Automation¶
OEE → Configuration → Automation
Everything in OEE can be logged by hand, and nothing on this page changes that. What it does is let Enture do the typing for you where your machines already know the answer:
- production counts, read from the counter parameters you've mapped on each Cycle Time Standard;
- downtime, worked out from each machine's runtime parameter;
- tomorrow's production plan, built from your cycle time standards.
The first two run together, once a day, at an hour you choose. The third is separate and runs each evening.
It fills in yesterday, not today
Automatic logging always works on yesterday's shifts — the ones that have finished. A shift row still running when it fires is left alone and picked up by the next day's run, so you never get a half-measured figure recorded as the day's result.
Before you can switch anything on¶
Automation reads real meters, so it needs to know which ones. Each switch stays greyed out, with a line saying what's missing, until its prerequisite is in place.
| To switch on | You need |
|---|---|
| Anything | A shift set under Shift Mapping. Automation works over that shift's rows |
| Generate production counts | At least one machine with counter parameters mapped on a cycle time standard — see Counts filled in from the machine |
| Generate downtime | At least one machine with a runtime parameter mapped, and a reason code to record generated stoppages under |
| Generate tomorrow's plan | At least one machine with an active cycle time standard |
Only machines that are actually mapped take part. A plant where three machines have counters and twelve don't will have those three filled in automatically and the rest logged as before — switching automation on is not all-or-nothing.
Automatic logging¶
Switch on Generate production counts, Generate downtime, or both. At least one has to be on before you can pick an hour.
Choosing the hour¶
Pick any hour from 00:00 to 23:00, in Indian Standard Time. One hour covers both kinds of logging.
Pick one late enough that yesterday's shifts have actually finished. If your last shift of the day runs until 06:00, an hour of 03:00 will find that row still open and leave it for tomorrow's run — which works, but means the figure shows up a day later than you might expect. 07:00 or later would catch it the same morning.
Generate production counts¶
For each machine with a plan for that shift row, Enture reads the counter parameters mapped on its cycle time standard and records good, scrap and rework.
- Where more than one product was planned on the same shift row, the measured total is divided between them in proportion to the planned quantities — the counters measure the machine, not the product, so the plan is the only thing that knows which pieces belong to which SKU.
- Counts already entered by hand are subtracted first. If an operator logged 200 good during the shift and the counter measured 480, automation records the remaining 280 — never the 480 again.
- A machine with no plan for that shift row is skipped. Without a plan there is nothing to say which product was running, and attributing the counts to a guess would be worse than leaving them out. This is the main reason to use automatic planning alongside it.
Generate downtime¶
For each machine with a runtime parameter mapped:
downtime = how long the shift row lasted − how long the meter says the machine ran − downtime already logged
If anything is left over, Enture records it as a single stoppage under the reason code you chose.
Nothing is generated when a row is fully accounted for — a shift where the machine ran the whole time, or where an operator already logged every stoppage, produces no entry at all.
The reason code for generated downtime¶
A stoppage has to carry a reason, and the meter cannot tell Enture what it was — only that the machine wasn't running.
So you choose one in advance. Pick something that reads plainly as not looked at yet — a reason like Unplanned → Unclassified → Not yet classified works well. Generated stoppages then show up in the Downtime Log and the downtime Pareto like any other, and anyone can open one and correct the reason to what actually happened.
Avoid a reason whose category is Planned
A reason code in the Planned category shrinks the planned production time instead of counting as lost running time (see how the figures work). Using a Planned reason as your automation default would quietly improve every machine's Availability instead of recording a loss. Choose an unplanned category.
Where the generated entries appear¶
In exactly the places you'd expect — the Downtime Log, Production Entry, and the OEE figures on the Summary — with a note on each entry saying it was generated automatically. They are ordinary entries: you can correct them, add to them, or delete them.
Reclassifying a generated stoppage sticks¶
Automation looks at each day more than once — the day after, and again the day after that, so that a shift row which hadn't finished the first time still gets picked up. A backfill over dates it has already covered does the same thing.
