Designing Modular Gas & Marine-Fuel Skids: Integration, Interfaces & FAT for EPC Delivery (2026)

The short answer for engineers in a hurry: a modular gas or marine-fuel skid succeeds or fails long before it reaches site. It is decided in three places — where you draw the module boundary, how cleanly you manage the mechanical, process, electrical and control interfaces to the surrounding plant or vessel, and whether the Factory Acceptance Test (FAT) at the fabricator proves the package works before it ships. Get those three right and integration at the EPC site or shipyard becomes a connect-and-commission exercise. Get them wrong and you inherit field rework, missed hold points, and a commissioning window that quietly consumes your schedule float.

This guide is the design-and-delivery companion to our earlier piece on how modular skid packages help EPC projects control delivery risk. That article looked at schedule. This one goes into the engineering — the design principles, the integration logic, the interface discipline, and the FAT regime that make a skid genuinely "plug-and-play."

The timing matters. As of early 2026, modular skid-based supply is moving from a convenience to the default delivery model for marine fuel and gas systems. Nikkiso Clean Energy & Industrial Gases launched a next-generation marine fuel gas supply system (FGSS) at the end of 2025 built around multi-phase pump-skid manufacturing programmes, and is expanding China-based engineering and fabrication capacity to deliver complex skid solutions at fleet scale (Nikkiso CE&IG, 2025). Shipowners are now ordering dual-fuel newbuilds across three parallel fuel paths — LNG, methanol and ammonia — each demanding its own modular supply system: FGSS for LNG, MFSS for methanol, AFSS for ammonia (Spectra by MHI, 2025). The regulatory frame is moving just as fast: the IMO approved Interim Guidelines for ships using ammonia as fuel (MSC.1/Circ.1687, 26 February 2025), and draft IGC Code amendments are expected at MSC 111 in May 2026 with entry into force targeted for 1 July 2028 (DNV, 2025; BIMCO, 2025).

Translation for the fabrication floor: the volume of modular gas and marine-fuel skids is climbing, the fuels are getting more hazardous and more tightly regulated, and the burden of proving a package works has shifted firmly to the FAT. This guide draws on real Lmart skid programmes — Wärtsilä reliquefaction and CIP units, TGE LNG fuel systems, Höglund methanol bunker skids, FPSO process modules, and LNG-terminal BOG skids — to show what good design, clean interfaces, and a rigorous FAT actually look like.

Table of Contents

1. What a Modular Skid Really Is — and Why It Wins
2. Design Principle 1: Shop-Fabrication vs Field Installation
3. Design Principle 2: Drawing the Module Boundary
4. Design Principle 3: Structural Frame & Baseplate Design
5. Design Principle 4: Transport & Lifting Constraints
6. Integration: From P&ID to Skid Layout
7. What Goes On a Skid: Component Integration
8. Interface Management: The Discipline That Saves Projects
9. Factory Acceptance Test (FAT): Proving It Works Before It Ships
10. Marine Gas Skids vs Onshore Petrochemical Skids
11. Certification: Class Society + Pressure Code Dual Track
12. The Design / Interface / FAT Checklist
13. Real Project Mapping
14. Frequently Asked Questions
15. Related Reading

---

1. What a Modular Skid Really Is — and Why It Wins

A modular skid is a complete, self-contained functional package — process equipment, piping, instrumentation and controls — pre-assembled onto a single structural frame in a workshop, then transported to site and connected to the surrounding plant or vessel through a defined set of interfaces. Instead of receiving loose vessels, pumps, valves and kilometres of pipe to be assembled in the field, the EPC contractor or shipyard receives a tested, documented unit that needs only utility tie-ins, electrical connection and commissioning.

The definition sounds simple. The engineering value is anything but.

Why modularization wins on real projects:

• The work moves to a controlled environment. Welding done in a climate-controlled shop with overhead cranes, jigging, and bench-mounted NDE equipment is faster, more repeatable and easier to inspect than welding on scaffolding 40 metres up an offshore module or inside a ship's engine room. Weld reject rates drop, and inspection throughput rises.
• Quality is provable at one point in time. A skid can be hydrostatically and pneumatically tested, function-tested, and run-tested as a complete system at the fabricator. That single FAT event captures evidence that would otherwise be scattered across months of site work.
• Site congestion drops. On a busy FPSO topsides build or a shipyard berth with dozens of trades competing for space, every system that arrives complete removes labour-hours, scaffolding and crane time from the critical path.
• Schedule parallelism. Skid fabrication runs concurrently with civil works, hull construction or foundation pours. The package and the place it goes are built at the same time, then married.

The trade-off is that all the integration risk is front-loaded into design. You cannot fix a badly drawn module boundary or a missed interface in the field without breaking the very advantages that justified going modular. That is precisely why this guide spends most of its words on boundaries, interfaces and FAT.

Quotable: A modular skid does not remove integration work — it relocates it. The integration happens in the design office and the fabrication shop instead of on the critical path at site. The discipline is in where you draw the lines and how you prove they hold.

---

2. Design Principle 1: Shop-Fabrication vs Field Installation

The first design decision on any skid programme is brutally practical: what gets built in the shop, and what is left for the field? The instinct is "build everything in the shop," but that instinct collides with three hard limits — transport envelope, lifting capacity, and access at the destination. The art is maximizing shop scope without exceeding those limits.

What belongs in the shop

• All pressure-boundary welding. Vessels, separators, buffer drums, intercoolers and the process piping that connects them should be welded, NDE'd and pressure-tested in the shop, where qualified welders, controlled WPS execution and bench NDE produce repeatable results. On Lmart's Wärtsilä 91K reliquefaction skids, every production weld coupon for the cryogenic 316L process circuit was impact-tested at –196 °C in the shop — a test regime that is impractical to administer on a ship's berth.
• Equipment mounting and alignment. Pumps, compressors and rotating equipment are mounted, aligned and shimmed on the frame in the shop, with alignment recorded. Doing this on a level shop floor under crane is far more accurate than on a moving vessel or an uneven foundation.
• Instrumentation, cable trays and junction boxes. Field instruments, local panels, cable trays and inter-device wiring should be installed and continuity-checked in the shop. On the TGE LNG fuel system skids, Lmart installed process piping, instrumentation, cable trays and insulation supports as a complete assembly before the package left the workshop.
• Insulation supports and cold boxes. Cryogenic skids need insulation cleats, stand-offs and (where applicable) cold-box structures fitted in the shop so that field insulation is a simple wrap-and-clad operation.

