Unattended plants
Equipment in auxiliary units that nobody goes into between rounds.
Teal Neural
How long it was really stopped.
The real machine stoppage time, not the one written down. And what was going on in front of it just before.
The report is written by whoever was there.
A plant's stoppage log gets filled in by a person, at the end of the shift, from memory. The long ones all get written down. The three-minute ones, hardly any.
And those are the ones that add up. Twenty three-minute stoppages are an hour of the shift that appears nowhere and that, therefore, nobody tries to fix.
When output does not add up at the end of the month, people go looking in the big stoppages. They are not there.
You can see the machine running or not.
From the camera you can tell whether the machine is working: whether the belt is running, whether the arm is moving, whether product is coming out. That gives you the real running and stopped time, with nobody writing it down.
Every stoppage is left with its time, its duration and its video. And at the end of the month there is a breakdown: how many, of what length and on which machine.
The conversation changes from 'the line is playing up' to 'this machine stops 34 times a shift in the same position'.
What it measures.
- Machine running or stopped, and the time in each state.
- Micro-stoppages and how they break down by hour and by shift.
- Visible faults and breakages: broken parts, loose belts, open guards.
- Changes in output: the same machine, slower than a month ago.
- Lubrication checks: grease present, dripping, build-up.
- Leaks of oil, air or coolant.
- Unattended machines: whether they are running when there is nobody in front of them.
- Response time: how long it takes to get there and to sort it out.
- Component temperature, with a thermal camera.
Lo que Teal Neural entiende
5 detecciones grabadas.
Producción y línea de fabricación.
Teal
Teal
Teal
Teal
Teal
Cada una es una grabación del sistema funcionando. Pasa el ratón por encima —o tabula hasta ella— para verla en movimiento.
A real case.
In a machining plant, the night shift report logged an average of 12 minutes of stoppage. What was measured was 41.
The difference was not bad faith: they were two- and three-minute stoppages that nobody counts as a stoppage. The hourly breakdown showed they were concentrated between 3:00 and 4:30, which is when the operator does the round on the other line.
Where it is used.
Compressors and refrigeration
Real running, leaks and component temperature.
Presses and stamping
Cycles, guards and tooling changeover time.
Packaging
Micro-stoppages and the rate falling away over the shift.
Plants with old machinery
Equipment that publishes no status and that today is only known to be stopped when someone walks past.
Night shifts
When there are fewer people and nobody writes down the short stoppages.
Machine shops
Real spindle running time against time with the machine switched on.
Lines with format changeovers
How long each changeover really takes, against what is estimated.
The first week is the one that convinces.
The conversation about machine availability usually gets stuck on whether the report is filled in properly or not. It is an argument with no way out because nobody has the figure.
With a week of measurement there is a number nobody has written: this many stoppages, of this length, at these times. From there the conversation is a different one.
And almost always it turns out the team was right on the qualitative side — 'this machine is trouble' — and well short on the quantitative one.
How you start.
- One machine, the one everybody points at.
- A low stoppage threshold, so the short ones show up.
- A week of measuring, telling nobody and changing nothing.
- Compare with the shift report: the difference is the argument.
The machines nobody looks at.
In almost every plant there is equipment that runs on its own and that nobody goes near: a compressor, a refrigeration unit, a pump in an auxiliary unit. You know it is fine because nobody has called.
Those are the ones that gain most. A camera looking at the machine tells you whether it is turning, whether it is dripping, whether it has got hot and whether anyone has been in front of it. And it tells you without anybody having to go.
In spread-out installations — a plant with separate buildings, an industrial estate, several sites — that is the difference between knowing today and knowing on Thursday's round.
Lubrication, which looks minor.
An ungreased chain sets off nothing. You notice it in the power draw, in the noise and, six months later, in a breakage.
From the camera you can see the state of the greasing: whether there is grease where there should be, whether it is dripping where it should not and whether dirt has built up on top of it. It is one of the dullest things in maintenance and one of the ones that prevents most faults.
