The Monday meeting
Every Monday, the production team looks at last week’s output. Line 3 was 12% behind plan again. The shift leader says there were “a lot of small stops”, maintenance says the machine was fine, and planning says the material arrived on time. Each explanation is reasonable, and nobody can prove or disprove any of them. The meeting ends with an agreement to “keep an eye on it”.
Production monitoring replaces that discussion with data.
Production and production monitoring
Production is the measurable result of manufacturing: of a good, or of a service. It is counted in parts, batches, kilograms or hours, and it is what the plan is measured against.
Production monitoring is the systematic follow-up of the manufacturing process, so that production can be carried out according to plan. It supports management and gives feedback in the form of production data. Where the plan says what should happen, monitoring shows what did happen, and why the two differ.
What to record
The core data is simple:
- Machine state: running, stopped, set-up, planned stop. With time stamps, so that every minute of the shift is accounted for.
- Stops and stop causes: when each stop started and ended, and why. Causes are chosen from a fixed list such as material, tool, quality, breakdown or no operator.
- Cycle times: how long each part or cycle takes, compared with the standard.
- Counts: produced, good, scrapped and reworked.
- Context: which machine, article, order and shift. Without this, the data cannot be compared across products or traced to an order.
State and counts can often be read automatically from the machine’s controller or a sensor. The cause usually needs a person, which is why entering it must be quick and close to the machine.
From stops to improvement
The figure shows how it works. One shift has three stops: a material change, a tool change and one stop without a cause. On its own, one shift says little. Summed over four weeks, the picture is clear, and the largest item is stops with unknown cause.
Worked example (illustrative values). Over four weeks, line 3 records these stop minutes:
- unknown cause: 55 minutes
- waiting for material: 40 minutes
- tool change: 30 minutes
- quality check: 20 minutes
A Pareto ordering like this tells the team where to start. Here the first action is not technical: it is to make sure operators record a cause for every stop. Once they do, the unknown minutes are redistributed among real causes, and the next action can be aimed at the largest of them.
How it connects to OEE
Production monitoring data is exactly what is needed to calculate OEE. Stops and planned time give availability, cycle times and counts give performance, and good versus scrapped parts give quality. ISO 22400-2:2014 defines these and other key performance indicators for manufacturing operations, which helps when results are to be compared over time.
Good practice
- Account for all the time. A timeline with gaps cannot be trusted. Every minute is either running, a stop with a cause, or planned non-production.
- Keep the cause list short. Ten clear causes are filled in; fifty vague ones are not.
- Set a threshold for short stops. Decide how short a stop must be to count as a minor stop, and treat those as a performance loss rather than asking for a cause each time.
- Show the data where the work is. A screen at the line showing current state and today’s stops changes behaviour more than a monthly report.
- Act on it. Review the largest losses regularly, decide on actions, and check whether they worked.
Limitations
Manual registration is slow and incomplete, especially for short stops. Automatic collection, as part of digitalisation and Industry 4.0, gives better coverage, but still needs people to supply causes. And production data shows where time is lost, not always why: the root cause of a recurring tool stop still needs investigation on the shop floor.
In short
Production monitoring shows what was produced, when the machines ran and why they stopped. Give every stop a time and a cause, sum it over weeks, and work on the biggest loss first.