What is deliberately left for the field

• Inter-skid tie-in spools. When a system is split across several skids (a bunker skid, an FPR skid and a scrubber skid, for example), the spool pieces that bridge them are often left as field welds or bolted spools, cut to length after the modules are positioned. This absorbs the small dimensional tolerances of foundation and hull positioning.
• Final utility connections. Cooling water, instrument air, nitrogen, electrical power and control-network connections are made at site to defined tie-in points.
• Soft connections that must flex. Expansion bellows, flexible hoses and anti-vibration mounts that need site adjustment after the skid is bolted down.

The governing rule

Every weld you move into the shop is a weld inspected under better conditions and removed from the critical path. Every connection you leave for the field is a connection that absorbs tolerance and avoids stress build-up. A good design pushes the boundary toward the shop until it hits a transport or lifting limit — and not a millimetre further. We return to those limits in Section 5.

---

3. Design Principle 2: Drawing the Module Boundary

If there is one design decision that determines whether a multi-skid system integrates cleanly or fights you at every tie-in, it is where you draw the module boundary. The boundary defines what is inside each skid, what crosses between skids, and what connects to the host plant or vessel. Every crossing is an interface — and every interface is a place where two parties' assumptions can fail to meet.

Principles for a clean boundary

1. Cut at natural functional seams. A boundary drawn through the middle of a tightly coupled process loop creates dozens of small interfaces; a boundary drawn at the natural inlet/outlet of a functional unit creates a handful of clean ones. The Höglund methanol system is a textbook case: it was split into a bunker skid (receiving), an FPR — fuel preparation room — skid (conditioning and supply to engines), and a WGS — waste gas scrubber — skid (exhaust treatment). Each module is a self-contained function. The crossings between them are a manageable, countable set of piping, electrical and control connections rather than a tangled web.

2. Cut where the interface is easiest to verify. A boundary at a flanged connection with a clear isolation point is easy to leak-test and easy to commission. A boundary mid-weld in an inaccessible location is neither.

3. Minimize live-process crossings. The fewer process lines that cross a boundary while the system runs, the fewer points of leak risk and the fewer items on the inter-skid FAT. Group equipment so that process intensity stays inside a module and only conditioned, controlled streams cross out.

4. Respect the destination's access. On marine projects, the boundary is partly dictated by what fits through an engine-room hatch. Alfa Laval's marine skids for Lmart had to pass through engine-room access hatches and bolt down under limited overhead clearance — so the module boundary was set by the largest piece that could physically be carried to its final position, not by process logic alone.

The boundary register

On any multi-skid system, the single most valuable design artefact is a boundary / interface register — a controlled list of every connection that crosses a module boundary, with its type (process / mechanical / electrical / control), its terminating detail (flange rating, cable spec, signal type), the responsible party on each side, and its verification method. On the Höglund three-skid programme, Lmart assigned a dedicated project engineer to own exactly this register, using Höglund's 3D model as the master reference and verifying every inter-skid piping connection, electrical penetration and control signal during concurrent fabrication.

Quotable: Draw module boundaries at functional seams, not through process loops. Every line that crosses a boundary is an interface someone must own, verify and sign off — so the fewer crossings, the lower the integration risk.

---

4. Design Principle 3: Structural Frame & Baseplate Design

The structural frame is the part of a skid that everyone takes for granted until it cracks a weld, transmits a vibration, or refuses to bolt down flat. On a process skid the frame is not just a transport pallet — it is a permanent structural element that must carry static equipment loads, resist dynamic and seismic loads, survive lifting and transport, and provide a stable foundation for rotating equipment alignment over the life of the plant or vessel.

What the frame has to do

• Carry the operating dead load of all mounted equipment, full of process fluid, plus piping, insulation and contents.
• Resist dynamic loads. On marine skids this means ship-motion accelerations (roll, pitch, heave), main-engine and propulsion vibration, and — per class rules — collision and grounding loads. Lmart's marine skids are routinely designed for roll of ±30°, pitch of ±10°, plus heave and engine-induced vibration. On FPSO topsides the numbers climb: the Jord NAG module for the Abigail Joseph FPSO was designed for vessel roll up to 22.5° and pitch up to 12.5° plus wave-induced accelerations — load cases that have no equivalent on a fixed foundation onshore.
• Maintain rotating-equipment alignment. A frame that flexes under load lets pump and compressor shafts drift out of alignment, driving bearing and seal failures. Frame stiffness under operating load — not just strength — is a design requirement for any skid carrying rotating machinery.
• Provide grounding, drainage and drip containment. Particularly for toxic or flammable fluids, the baseplate often serves as a drip tray with bunding and a drain point.

Frame design choices

Frame element	Onshore process skid	Marine / offshore skid
Primary structure	Rolled I-beams / channels, bolted or welded	Box-section or heavy channel, fully welded, class-approved
Design loads	Dead + live + seismic (site-specific)	Dead + ship motion (roll/pitch/heave) + vibration + collision/grounding per class
Surface protection	Industrial epoxy / inorganic zinc per spec	Marine coating system (multi-coat epoxy), salt-spray rated
Lifting provisions	Lifting lugs, fork pockets, jacking points	Certified, class-witnessed lifting lugs with calculated SWL
Deflection limit	Per equipment-vendor tolerance	Tighter, governed by alignment + class requirements

A note on coatings

Marine and offshore frames live in a salt atmosphere for decades. Every Lmart marine skid frame uses a marine coating system — typically a multi-coat epoxy build over blast-cleaned steel — specified to the project's corrosion category. On the CNOOC Yingkou BOG terminal skids, by contrast, the frame was Q235B carbon steel with an industrial coating appropriate to an onshore LNG-terminal environment. The structural logic is similar; the coating spec is driven by the destination's corrosion class.