And output that falls away on its own.
A machine rarely stops all at once: first it goes a bit slower, then a bit slower still. Because the reference is 'normal' and normal keeps shifting, nobody notices.
With cycle time measured daily, you can see the degradation curve. And with it you can plan the work before the machine decides for itself when to stop.
Casos grabados
Funcionando, en sitios de verdad.
Botellas atascadas - Optimiza tu línea de producción con Infinity Neural
Optimización de la Producción Frutícola con Teal Neural
Monitorización de manómetros con Teal Neural
Teal Neural: Optimizando la producción industrial
Teal Neural: Optimización y control en cintas de transporte
What it does not do.
It does not know WHY it stopped. It knows when and how long, and it shows you the video. The cause is put in by whoever looks at it.
And it does not replace the controller's own signal where there is one: if the machine already publishes its status, that is more reliable. This covers the ones that do not publish it, which tend to be the old ones.
And what sits alongside it.
Other things the same family understands, on the same cameras.
What people ask us.
▸Do we have to replace the cameras we already have?
Almost never all of them, and sometimes none. What decides it is not the brand or the megapixels: it is the angle, the distance and the light. A modest camera well placed works better than a good one looking from the wrong spot.
The first thing we ask for is an image from the camera exactly as it is. With that you can see whether it works, whether it has to be moved or whether another one is needed. And if another one is needed, we say so then and not after signing.
▸Where is the video processed?
Wherever it needs to be. It can be on a machine on your own site, with the image never leaving it, or on a server. On sites where the video cannot leave for reasons of policy or regulation, it is processed inside, full stop.
What goes out in that case is not image: it is the alerts and the figures.
▸And data protection?
None of this identifies people. It tells apart figures, postures, objects and vehicles, not faces or names.
Even so, a system that processes images of a workplace has its obligations: informing people, defining what it is used for and how long things are kept. That gets sorted at the start, with whoever handles data protection at the company, and it is not paperwork left to the end.
▸Is it any use for working out OEE?
It gives you availability measured rather than estimated, which is the leg that is known least well. Performance comes from counting product and quality from visual checks, which are the other two capabilities in this family.
Together they give you an OEE built on data and not on the shift report.
▸How many machines per camera?
One, if you have to tell each machine's state apart reliably. A camera covering a whole line tells you whether the line is running, not which machine stopped it.
On short lines with well-separated machines, two or three can work. We look at it with a photo.
▸And if the machine is inside a closed cabin?
Then it cannot be seen, and this is no use. What you can do is watch what goes in and what comes out, which tells you whether it is working even if you cannot see inside.
▸Does it detect the reason for the stoppage?
It does not classify it on its own. It gives you the moment and the video, and whoever looks at it recognises it straight away.
Over time, and seeing that the stoppages repeat at the same point, the most frequent ones can be labelled so they come out already classified.
▸Does the machine's own system not give us this already?
If the machine publishes its status and it is reliable, yes, and then you do not need this. The problem is the ones that do not publish it, which tend to be the oldest and the ones that stop most.
It is also useful as a cross-check: there is equipment taken to be running with the contactor closed even though the shaft is not turning.
▸What counts as a stoppage?
Whatever you decide. You set a threshold — ten seconds without movement, for example — and from there it counts as a stoppage.
What is interesting is usually to lower that threshold: stoppages under a minute are the ones that appear in no report and the ones that add up most.
▸Does it tell why it stopped?
No. It says when and how long, and it shows you the video of that moment. The cause is put in by whoever looks at it, who recognises it straight away anyway because they are watching it.
▸Is it any use for working out OEE?
It gives you the availability part with measured figures instead of estimated ones. Performance and quality come from counting product, which is the other capability in this family.
Get started
Do you want this on your site?
This is already recorded working. What is left is to see whether your camera allows it: where it is fitted, what angle it has and what can be seen from there. If it will not work, we will tell you.