A factory acceptance test on a mooring winch in a Chinese yard does not begin on the test bench. By the time the prime mover is energised, the verifiable failure modes have already been decided in the document review.

Below is the pre-FAT sequence we run on Type 4 winches and equivalent deck machinery, refined across yards in Jiangsu, Zhejiang, and Liaoning over the past several years. Senior procurement teams from EPC contractors and operators in OCIMF MEG4 environments will recognise the pattern.

1. Brake PQR and dynamic friction curve

OCIMF MEG4 paragraph 4.6 sets the dynamic brake holding requirement at 60% of MBL for a single-layer wrap. The certificate the yard hands you at FAT typically shows a static holding test result. That is not the same data point.

What we ask for, and verify against the actual brake assembly serial number on the unit on the bench:

• Brake material PQR (Procedure Qualification Record) issued by the brake lining manufacturer, traceable to a specific production batch.
• Dynamic friction coefficient curve across the operating temperature window the winch will see in service. This is not a single number. A brake that passes at 20 degrees C ambient can fade at 70 degrees in a tropical service condition.
• Comparison of the as-built brake clearance to the design tolerance band, witnessed before the load test, not after.

If the yard cannot produce these documents at the bench, the FAT pass is a performance demonstration, not a verification. We mark the hold point and stop.

2. Hydraulic cleanliness to ISO 4406

The hydraulic system on a deck winch carries a cleanliness target. The realistic working target on new-build commissioning is ISO 4406 18/16/13 or better. Anything looser than 20/18/15 will accelerate wear on the proportional valves and the swash-plate of the variable displacement pump.

Required documents and observations:

• Flushing record showing flow rate, duration, and the particle counter reading at sign-off, taken from a sampling point downstream of the return filter.
• Filter element serial numbers and beta ratings (typically beta 200 at 10 microns or finer for new-build commissioning).
• The particle counter calibration certificate, traceable to a recognised national metrology institute.

A system loaded for FAT before adequate flushing can pass the load test and fail in service inside 500 operating hours. The cost difference is borne entirely by the operator.

3. Class surveyor pre-FAT log

If the winch is supplied for class — DNV, ABS, BV, LR, CCS, RINA, KR, or NK — the surveyor will have visited the yard at scheduled hold and witness points before the FAT. The pre-FAT log is the audit trail for those visits.

What to verify:

• Each hold point in the approved Inspection and Test Plan has a surveyor signature, the date, and the surveyor's reference number.
• Witness points are distinguished from hold points. A witness point allows the manufacturer to proceed if the surveyor is delayed; a hold point does not. The ITP must show which is which.
• Surveillance points (lower than witness) are noted and counter-signed by the QA inspector.

An ITP with unsigned hold points is not a valid quality record. We have caught yards presenting an ITP at FAT where the hold points were signed retrospectively the day before the test. That is a procedural non-conformance and class will not accept it on appeal.

4. Load cell traceability

The number on the digital readout during the load test is only as credible as the calibration chain on the load cell.

What we record at FAT:

• Load cell serial number and calibration certificate, with traceability to a recognised national standard. NIST in the United States, NIM in China, NPL in the United Kingdom, PTB in Germany. The certificate must show the calibration date, the next-due date, and the standard reference.
• Calibration must be in-date on the FAT day. We have seen certificates back-dated to keep them inside validity. We photograph the unit, the certificate, and the seal.
• The accuracy class of the load cell against the precision required by the test method. A 0.5% accuracy class is the minimum we accept for verification of MBL-related figures.

The point

FAT is not a moment. It is the closing of an audit trail that started at material procurement. By the time we are at the test bench, the verifiable failure modes are already decided. Either the documentation chain holds, or the test demonstrates a number we cannot defend in service.

If you are sourcing deck machinery from China and the FAT date is approaching, we are happy to share the longer version of this checklist on request.