---

5. Design Principle 4: Transport & Lifting Constraints

This is where modular design philosophy meets the unforgiving reality of roads, cranes, hatches and decks. The most elegant integrated skid is worthless if it cannot reach its final position. Transport and lifting constraints are design inputs that must be fixed before the module boundary and frame are finalized — not discovered afterwards.

The constraints that govern module size

• Road / sea transport envelope. Over-the-road shipping has width, height and length limits before a load becomes "abnormal" and requires permits, escorts and route surveys. Sea and inland-waterway transport relaxes some limits but introduces others (container dimensions, breakbulk lifting, lashing).
• Crane capacity at destination. The shipyard's or site's available crane (or the vessel's own deck crane) sets the maximum lift weight. A skid that exceeds the lift capacity at its destination is an immovable object.
• Access route at destination. On marine projects this is decisive. As noted for the Alfa Laval skids, the module had to fit through engine-room access hatches with limited overhead clearance. The largest single piece that can be physically carried, lowered and manoeuvred into final position is a hard cap on module size — often far smaller than what transport alone would allow.
• Lifting geometry. Lift lugs must be positioned so the skid lifts level, with the centre of gravity calculated and the lug SWL verified. On class projects the lifting arrangement is itself a witnessed, approved design.

How constraints flow back into design

When the destination crane caps a lift at, say, the weight of a fully dressed module, the designer has three levers: split the module (creating more interfaces — see Section 3), defer some equipment to field installation (reducing shop scope — see Section 2), or strip removable items (motors, coolers) for separate shipment and site re-fit. Each lever trades one advantage for another. The decision is an engineering judgement, made with the transport and lifting data in hand from the start.

Quotable: Transport envelope, crane capacity and destination access are not logistics afterthoughts — they are primary design inputs. Fix them before you draw the module boundary, because they cap how much integration you can legally and physically deliver as one piece.

---

6. Integration: From P&ID to Skid Layout

A P&ID tells you what connects to what. It says nothing about where the equipment sits, how the pipe routes in three dimensions, whether a valve handwheel is reachable, or whether a pump can be pulled for maintenance without dismantling half the skid. Translating a P&ID into a buildable, operable, maintainable skid layout is the core integration discipline — and it is where most of the engineering hours on a skid live.

The translation workflow

Step 1 — P&ID freeze and equipment list. Integration cannot begin against a moving P&ID. The process design must be frozen, and every tagged item (vessels, pumps, valves, instruments, heat exchangers) listed with its physical envelope, weight, nozzle orientation and maintenance access requirement.

Step 2 — Block layout within the envelope. Equipment is arranged within the frame envelope (capped by the transport/lifting limits) to satisfy several competing demands at once: process flow logic (minimize pipe runs and pressure drop), maintenance access (pull space for pump rotors, tube-bundle withdrawal for heat exchangers, valve operability), centre-of-gravity for lifting, and — on marine skids — fit within the host envelope.

Step 3 — 3D modelling and clash detection. This is non-negotiable on a dense skid. Every Lmart marine skid is modelled in 3D before fabrication. On the Wärtsilä 91K reliquefaction skids, all piping was modelled in 3D and verified against Wärtsilä's vessel general arrangement drawings for clash detection — catching interferences between pipe, structure, cable tray and the surrounding vessel in the model, where a clash costs an hour to fix, rather than in the shop, where it costs a re-fabrication.

Step 4 — Pipe stress and support design. Especially on cryogenic and high-temperature skids, thermal growth and contraction must be analysed and accommodated. On the CNOOC Yingkou BOG skids, inlet gas arrives at –160 °C while temperature rises through the compression stages, creating large thermal gradients across the skid that were managed through expansion joints and flexible connections. Pipe supports, guides and anchors are designed from this analysis — not eyeballed.

Step 5 — Operability and maintainability review. Before fabrication, a layout review confirms that every operating valve is reachable, every instrument is readable, every item that needs periodic removal can be removed, and that escape and access routes around the skid are clear.

The maintainability test

A useful rule for any skid layout: can a technician do every routine maintenance task without removing a single item that was not the subject of the task? If pulling a pump rotor requires removing a length of permanent process pipe, the layout has failed the maintainability test and should be revised before metal is cut.

---

7. What Goes On a Skid: Component Integration

A gas or marine-fuel skid is a small plant in itself. The integration challenge is making a heterogeneous set of components — each with its own vendor, its own standard, its own failure modes — function as one tested unit. Here is what typically lives on a process skid and what integration each demands.

Process equipment

• Heat exchangers (intercoolers, aftercoolers, vaporizers, condensers, reboilers). Shell-and-tube or plate; require nozzle orientation that suits the pipe routing, support that allows thermal growth, and clearance for tube-bundle withdrawal. Lmart's gas-handling skids regularly carry intercoolers and aftercoolers integrated with the compression train.
• Separators and buffer drums (suction/discharge separators, knock-out drums, surge buffers). Pressure vessels in their own right, code-stamped, with level instrumentation and drain provisions. The CNOOC Yingkou BOG skid integrated suction/discharge separators with the compressor package; Lmart's CCUS-project work for an offshore platform integrated 1st- and 2nd-stage gas-liquid separator skids with after-cooler and flash-economizer skids.
• Pumps and compressors. The rotating heart of the skid. Demand precise alignment on a stiff frame, vibration isolation, lube-oil and seal-gas auxiliary systems, and pulsation control on reciprocating machines. Integration includes the entire auxiliary chain, not just the bare machine.
• Valves and actuators. Manual, pneumatic and motor-operated valves, plus emergency shutdown (ESD) valves on hazardous-fuel skids, all with operability access and actuator air/signal supply.

