FAT / SAT: how to stop acceptance from becoming a last-minute negotiation.

The real problem: criteria are defined too late

FAT or SAT rarely fail because nobody tested anything. They fail because the parties did not define early enough what “acceptable” means: throughput on which product, for how long, with what reject rate, in which mode, with what data and which deviations.

1. Turn each requirement into a test

A good acceptance matrix links each requirement to a test method, expected result, owner and evidence. Vague statements such as “smooth operation” or “satisfactory quality” should be replaced by observable criteria.

2. Separate FAT from SAT

FAT validates what can reasonably be tested at the supplier: sequences, interlocks, safety, recipes, alarms, documentation, sample cycles and simulated interfaces. SAT validates what depends on the real site: upstream/downstream flow, utilities, network, real product, operators, environment and production speed.

3. Manage open points explicitly

An open point is not automatically a failure. It becomes a risk when it has no owner, date or known impact on acceptance. A punch list should distinguish safety blocker, performance blocker, documentation reserve and non-blocking improvement.

4. Prepare restart and maintenance

Technical acceptance should not stop at the automatic cycle. Test stop/restart, power loss, fault acknowledgement, manual mode, changeover, backup, restore and maintenance-relevant procedures.

A good FAT reduces SAT risk; it does not replace SAT

The objective is not to accumulate signatures. It is to move as much uncertainty as possible into an environment where corrections are cheaper. The earlier interfaces and criteria are tested, the less commissioning becomes a discovery phase.

General technical scoping note. Regulatory applicability, performance and responsibilities must be confirmed project by project.