Site Acceptance Test

A device may work in a supplier's factory but fail after installation, because site power, network, other equipment and operating conditions can change the result. A site acceptance test, often called SAT, checks the delivered system in its real environment.

The test plan should be agreed before running it, listing requirements, methods, pass criteria, responsible people and evidence, since a vague "works as expected" leaves both sides unsure what acceptance means. The International Society of Automation lists a current standard covering factory acceptance, site acceptance and site integration tests for process-industry automation, though its scope is specific and not every small equipment purchase needs that standard.

Factory acceptance testing and SAT answer different questions: factory tests examine the built system before shipment, SAT checks installation and site operation, and site integration testing may separately cover connected systems. Siemens' installation guidance for one IoT system says to use an SAT template after planning, commissioning and installation, and to include the test in customer handover, which is a product example, not a universal legal sequence.

Preconditions matter, as power, calibration, access, approved drawings and trained operators may be needed, and deviations should be recorded rather than forcing a test under unsuitable conditions. The test should include normal operation and relevant failure responses, because a system that starts once may still fail an alarm or safe shutdown, and risk determines the required depth.

It should use the installed configuration, recording software versions, settings, serial numbers and interfaces tested, since changes after the test may require targeted retesting. Witnesses should be named in the contract or protocol: a supplier technician can run a test, while buyer, engineer or quality staff may witness and approve it, and their signatures have distinct meanings.

A failed test should generate a defect or punch-list item with an owner and retest rule, and a failed criterion must not be turned into a pass by changing the wording after the fact. Minor open items may be carried into a conditional handover if the contract and risk assessment permit it, but safety-critical failures should not be casually deferred, and the record should state what is accepted and what remains.

Site acceptance is distinct from final commercial payment, as the contract may tie payment to SAT, training or additional milestones, so the actual terms should be checked before claiming a signed test releases all money. Testing should generate evidence such as readings, screenshots, logs and calibrated instrument records, because a ticked checklist without measurements may be insufficient for a quantified criterion.

A technically passing system can still need operating instructions and maintenance plans, and handover documents should match the installed version. SAT should not be confused with regulatory certification, since required inspections or permits may be separate and a private acceptance document does not replace a statutory approval.

Realistic time for repair and retesting belongs on the project schedule, because a single-day ceremonial test with no allowance for issues can lead to pressured signatures. After sign-off, the protocol and results should be retained for warranty, maintenance and disputes, recording who accepted which version and any conditions, with no silent edits.