Piping, instrumentation and controls

• Process piping sized and routed from the stress analysis, with material matched to service (316L for cryogenic and methanol-wetted service; carbon and low-temp grades elsewhere).
• Instrumentation — pressure, temperature, level, flow transmitters; gas detectors on hazardous-fuel skids; local gauges. Tubing in SS316 on marine skids. All wired to junction boxes and continuity-checked at FAT.
• Electrical and control — local control panels, cable trays, cable glands, and the wiring that ties field devices to the panel and ultimately to the host control system. On hazardous-area skids, all electrical equipment is rated for the area classification (Ex-rated), a point that drives equipment selection on every FGSS and methanol skid.

The auxiliary systems that come with rotating equipment

The single most underestimated integration item on a gas skid is the auxiliary chain that a compressor or pump drags along with it. A bare reciprocating or screw compressor is a small fraction of the skid scope. Around it sits a lube-oil system (reservoir, pumps, coolers, filters, and the small-bore tubing that connects them), a seal-gas system on machines handling hydrocarbon or toxic gas, a cooling system (often integrated intercoolers and aftercoolers, as on the CNOOC Yingkou BOG skid), pulsation dampeners on reciprocating machines, and a dedicated control and protection package. Each auxiliary is itself a mini-skid of vessels, instruments and piping that must be integrated, routed and tested. On Lmart's LNG-terminal BOG skids, the package integrated the compressor, suction/discharge separators, intercoolers, aftercoolers, lube-oil system, seal-gas system, and a full control and safety system — all factory-assembled and tested as one unit while staying within transport dimensions. When estimating integration effort, the rule of thumb is that the auxiliaries and their connective piping often exceed the core machine in design hours.

Thermal management across the skid

Gas skids almost always have a thermal story. Cryogenic skids contract sharply on cooldown; compression skids build large temperature gradients across the stages. The CNOOC Yingkou BOG skid illustrates both ends on a single frame — inlet gas at –160 °C while temperatures rise through the compression stages — so the design had to absorb the differential growth between cold inlet piping and hot discharge piping through expansion joints and flexible connections, with supports that guide thermal movement rather than fight it. Ignore this and the first cooldown or the first hot run will load nozzles and supports beyond their design, cracking welds or pulling equipment out of alignment. Thermal management is therefore not a piping afterthought but a layout driver: cold and hot sections are positioned, supported and connected with their movement designed in from the start.

The integration mindset

The components are not the hard part — competent vendors supply competent equipment. The hard part is the connective tissue: the piping that links them, the supports that hold them through thermal cycling and ship motion, the instrument wiring that makes them controllable, and the layout that keeps them maintainable. On the TCS marine gas-handling skids, each skid integrated process vessels, compressors or pumps, heat exchangers, piping networks, instrumentation, electrical systems and control panels — and the design challenge Lmart called out was precisely ensuring that all those subsystems function together as an integrated unit, not as a collection of correctly specified parts.

Quotable: Specifying good components is necessary but not sufficient. A skid is integration risk, not equipment risk — the project is won or lost in the piping, supports, wiring and layout that turn a parts list into a working system.

---

8. Interface Management: The Discipline That Saves Projects

If the module boundary defines where the skid ends, interface management defines how it connects — to other skids and to the host plant or vessel — without surprises at site. This is the single discipline that most reliably separates a smooth integration from a painful one. It is also the discipline most often under-resourced, because it produces no visible hardware — only documents, checks and sign-offs.

The four interface types

Every connection that crosses a skid boundary falls into one of four categories, and each needs explicit definition and an owner on each side:

1. Mechanical / process interfaces. Piping tie-ins, flange faces and ratings, gasket specs, bolt-up torque, and the physical position and orientation of every termination. The classic failure mode is a flange that does not line up — wrong rating, wrong face, wrong elevation, or a few millimetres of misalignment that forces a field cut. Defined by a termination-point schedule with coordinates, ratings and orientation.

2. Electrical interfaces. Power supply (voltage, phase, load), cable entries, glanding, earthing, and the cable schedule that says which cable lands where. On hazardous-fuel skids, the Ex-rating of every penetration is part of the interface definition.

3. Control / instrumentation interfaces. The signal list — every hardwired I/O point and every fieldbus / network connection between the skid's local controls and the host control system. Signal type (4–20 mA, digital, Modbus, Profibus, etc.), tag, direction and the responsible party for each. This is where a skid that mechanically fits can still fail to commission because the engine maker's fuel-pressure setpoint, the system integrator's process logic, and the fabricator's instrument wiring do not agree.

4. Utility interfaces. Cooling water, instrument air, nitrogen, drains, vents — the supporting services the skid needs from the host, each with its own tie-in point, flow, pressure and quality requirement.

Why marine fuel skids make interfaces life-or-death

On a marine fuel system the control interface is not a convenience — it is a commissioning gate. The TGE LNG fuel system skids had to deliver LNG at precisely controlled pressure and temperature to the dual-fuel engine's gas admission valves. That required tight coordination between TGE's process design, the engine maker's fuel specification, and Lmart's mechanical fabrication — and as the project team explicitly noted, any interface mismatch can delay commissioning at the shipyard. The interface is where three companies' work has to meet exactly, on a date when the ship is waiting.

How to manage interfaces in practice

• One register, one owner. Maintain a single interface register (see Section 3) and assign a named engineer to own it. On the Höglund three-skid programme, Lmart did exactly this — a dedicated project engineer managing interfaces across the three skid types against the OEM's 3D model.
• Master 3D model as truth. All parties reference one 3D model. Clash detection (see Section 6) verifies physical interfaces; the signal list verifies control interfaces.
• Verify before shipment, not after. Inter-skid connections, electrical penetrations and control signals are verified during fabrication and at FAT — for the Höglund project, a final system-integration check was performed with all skids positioned together in the workshop before shipping, precisely so that interface mismatches surfaced in the shop and not on the berth.

