What a Factory Acceptance Test Should Cover
A factory acceptance test checklist should cover the agreed design, input and output checks, operating sequences, alarms, failure behaviour, outstanding defects and sign-off. Each test should state what to do, what result is expected and what evidence to record. Agree which checks can be completed before delivery and which need the installed equipment at site.
What is the difference between FAT and SAT?
A factory acceptance test (FAT) evaluates equipment or a control system before delivery, while a site acceptance test (SAT) evaluates it after installation at the operating site.
Use the project contract and approved test plans to define the boundary. Do not leave it to interpretation on the test day.
| Area | Include in the FAT plan | Reserve or repeat at SAT |
|---|---|---|
| Control logic | Sequences using available equipment or simulated signals | Behaviour with the installed process |
| Inputs and outputs | Available panel wiring, signal mapping and scaling | Field wiring, instruments and actuator response |
| Operator screens | Commands, indications, permissions and messages | Operation from installed workstations |
| Interfaces | Available connections and agreed simulations | Connections to actual site systems |
| Failure behaviour | Controlled, agreed fault scenarios | Site-dependent scenarios under an approved method |
Record what each simulation represents and what it leaves untested. A simulated valve response should not close a site check for the actual valve's direction, travel or feedback.
How should the test plan come from the FDS?
Build the plan from the approved functional design specification (FDS). For each requirement, identify a test, its expected result and the evidence needed to decide whether it passed.
If the FDS says a pump may start only when specified conditions are satisfied, test both permission to start and refusal to start. If it describes restart behaviour after a trip, test that separately.
Each test record should include:
- The requirement reference and test identifier.
- Preconditions, operating mode and starting state.
- Signals, commands or faults to introduce.
- Expected equipment, screen and alarm responses.
- Actual results and supporting evidence.
- Pass, fail or not tested status.
- Witness and retest details where required.
Resolve missing or conflicting requirements before testing. Our guide to what a functional design specification should contain explains the decisions that should be written down first.
What should the factory acceptance test checklist include?
Use the following checklist as a starting point, then adapt it to the equipment and agreed scope. Assign an owner and an evidence record to every applicable item.
| Check | Acceptance question |
|---|---|
| Test baseline | Are the drawings, software versions and FDS revisions identified? |
| Panel inspection | Does the available equipment match the agreed drawings and scope? |
| I/O checks | Does each tested signal reach the intended destination with the expected meaning? |
| Operating modes | Do mode changes follow the agreed rules? |
| Sequences | Do normal, interrupted and abnormal paths behave as specified? |
| Alarms | Are triggers, messages, acknowledgement and clearing correct? |
| Failure modes | Do faults produce the agreed response and recovery conditions? |
| Interfaces | Are valid, missing and invalid exchanges handled as specified? |
| Access | Can each role perform only its agreed actions? |
| Handover | Are backups, records and outstanding work identified? |
| Acceptance | Are defects, retests and sign-off conditions recorded? |
Mark exclusions explicitly. Distinguish an agreed exclusion from a test that could not be completed.
How should inputs and outputs be checked?
For each input/output (I/O) point, agree the test boundary. Identify whether the check covers software mapping, panel terminals, connected equipment or the complete field loop.
For digital inputs, apply both states and confirm the intended indication and logic response. Check the signal's meaning: does an active state mean healthy, running, open or faulted?
For digital outputs, confirm the intended destination and indication using an approved test arrangement. Plan output testing so that it cannot cause unintended movement or energisation.
For analogue signals, check agreed values across the working range, engineering units, scaling and display precision. Include out-of-range or invalid-signal behaviour where specified.
Record temporary links, overrides and forced values. Before closing the session, remove them and verify that the system has returned to the agreed state.
Which sequences and operating modes should be tested?
Test the paths operators are expected to use, including the awkward ones.
For each applicable mode, check:
- Start permission with every required condition satisfied.
- Refusal to start when a required condition is absent.
- Normal operation and normal stop.
- Pause, abort and interrupted operation where provided.
- Mode changes while equipment is stopped and running.
- Recovery following a trip or loss of a required condition.
- Restart permission and any required operator action.
Illustrative example: a transfer sequence requires a receiving vessel to be available and a discharge valve to confirm open before starting a pump. Test the normal transfer, an unavailable vessel, missing valve feedback and loss of feedback during transfer. Take expected responses from the FDS; do not invent them during the demonstration.
Record timing requirements where the design defines them. “It looked right” is not an acceptance criterion.
What should alarm testing prove?
For each alarm in scope, test the trigger and the complete operator-facing behaviour.
Check the message, equipment reference, priority, timestamp, acknowledgement, clearing condition and history entry where those functions are specified. Confirm any delay or suppression rule against the approved design.
Separate acknowledgement from reset in the test steps. Verify whether either action changes equipment behaviour, and whether a reset is refused while the fault remains.
Review the messages with operations staff: the text should explain what needs attention and support the agreed response. See why operators ignore alarms for the wider review of alarm usefulness.
Which failure modes belong in the FAT?
Choose fault scenarios from the design and project risk review. Agree the method before introducing them.
Consider loss of communications, unavailable devices, invalid instrument values, missing feedback, rejected commands, interrupted interfaces and loss of power where the test arrangement permits it.
For every scenario, answer:
- What state should the equipment enter?
- What should the operator see?
- What commands should be blocked or permitted?
- What data should be retained?
- What must happen before recovery or restart?
For safety-related functions, develop the design and validation approach with a qualified assessor. Identify the dedicated validation work and evidence separately in the acceptance plan. Do not treat a general FAT tick as sufficient acceptance of a safety-related function.
How should the punch list and sign-off work?
Give each unresolved item a unique reference, a clear description, the affected requirement, an owner and closure evidence.
Separate defects from requested changes and site-dependent tests. A new preference should not silently rewrite an approved acceptance criterion.
Agree which findings block acceptance and which may remain open under stated conditions. For each accepted open item, record the consequence, responsible person and closure condition.
Before sign-off, confirm:
- The tested configuration and document revisions.
- Completed, failed and untested checks.
- Retest results after corrections.
- Outstanding items and any restrictions on shipment or use.
- Responsibilities for site testing and final acceptance.
- The scope of each signature.
After a correction, repeat the failed test and any affected tests identified by reviewing the change. Keep the results with the acceptance record.
What common mistakes should the team avoid?
Testing only the happy path. Include refusal, interruption, fault and recovery cases.
Accepting a screen demonstration as a complete I/O check. State exactly which wiring and equipment were exercised.
Writing expected results after seeing the behaviour. Approve acceptance criteria before execution.
Leaving temporary overrides behind. Include their removal and verification in test closure.
Treating FAT sign-off as permission for every later activity. Define the separate conditions for shipment, installation, energisation and production use.
For retrofit work, connect remaining site tests to the shutdown plan. Our guide to planning a PLC retrofit around a shutdown window covers that preparation.
What questions come up before a FAT?
Who should witness the test?
Agree witnesses by responsibility. Include people who can judge the design, operation and maintenance requirements and authorise the relevant acceptance decisions.
Can a FAT proceed without all field equipment?
Define a reduced test scope if needed. Record simulations, limitations and the checks still required with the actual equipment.
Does every punch-list item need closing before shipment?
Set shipment conditions in advance. Record any permitted open items and their consequences; leave blocking findings unresolved only if shipment remains on hold.
What should the plant receive afterwards?
Request the completed test plan, results, defect and retest records, tested software backup, configuration details and agreed site-test scope.
For help defining a project's acceptance scope, contact our team with the equipment scope and available design documents.