Each time, it revises the entry it created rather than adding a second one, and recomputes it from the meter. So it is worth knowing what a later pass will and won't change:
| Kept | The reason code and the notes. Once the entry exists these are yours — the meter knows the machine stopped, it can never know why, so your classification is never overwritten |
| Refreshed | The times and the counts. These are what the machine measured, and re-reading it is the point of running again |
So you can reclassify a generated stoppage from Not yet reviewed to Breakdown → Motor failure the morning after, and it will still say that next week.
Counts you correct by hand on a generated entry may be refreshed
If you want to adjust a generated production count, add a separate entry for the difference rather than editing the generated one — a later pass re-reads the counter and will overwrite the figure on its own row. Separate entries are always subtracted first, so they are never double-counted.
The shift's history is updated too
Generated entries feed the OEE history straight away, so the Summary's history grid, trend chart and Analyse view include them as soon as the job has run. There is nothing to refresh or rebuild.
Automatic production planning¶
This half is separate from automatic logging and runs on its own schedule — a fixed time each evening, before the next day begins. It does not use the hour you chose above, and you can enable it on its own without switching any logging on.
Each evening it builds tomorrow's plan in the Machine Production Planner:
- Every machine with an active cycle time standard is planned, across every row of your mapped shift.
- Where a machine makes several products, its shift row is shared between them, so the rows for one machine and shift row add up to that row's length rather than each claiming all of it.
- Planned quantity is worked out from the standard's cycle time, exactly as the Planner's own Auto Generate does.
- Forecast downtime starts at zero. A planned stop is something a person knows about in advance, so it is left for you to enter.
A plan that already exists is never changed¶
This is the one way the nightly job differs from doing it yourself on the Auto Generate screen. There, you see existing plans in the review grid and can deliberately revise them. Here, nobody is reviewing anything — so a plan you wrote by hand, uploaded from a spreadsheet, or generated last night and then corrected is left exactly as it is. The job only ever adds the rows that are missing.
Its minutes are still taken into account: if one product on a machine already has 120 minutes planned, the other products on that shift row share the remaining time rather than the whole row.
Everything generated stays fully editable in the Planner afterwards. Nothing is locked.
Switching it on does not fill in the past¶
Automation starts from the next run onwards. Yesterday gets filled in tomorrow; the weeks before you switched it on stay as they were.
If your machines' counters and runtime meters were already recording during that period, that history can be filled in on request — it is a supported operation, run by Nebeskie support against the dates and the organisation you name, and it uses exactly the same calculation the nightly run uses. It is deliberately not a button here: generating weeks of production and downtime records is not something to set off by accident, and it is worth checking one day's results before committing to a month.
Worth knowing either way:
- It only produces figures for dates where the parameter data genuinely exists. Where it doesn't, those days stay empty rather than being filled with zeros.
- Anything already logged by hand on those dates is left alone and subtracted first, exactly as in a normal run — a backfill cannot double-count a shift someone already recorded.
- The OEE history for the backfilled dates is rebuilt too, so the Summary's history grid and Analyse view pick them up.
Switching it off¶
Turn the switches off and click Save. No further data is generated from the next run onwards.
Entries already generated are not removed — they are real log entries describing real shifts, and deleting them would change history. If you want them gone, delete them from the Downtime Log or Production Entry as you would any other entry.
Who can change these settings¶
Automation settings are engineer-level. Operators with shopfloor logging access won't see this page, and can continue logging by hand exactly as before.
If nothing is being generated¶
Work down this list:
| Check | Why |
|---|---|
| Has the hour passed yet today? | Automatic logging runs once a day, at the hour you chose |
| Are the shifts you expect finished? | A row still open when the job ran is left for the next day |
| Does the machine have the right parameter mapped? | Counter parameters on its cycle time standard; a runtime parameter under Runtime Parameters |
| Is there a plan for that machine and shift row? | Production counts need one to know which product ran. Downtime does not |
| Is the parameter's unit a duration? | A runtime meter whose unit isn't hours, minutes or seconds produces no figure — see Runtime Parameters |
| Did the counters record anything for that shift row? | If the shift's parameter data hasn't been processed, there is nothing to read yet |
A missing reading never becomes a zero. Enture leaves the count blank or generates no stoppage, rather than recording a figure the machine never measured — a fabricated zero scrap would read as perfect Quality, and a fabricated zero downtime as a machine that never stopped.