Augos Factory / Proof / Flour mill
Case study · Food and beverage · Gauteng
Thousands of stops by symptom. 445 by part.
The mill logged thousands of stops a month, but the labels named symptoms, not failed parts. Step by step, one machine and asset at a time, it changed how stops were recorded.
The situation
Every stop recorded, few explained
Stops were labelled "PLC fault", "SCADA issue" or "No production". When a sifter blocked product flow, it went in as "Choke / Blocked".
The record showed that the mill had stopped, not what had failed. Engineers started each investigation from scratch, so maintenance stayed reactive.
What the record showed
Fewer entries, each one named
Stops recorded in one month, before and after the labels changed:
| How stops were labelled | Stops recorded in a month |
|---|---|
| By system or symptom | 3,372 |
| By asset and component | 445 |
The count fell as recording changed, so it does not measure fewer failures. It measures stops that now name a part.
One belt
Over 20% of monthly downtime came from one conveyor belt on the roller mills, once each stop named its part.
Source: Augos platform, Downtime Report. One anonymised flour mill, before and after the change. Site, products and people removed.
What the mill changed
Four steps, in the mill's own order
Step 1: the starting point
3,372 stops in one month, labelled by system or by the symptom on the floor.
Step 2: tie every stop to an asset
The team began linking each stop to a specific machine and its asset tag.
Step 3: the pattern shows
One belt on the roller mills held over 20% of monthly downtime. Stops recorded fell by more than half.
Step 4: every stop names its part
445 stops a month, by asset and component. Planned non-running time is labelled "dark hours", apart from failures.
The result
Same mill, a different record
| Measure | Before | After |
|---|---|---|
| Stops recorded in a month | 3,372 | 445 |
| How a stop is labelled | By system or symptom | By asset and component |
| Electrical stops | "PLC fault", "SCADA issue" | Motor failure or drive fault, by asset |
| Mechanical stops | "Choke / Blocked" | Idler pulley, gearbox or a named bearing |
| Planned non-running time | Mixed in with failures | Labelled "dark hours" |
From the mill's Downtime Reports, before and after the change.
With stops named by part, the maintenance team can track the time between failures for each component and plan spares around it.
Method
- Source: Augos platform Downtime Report, the mill's own records.
- Periods: one month before the change, and months after it.
- Mill: one flour mill, with roller mills, sifters and conveyors.
- Counts: every stop the mill recorded, as recorded at the time.
- Rand value: none shown, since this case reports counts rather than hours.
What this case leaves out
- Counts only: the case reports stops recorded, not hours or rand.
- A falling count reflects better labels, not fewer failures.
Measured figures only, as the platform recorded them.
Questions
This case, answered
Does a lower stop count mean fewer failures?
No. The count fell as recording changed. It measures stops that now name a part, not how often things failed.
What changed in how stops were recorded?
The team linked each stop to a specific machine and its asset tag. Electrical and mechanical stops now name the motor, drive, pulley, gearbox or bearing. Planned non-running time is labelled dark hours, apart from failures.
Why is there no rand value?
This case reports counts of stops recorded, not hours or rand.
Find losses like these on your own line.
Thirty minutes with a manufacturing specialist, on your own lines and products.
- Every stop timed and given a reason
- Read by your own team
- Prices in rand, ex VAT
A manufacturing specialist replies within two business hours to set a 30-minute call.