Interface verification matrix

Interface type	Defined by	Verified at FAT by	Common failure if neglected
Mechanical / process	Termination-point schedule (coordinates, flange rating, orientation)	Dimensional check vs host GA; trial fit of inter-skid spools	Misaligned flange → field cut, schedule hit
Electrical	Cable schedule, glanding spec, Ex-rating	Continuity test, insulation test, Ex-equipment check	Wrong gland / rating → re-work, safety hold
Control / instrumentation	Signal list (tag, type, direction, party)	Loop check, signal simulation to host stub	Setpoint / protocol mismatch → commissioning stall
Utility	Utility tie-in schedule (flow, pressure, quality)	Connection check, flush/blow-through	Undersized service → derated performance

---

9. Factory Acceptance Test (FAT): Proving It Works Before It Ships

The Factory Acceptance Test is the moment a skid stops being a fabricated object and becomes a proven system. It is the single most valuable event in the modular delivery model, because it converts the entire integration effort into documented evidence — captured at one place, at one time, with the right people watching — instead of being discovered piecemeal during site commissioning. A skid that passes a rigorous FAT arrives at site as a known quantity. A skid that ships on a weak FAT arrives as a question.

What a FAT actually covers

A complete skid FAT is a layered programme, working from the components outward to the whole system:

1. Pressure / leak integrity testing
- Hydrostatic test of the pressure boundary to the code test pressure (1.3 × design per ASME VIII Div.1 / 1.43 × per PED, applied per the governing code), proving structural integrity.
- Pneumatic test where hydro is impractical (e.g. systems that must stay bone-dry), under controlled safety procedures.
- Helium leak / mass-spectrometer testing on critical joints for cryogenic and toxic-fluid service, where a hydrostatic test is not sensitive enough to the tiny leak rates that matter. On the Wärtsilä 91K reliquefaction skids, critical joints underwent helium leak testing in addition to standard hydrostatic testing; the TGE LNG skids ran a pneumatic-hydrostatic test followed by helium mass-spectrometer leak testing of the complete LNG circuit before shipment.

2. Functional testing
- Operating every valve, confirming actuator stroke and ESD action.
- Loop-checking every instrument from field device to local panel.
- Verifying interlocks, alarms, trips and the control logic that protects the system.

3. Performance / run testing
- Where the contract requires it, running rotating equipment (pumps, compressors) to confirm flow, head, vibration and bearing behaviour against the datasheet.
- Confirming system performance points — pressure, temperature, capacity — under simulated operating conditions.

4. Dimensional and interface verification
- Confirming the skid's overall dimensions, lift-lug positions, centre of gravity, and — critically — that every termination point matches the host general arrangement. This is the FAT step that prevents the field-fit nightmare.

5. Third-party witness
- For class and code work, the FAT is witnessed. A DNV / ABS / BV / CCS surveyor and (for PED) a Notified Body representative attend the hold points defined in the Inspection & Test Plan (ITP). On Lmart's marine projects a single project ITP covers both the class and pressure-code approval tracks, with the surveyor attending material inspection, fit-up, NDE, pressure test, dimensional check and FAT — reducing scheduling conflict between the two parallel approval streams.

The integration-test edge case

The most valuable FAT step on a multi-skid system is the one that is easiest to skip: testing the skids together. When a system is split into multiple modules that will only meet for the first time at site, positioning them together in the shop for a combined interface check is the closest you can get to a site-commissioning rehearsal. Lmart did this on the Höglund three-skid methanol programme — a final system-integration check with all three skids positioned together before shipping. The cost is a few days of shop floor space. The return is finding an interface mismatch where it costs nothing to fix.

FAT deliverables

A FAT is only as good as its documentation. The output should include: test certificates (hydro, pneumatic, helium), the loop-check and function-test records, the run-test data sheets, the dimensional report, the signed ITP with all hold points cleared, the as-built/red-line drawings, the welder and WPS traceability records, and the surveyor's/Notified Body's release. Bundled, this is the data book that lets the EPC contractor or shipyard accept the skid and plan its commissioning with confidence.

Quotable: FAT is where integration becomes evidence. A skid that passes a rigorous, witnessed FAT — including a combined multi-skid interface check — arrives at site as a known quantity; everything not proven at FAT becomes a risk discovered during commissioning, on the critical path.

---

10. Marine Gas Skids vs Onshore Petrochemical Skids

The same modular philosophy governs a marine fuel skid and an onshore petrochemical skid — but the constraints, codes and acceptance regimes diverge sharply. Understanding the difference is essential, because a design or procurement assumption carried from one world to the other is a reliable source of error. The table below summarizes the divergence; the discussion that follows explains why it matters.

Dimension	Onshore petrochemical skid	Marine gas / fuel skid (FGSS / CIP / reliquefaction)
Governing pressure code	ASME VIII Div.1/2, PED 2014/68/EU, or GB/T 150	Same pressure code plus classification society rules (DNV/ABS/BV/CCS)
Safety regime	Onshore process safety, local regulations	IMO IGF Code (low-flashpoint fuels), IGC Code (gas carriers), SOLAS
Dynamic loads	Static + seismic (site-specific)	Ship motion (roll/pitch/heave) + vibration + collision/grounding
Space envelope	Generous; defined by plot plan	Severe; defined by engine room / fuel-prep room and access hatches
Approval tracks	Pressure code (single track)	Dual track — class society and pressure code, run concurrently
Inspection witness	Owner/EPC + Notified Body (PED)	Class surveyor + Notified Body (PED), both attending the ITP
Coating / corrosion	Industrial coating per site corrosion class	Marine multi-coat epoxy, salt-spray rated
Materials (cryogenic)	As process demands	316L mandatory for cryogenic wetted parts, impact-tested at –196 °C

Why marine is harder

Dual-track certification. This is the defining difference. An onshore PED skid satisfies one code and one Notified Body. A marine gas skid must satisfy both a classification society (for structural integrity, vibration, fire safety, ship-motion loads) and the pressure code (PED or equivalent) — concurrently. As Lmart's project teams describe repeatedly across the Wärtsilä CIP, reliquefaction, methanol and TGE LNG programmes, this means two parallel approval tracks: class plan approval and workshop surveys alongside Notified Body involvement for pressure-equipment conformity — and any design change requires re-approval from both parties. The administrative and schedule burden of running two approval streams in lockstep is something onshore work simply does not carry.

