product running at the machines maximum speed.

## Product Allocation

Not everyone can afford a different machine for each product, and thus we allocate also products to a machine when this is maybe not the ideal combination. So we accept a certain loss in order to be able to run that product.

### OEE Top

This is where the OEE Top value comes in. OEE Top calculates the performance rate ALWAYS based upon the NPC. So also for products that run slower and have a lower Standard.

If we can run product B at only halve the speed of product A, the performance of product B would drop to 50%.

### Allocate product at more suitable machine

Looking at the cutter example, this indicates you might want to move this product to a heavier cutter that can cut this product in one stroke. This also works in the opposite direction. Imagine there is a product C, half the thickness of product A. What would be the standard?
I would say it has a standard of 120 because the cutter could cut two of these pieces in one stroke.

### determining the NPC and the maximum speed

In a technical sense, determining the NPC and the maximum speed for each product on a machine is probably the most difficult part of defining your OEE settings. Always make sure you look for the theoretical maximum because the goal of OEE is to make losses, and specially those hidden losses, visible. Making them visible is the first step on overcoming the losses! Hiding them away between excuses like ‘this is not realistic’ closes the road for improvement…