Process reasoning
A centrifuge result combines machine behaviour, feed variation and measurement error. Cake moisture can vary across a discharge; a liquid sample can change as solids settle in its bottle. Use the same sampling interval and analytical method for each trial. Record the stable operating period rather than combining start-up, steady operation and shutdown into a single unexplained average. A change smaller than the variability of the test may not justify a purchase decision.
ILLUSTRATIVE EXAMPLE
Apply the reasoning
Consider three illustrative cake dry-solids results for option A: 24.0%, 24.8% and 24.2%. Option B reports a single result of 24.6%. The available evidence does not establish that B consistently makes drier cake. Request comparable repeated runs, feed properties and operating loads. Also compare solids capture and power: a small dryness improvement may come with a different process cost.
What to record
| Item | Basis to document |
|---|---|
| Measurement | Document the method, sample location and units |
| Repeat runs | Keep feed and machine settings comparable |
| Outliers | Investigate the cause; do not quietly remove inconvenient results |
| Acceptance | Agree limits and treatment of variability before the trial |
Review checklist
- Collect feed and outlet samples over the same interval.
- Record instrument resolution and the laboratory method.
- Separate transient periods from stable production.
- Retain individual results alongside averages.
Your next decision
Use a trial summary with a range, a mean where meaningful, and the reasons for any excluded data. Ask the supplier to explain which variables were controlled and which were observed. A larger pilot campaign is worthwhile when the difference between options is small but its lifetime economic effect is large. Do not present the illustrative numbers here as test results for either brand.