The IGF / IGC safety overlay. Marine fuel systems on non-gas-carrier vessels must comply with the IMO IGF Code (International Code of Safety for Ships using Gases or Low-flashpoint Fuels), which imposes double-wall piping, gas detection, mechanical ventilation and emergency shutdown on the skid design. Gas carriers fall under the IGC Code. Critically, the regulatory frame is still maturing for the newest fuels: as of 2026 the IGF Code does not yet formally cover methanol, ethanol, LPG, ammonia or hydrogen — methanol/ethanol guideline revisions are targeted for approval in 2027 — so projects rely on IMO interim guidelines (such as MSC.1/Circ.1687 for ammonia, February 2025) and individual class-society fuel rules (DNV, 2025). A methanol skid built today is built to interim guidance plus DNV/ABS methanol fuel rules, not a settled code.

Hazardous-fuel specifics. Methanol is toxic (TLV-TWA 200 ppm), flammable (flash point 11 °C) and burns with an invisible flame — driving double-wall piping with inter-wall leak detection, Ex-rated electrical equipment, gas detection and automatic isolation. On the Höglund and Wärtsilä methanol skids, Lmart fabricated double-wall piping sections with leak-detection provisions between inner and outer walls and welded all wetted components with argon-purged GTAW to preserve corrosion resistance, with 100 % radiographic examination on pressure welds. Ammonia adds toxicity and material-compatibility demands that are pushing the next wave of skid design as the fuel enters service.

The onshore world is not "easy" — high-pressure, high-temperature and corrosive-service skids carry their own severe demands. But the marine gas skid is uniquely constrained by the combination of dual-track certification, the IGF/IGC overlay, ship-motion loads, and a space envelope measured in centimetres.

---

11. Certification: Class Society + Pressure Code Dual Track

Because dual-track certification is the defining challenge of marine gas skids, it deserves a closer look at how it is actually managed — because the difference between a smooth project and a stalled one is often nothing more than how well the two approval streams are synchronized.

The two tracks

Track A — Classification society (DNV, ABS, BV, CCS, LR, NK, KR). The class society approves the skid for service on a classed vessel. This covers structural design for ship-motion and accidental loads, vibration, fire safety, materials, and — for fuel systems — the IGF/IGC and class fuel-rule requirements. The process is: plan submission → plan approval → workshop survey at defined hold points → final survey and certification.

Track B — Pressure equipment code (PED 2014/68/EU via a Notified Body, or ASME VIII via an Authorized Inspector, or GB/T 150 domestically). This covers the pressure-retaining integrity of the vessels and piping: design calculation, material certification, welding qualification, NDE, and pressure test. For PED, a Notified Body conducts the conformity assessment.

The synchronization problem — and the fix

The risk is obvious: two independent approval bodies, each with its own hold points, its own document requirements, and its own surveyor calendar. Run them as separate streams and you get scheduling conflicts, duplicated inspections, and design changes that have to be re-approved twice.

Lmart's standard fix, applied across the Wärtsilä, TGE and Höglund marine programmes, is a single integrated ITP that maps both tracks onto one inspection-and-test plan. The class surveyor and the Notified Body representative attend overlapping hold points — material inspection, fit-up, NDE, pressure test, dimensional check, FAT — so that one inspection event satisfies both tracks where possible, and the two release chains advance in step rather than colliding. A dedicated project engineer owns the document flow to both bodies. The result, as the project teams report, is reduced scheduling conflict between the two approval streams and a single coherent release.

Material and welding rigour under dual track

Both tracks converge hardest on welding and materials, especially for cryogenic and hazardous service:

• Welder qualification. On the Wärtsilä 91K and TGE LNG skids, all welders held valid 316L cryogenic qualifications (GTAW root, SMAW/FCAW fill, with cryogenic endorsements).
• Weld traceability. Each weld carries a unique number traceable to the welder, the WPS, the filler-metal heat and the NDE records — a requirement both tracks demand and the data book must demonstrate.
• Impact testing. Production weld coupons are Charpy V-notch impact-tested at –196 °C for cryogenic service, with values confirmed above the project minimum.
• NDE coverage. Enhanced NDE — including 100 % radiography on pressure welds for methanol and cryogenic skids — beyond the code minimum, reflecting the consequence of failure in marine fuel service.

What the buyer should ask for

If you are an EPC contractor or shipyard accepting a marine gas skid, the certification documents you should require at handover are concrete and checkable:

• The class society design appraisal / type approval letter and the final survey report against the approved drawings.
• The PED Declaration of Conformity (or ASME Manufacturer's Data Report / U-1A, or GB/T 150 quality certificate) with the Notified Body / Authorized Inspector reference.
• The signed ITP with every hold point cleared and witnessed, showing the class and pressure-code signatures side by side.
• The weld map and traceability matrix linking every weld to welder, WPS, filler heat and NDE result.
• The material test certificates (EN 10204 3.1 / 3.2) for all pressure-retaining and cryogenic components, with impact-test results where required.
• The FAT report and data book (Section 9) as the integrating document.

A skid that cannot produce this set at FAT is not ready for release, regardless of how complete it looks on the floor. The documents are the deliverable as much as the steel.

Quotable: On a marine gas skid, certification is not a single gate at the end — it is two parallel approval tracks (class society + pressure code) that must be synchronized through one integrated ITP from day one. The fabricators who deliver on time are the ones who manage the two streams as one.

