How to calculate OEE: the formula with a worked example

What OEE measures

A line that runs all day can still make far less than it could. It stops for half an hour for a changeover, runs slower because the labeler keeps jamming, and some of the bottles are rejected. OEE captures those three losses in one figure.

The figure is made up of three factors, each a percentage:

• Availability: which share of the planned time the line actually ran, so after subtracting downtime
• Performance: how fast the line ran compared with its ideal pace, so after subtracting speed loss and short stops
• Quality: which share of what the line made was right first time, so after subtracting scrap and rework

The formula step by step

Start with planned production time: the time the line was scheduled to produce. Breaks and periods without orders are left out; changeovers and breakdowns are included.

Then calculate the three factors and multiply them:

• Availability = run time / planned production time, where run time is planned production time minus all downtime
• Performance = (units made × ideal cycle time) / run time
• Quality = good units / units made
• OEE = availability × performance × quality

A worked example

Line 2 fills bottles in an eight hour shift, 480 minutes. There is a 30 minute break, so planned production time is 450 minutes. A changeover takes 35 minutes and a labeler fault 25 minutes. Run time is therefore 390 minutes and availability is 390 / 450 = 86.7%.

The ideal pace of the line is 60 bottles per minute. The shift made 19,500 bottles. At ideal pace that would have taken 19,500 / 60 = 325 minutes. Performance is 325 / 390 = 83.3%: the line lost 65 minutes to a slower pace and short stops.

Of the 19,500 bottles, 390 were rejected on fill weight, so 19,110 were good. Quality is 19,110 / 19,500 = 98.0%.

OEE is 0.867 × 0.833 × 0.980 = 70.8%. You can check it with the short route: 19,110 good bottles take 318.5 minutes at ideal pace, and 318.5 / 450 is also 70.8%. So of the 450 planned minutes, just over 318 produced good product.

Common OEE mistakes

The formula is simple. The mistakes are in what you count:

• Treating changeovers as planned time, which makes availability look better than it is
• Taking the nameplate speed of the machine as ideal when the line never reaches it, or choosing an ideal that is too low
• Accepting performance above 100%: it means the ideal pace is wrong
• Counting reworked product as good, although it was not right the first time
• Not recording short stops, so they disappear unnoticed into low performance
• Averaging the OEE of different lines instead of weighting by planned time

From OEE figure to action

A single OEE figure says little. It becomes useful when you see which factor is low and why: per line, per order and per shift, with the reason for every stop. Then you see, for example, that changeovers on line 1 cost more time than all breakdowns together.

Compare mainly with yourself: the same line, the same products, week on week. A line with many short runs will always have more changeover time than a line that runs the same product all week. Also look at the data itself: an OEE based on sheets filled in by hand at the end of the shift is less reliable than one recorded during production.

Where Frontlink fits

Frontlink measures speed, downtime and scrap per line, per order and per shift, and calculates OEE from them, live and after the fact. When a line stops, Frontlink asks the operator for a reason. For every figure you see which data it is based on and where data is missing.

You can start with operators recording on a tablet and connect machines later through the edge gateway. A Frontlink engineer builds it at your site, on your own data and ERP.