Constraint-Level track. Three inline buffers between the stages. The recorded result is that system efficiency does not move — and a live sweep shows just how far that holds.
At a glance
| Experiment | Multi Run |
| Duration | 90 days |
| System efficiency | 54.1% |
| Production | 70,113 pallets |
| Output unit | pallets |
What this model is
The Converter model with three inline buffers added: between Capper and Labeler, between Labeler and Case Packer, and between Case Packer and Palletizer.
Buffers decouple. A stage with somewhere to put its output does not block when the next stage stops, and a stage with a supply does not starve. The reasonable expectation is that adding them raises output.
The line
| Node | Name |
|---|---|
c0 |
Capper |
c1 |
Labeler |
n0 |
Filler |
n1 |
Case Packer |
n2 |
Palletizer |
The interrupts
| Interrupt | Machine | Time to failure | Time to repair |
|---|---|---|---|
| Palletizer | Palletizer | Weibull(30, 1.00) | Weibull(5, 1.00) |
| Capper | Capper | Weibull(30, 1.00) | Weibull(5, 1.00) |
| Case Packer | Case Packer | Weibull(30, 1.00) | Weibull(5, 1.00) |
| Labeler | Labeler | Weibull(30, 1.00) | Weibull(5, 1.00) |
| Filler | Filler | Weibull(30, 1.00) | Weibull(5, 1.00) |
Parameters are read from the model file as it ships. Open it in the sandbox and you will see these values.
The recorded result
System efficiency is unchanged against the Converter model without buffers. Buffers decouple stages from each other; they do not fix a constraint that is limited by its own internal downtime. There was no starvation to absorb.
What the run produced
- Bottleneck operating availability: 85.1%
- System OEE: 54.1%
- Bottleneck total downtime: 204.5 hours across 2,415 events
- Mean time between failures: 29.0 min · Mean time to recover: 5.1 min
- Throughput: 70,113 pallets (≈ 7.01M bottles)
- Blocked or starved: 787.0 hours (36.4%)
Run it yourself
What 80× more buffer actually buys
A live sweep run against the engine now — seven runs, 90 simulated days each, all three inline buffers set to the same capacity. Separate from the recorded figures above.
pallets producedheld in the buffers at the end of the run
Output is identical at every buffer size. The rising amber bar is not production — it is material sitting in the buffers when the run ended, and it scales exactly with capacity. Eighty times the buffer, eighty times the work in progress, and not one extra unit out of the door.
This is a live run and its figures are its own; the recorded results above come from a separate transcription and the two are not expected to match digit for digit.
Where to go next
The same experiment against real named failure modes: Buffer Options — LEDS.
Open it
This model ships with ReliaSim® and runs in the browser sandbox — no install required. Open + Buffer Options in the sandbox.
Every recorded figure on this page is the published result for the model as it ships.