---

12. The Design / Interface / FAT Checklist

A practical, copy-and-use checklist for anyone specifying, designing, or accepting a modular gas or marine-fuel skid. Treat each unchecked box as an open risk.

Design checklist

• [ ] Transport envelope (road/sea), destination crane SWL, and destination access route confirmed before module boundary is fixed.
• [ ] Module boundaries drawn at functional seams, not through process loops; crossings minimized and counted.
• [ ] Structural frame designed for the full operating dead load plus the correct dynamic load case (seismic onshore; roll/pitch/heave/vibration/collision per class for marine).
• [ ] Frame stiffness verified for rotating-equipment alignment under operating load, not just strength.
• [ ] Coating system specified to the destination's corrosion class (marine multi-coat epoxy for offshore/marine).
• [ ] Lift lugs positioned for level lift; centre of gravity calculated; SWL verified (class-witnessed for marine).
• [ ] P&ID frozen before layout begins; full tagged equipment list with envelopes, weights, nozzle orientation.
• [ ] 3D model built; clash detection run against host general arrangement.
• [ ] Pipe stress / thermal analysis complete for cryogenic and high-temperature service; supports designed from it.
• [ ] Maintainability test passed: every routine task achievable without removing unrelated permanent items.
• [ ] Materials matched to service (316L for cryogenic and methanol-wetted; impact-tested where required).
• [ ] Hazardous-area classification applied; all electrical equipment Ex-rated where required.

Interface checklist

• [ ] Single interface / boundary register established, with a named owner.
• [ ] Every crossing classified (mechanical / electrical / control / utility) and owned on both sides.
• [ ] Termination-point schedule with coordinates, flange ratings and orientation issued.
• [ ] Cable schedule and glanding/Ex spec issued.
• [ ] Control signal list (tag, type, direction, responsible party) agreed across all parties — including the engine maker / host control system for fuel skids.
• [ ] Utility tie-in schedule (flow, pressure, quality) confirmed against host supply.
• [ ] Master 3D model designated as the single source of truth for all parties.

FAT checklist

• [ ] Pressure / leak test plan defined: hydrostatic to code; pneumatic where required; helium leak test on cryogenic/toxic critical joints.
• [ ] Functional test: all valves, ESD action, loop checks, interlocks/alarms/trips.
• [ ] Performance / run test for rotating equipment where contract requires.
• [ ] Dimensional and termination-point verification against host GA.
• [ ] Combined multi-skid interface check (skids positioned together) for multi-module systems.
• [ ] Witness arranged: class surveyor + Notified Body / Authorized Inspector at ITP hold points.
• [ ] Single integrated ITP covering both class and pressure-code tracks.
• [ ] Data book compiled: test certificates, loop/function records, run data, dimensional report, weld/WPS traceability, as-built drawings, surveyor/NB release.

---

13. Real Project Mapping

The principles above are not abstractions — they were learned and proven on real skid programmes. The table maps each design lesson to a project where it mattered. (Customer-public references only; vessel and hull identifiers omitted.)

Design / interface / FAT lesson	Where it was proven
Module boundary at clean functional seams	Höglund methanol system split into bunker / FPR (fuel preparation room) / WGS (waste gas scrubber) skids — each a self-contained function
Dedicated interface owner + master 3D model	Höglund three-skid programme — a dedicated project engineer managed all inter-skid piping, electrical and control interfaces against the OEM 3D model
Combined multi-skid interface check before shipment	Höglund — final system-integration check with all three skids positioned together in the workshop
Control interface as commissioning gate	TGE LNG fuel system (PCTC) — LNG delivered at precise pressure/temperature to dual-fuel engine gas-admission valves; interface mismatch delays shipyard commissioning
3D clash detection vs vessel GA	Wärtsilä 91K reliquefaction & booster-pump skids — all piping 3D-modelled and verified against vessel general arrangement
Cryogenic welding + helium leak FAT	Wärtsilä 91K reliquefaction — 316L cryogenic welds, coupons impact-tested at –196 °C, critical joints helium leak-tested
Full shop pre-assembly + system leak test	TGE LNG — complete skid assembled with piping, instrumentation, cable trays and insulation supports; pneumatic-hydrostatic + helium mass-spec leak test before shipment
Single integrated ITP for dual-track approval	Wärtsilä CIP / reliquefaction / methanol and TGE LNG — one ITP covering DNV class + PED, surveyor and Notified Body at shared hold points
Double-wall piping + Ex + 100 % RT for methanol	Höglund & Wärtsilä (CIMC Raffles) methanol skids — double-wall piping with inter-wall leak detection, argon-purged GTAW, 100 % radiography on pressure welds
FPSO dynamic-load frame design	Jord NAG module, Abigail Joseph FPSO — frame designed for roll up to 22.5°, pitch up to 12.5° plus wave-induced accelerations
Destination access caps module size	Alfa Laval marine skids — modules sized to pass through engine-room access hatches under limited overhead clearance
Multi-system integration on one skid	TCS marine gas-handling and CNOOC Yingkou BOG skids — vessels, compressors/pumps, exchangers, instrumentation, electrical and controls integrated and tested as one unit
Onshore code vs marine class divergence	CNOOC Yingkou BOG (GB/T 150, onshore terminal) vs the marine class+PED programmes — same modular logic, different code and acceptance regime

What ties these together is consistency of method: freeze the design, draw clean boundaries, own the interfaces against a master model, prove the package at a witnessed FAT, and run the dual approval tracks as one. Lmart maintains 600+ PQR welding qualifications and an 8,000 m² assembly workshop precisely to execute this method across cryogenic marine, hazardous-fuel and onshore petrochemical skids — with full in-house design, fabrication, assembly, testing and documentation.

