Glotech.io
What is OEE and how is it calculated?
OEE stands for Overall Equipment Effectiveness. It shows how much of a machine's planned production time was actually spent making good parts, and it is calculated by multiplying three rates: Availability, Performance and Quality.
The idea comes from Total Productive Maintenance (TPM), developed in Japan. Today it is widely used to track efficiency by machine and by line.
The three rates in OEE
Each of the three rates shows a different loss.
Availability shows how much of the planned time the machine actually ran. Breakdowns, setups and changeovers, and waiting for material all lower it. It is run time divided by planned production time. Whether breaks or planned maintenance count as planned time is your decision; what matters is that it stays the same from one calculation to the next.
Performance shows how fast the machine produced while it was running. Multiply the parts produced by the ideal cycle time, then divide by run time. It drops when the machine runs slowly or has short stops that nobody writes down.
Quality shows how many of the parts produced were good. It is good parts divided by total parts.
A worked example for one shift
The figures below are an example chosen only to show the calculation.
Say a machine has 480 minutes of planned production time. During the shift it stops for 80 minutes in total for a breakdown and a setup, so it runs for 400 minutes. In that time it makes 720 parts with an ideal cycle time of 30 seconds, and 684 of them are good.
| Rate | Calculation | Result |
|---|---|---|
| Availability | 400 / 480 | 83.3% |
| Performance | (720 × 0.5 min) / 400 | 90% |
| Quality | 684 / 720 | 95% |
| OEE | 0.833 × 0.90 × 0.95 | about 71% |
There is a quick way to check the result: multiply good parts by the ideal cycle time and divide by planned time. Here, 684 × 0.5 / 480 also comes to about 71%. In other words, the good parts from the shift are what the machine could have made at ideal speed in about 71% of the planned time.
Where the numbers come from
The formula fits in a few spreadsheet cells. Whether the result means anything depends on the numbers that go into it.
Part and scrap counts usually come from a counter, a production form or the ERP. Downtime usually comes from forms that operators fill in. Short stops may never get written down, and some entries are filled in after the job from memory. In that case Availability is an estimate too. Because the three rates are multiplied, that estimate carries through to the whole OEE figure.
Measuring Availability from the tower light
Whether a machine is running already shows on its tower light. Kaya reads that light through the cameras already in your factory and records how long each machine ran and how long it stood still. No sensors are fitted to the machine, and operators don't enter anything. If your cameras can't see the lights clearly, you may need to change a camera's angle or add a camera.
When the light is amber, Kaya also checks whether anyone is at the machine, so time with someone at the machine is reported separately from time the machine waited with nobody there. If a camera can't see the machine for a while, Kaya doesn't fill that time with a guess; it records it as "no evidence". If an hour looks wrong to you, open the camera frames behind it and check it yourself.
Kaya measures the runtime and downtime that the Availability part of OEE needs. Performance needs part counts and an ideal cycle time, and Quality needs good and scrap counts. Kaya is not a complete OEE system on its own, but it can be integrated with OEE systems.
The first 30 days are free
You can try Kaya free for 30 days on your existing cameras. It measures how long your machines run and stop, and at the end we go through the results with you. The Kaya page lists where Kaya isn't a fit, and other questions are answered in the FAQ.
Do your machines have tower lights? Can your cameras see them?
Write to info@glotech.io.