---

14. Frequently Asked Questions

What is the difference between a skid and a module?
The terms overlap, but a useful distinction is scale and structure. A skid is a single integrated package on one structural frame — process equipment, piping, instrumentation and controls — sized to be transported and lifted as one piece. A module is often larger, may comprise several skids and structural steel, and is typically associated with offshore topsides or large onshore plant blocks. In practice, a marine fuel system might be delivered as several skids (bunker, fuel preparation, scrubber) that together form one functional system, while an FPSO topsides arrangement is built from modules. The engineering disciplines — boundaries, interfaces, FAT — apply at both scales.

What does a Factory Acceptance Test (FAT) actually prove?
A FAT proves that the assembled skid functions as a system before it ships — not just that its parts are individually correct. A complete skid FAT layers pressure and leak integrity testing (hydrostatic, pneumatic, and helium leak testing on cryogenic or toxic joints), functional testing (valves, ESD, instrument loops, interlocks), performance/run testing of rotating equipment where required, and dimensional verification of every termination point against the host general arrangement. For class and code work the FAT is witnessed by a classification surveyor and a Notified Body or Authorized Inspector. The deliverable is a data book that lets the buyer accept the skid and plan commissioning with confidence.

Why do marine fuel skids need both class society and pressure code approval?
Because they serve two distinct safety regimes at once. The classification society (DNV, ABS, BV, CCS and others) governs the skid's fitness for service on a classed vessel — structural design for ship-motion and accidental loads, vibration, fire safety, and IGF/IGC fuel-system requirements. The pressure code (PED, ASME VIII or GB/T 150) governs the integrity of the pressure-retaining vessels and piping. These are separate approvals from separate bodies, and a marine gas skid must satisfy both concurrently. The practical key is a single integrated Inspection & Test Plan that synchronizes the two tracks so design changes, inspections and releases advance in step rather than colliding.

How do you keep interfaces from causing problems at site?
By treating interface management as a named, owned discipline rather than an assumption. Establish a single interface register listing every connection that crosses a skid boundary, classified as mechanical/process, electrical, control or utility, with an owner on each side and a defined termination detail. Reference one master 3D model across all parties, run clash detection to verify physical interfaces, agree a control signal list (including with the engine maker or host control system on fuel skids), and verify everything at FAT — ideally including a combined check with multi-skid systems positioned together in the workshop before shipment. The goal is that interface mismatches surface in the shop, where they are cheap to fix, not on the berth, where they sit on the critical path.

What determines the maximum size of a skid module?
Three hard constraints, all of which must be fixed before the module boundary is finalized: the transport envelope (road or sea shipping limits), the lifting capacity of the crane available at the destination, and the access route at the destination — which on marine projects often means fitting through an engine-room hatch under limited overhead clearance. The largest single piece that can be physically transported, lifted and manoeuvred into final position caps module size, frequently below what process logic alone would suggest. When the cap is binding, the designer either splits the module (adding interfaces), defers equipment to field installation, or ships removable items separately for site re-fit.

Which materials are standard for cryogenic and methanol marine fuel skids?
For cryogenic LNG/BOG service (down to roughly –162 °C), all wetted process components are austenitic stainless steel, typically AISI 316L, with welding procedures qualified for cryogenic service and production weld coupons Charpy V-notch impact-tested at –196 °C. For methanol fuel service, wetted components are likewise 316L, verified for methanol compatibility, with double-wall piping in enclosed spaces and argon-purged GTAW welding to preserve corrosion resistance and 100 % radiography on pressure welds. Structural frames are carbon steel with a marine coating system. Onshore terminal BOG skids may use low-temperature carbon steels (e.g. A333 Gr.6) and domestic vessel steels (Q345R) under GB/T 150 where the service and code permit.

Can a modular skid be retrofitted to an existing vessel or plant?
Yes, and retrofit is a growing driver — dual-fuel conversions of existing vessels are a significant part of the 2025–2026 marine fuel market, with OEMs developing skid-based fuel-supply solutions specifically for conversion programmes (Nikkiso CE&IG, 2025). Retrofit raises the interface and access challenges to their maximum: the skid must fit an existing, already-congested space; the tie-ins must match existing systems whose as-built condition may differ from drawings; and the access route is whatever the existing vessel or plant provides. A laser scan of the installation space and meticulous interface verification against as-built conditions become essential, because the margin for field adjustment is even smaller than on a newbuild.

---

15. Related Reading

• Gas Compressor Package Selection: Recip vs Screw vs Centrifugal
• Air Products Korea Hydrogen: PSA & Shift Reactors
• Pressure Vessel Hydrotest: A Complete Guide
• EPC Procurement Hub — pillar guide for buyers

---

Author: Qiangbin Chu, Suzhou Lmart Energy Equipment Co., Ltd. Lmart is an ASME U-Stamp manufacturer delivering modular gas-processing and marine-fuel skids, pressure vessels, heat exchangers and compression packages to EPC and shipyard projects worldwide, with classification approvals (DNV/ABS/BV/CCS) on a project basis. This article is an engineering reference and does not disclose project-specific commercial terms.

---

Source Anchors (industry hotspots referenced)

• Nikkiso Clean Energy & Industrial Gases — next-generation marine fuel cryogenic pump skids for dual-fuel conversions (2025): https://nikkisoceig.com/news/nikkisos-next-generation-marine-fuel-cryogenic-pump-skids-to-support-dual-fuel-vessel-conversion
• Spectra by MHI — flexible marine fuels (LNG / methanol / ammonia) cutting shipping emissions (2025): https://spectra.mhi.com/energy-transition/how-flexible-marine-fuels-can-help-cut-emissions-from-shipping
• DNV — IMO CCC 11 interim guidelines for low-flashpoint / alternative fuels; IGC Code amendment timeline (2025): https://www.dnv.com/news/2025/imo-ccc-11-interim-guidelines-for-hydrogen-as-fuel-completed/
• BIMCO — IMO interim guidelines for the safety of ships using ammonia as fuel, MSC.1/Circ.1687 (27 March 2025): https://www.bimco.org/news-insights/bimco-news/2025/03/27-ammonia/