Gas Compressor Package Selection 2026: Reciprocating vs Screw vs Centrifugal for Petrochemical & Marine BOG Service
Gas Compressor Package Selection 2026: Reciprocating vs Screw vs Centrifugal for Petrochemical & Marine BOG Service
In December 2025, LNG Industry ran a special report titled "Managing the pressure: LNG boil-off gas compression," and the timing was not an accident. Earlier that year IMO's EEDI Phase 3 took effect (January 2025), and MEPC 80 had already locked the methane-slip ceiling at 3 g/kWh for gas-fuelled ships — two rules that quietly rewrite how boil-off gas (BOG) gets handled on every new LNG carrier and gas-fuelled vessel (LNG Industry; DNV).
What does that mean on the ground? Boil-off can no longer be vented or burned the way it used to be. It has to be compressed, re-routed, re-injected, or reliquefied — which puts the gas compressor package at the center of both the emissions ledger and the project schedule. And the market has noticed: the BOG compressor segment is estimated at USD 1.62 billion in 2025, heading for USD 2.21 billion by 2030 at a 6.4% CAGR, with petrochemicals and industrial users growing fastest at 8.2% (Mordor Intelligence).
From where we sit at Suzhou Lmart Energy Equipment, the regulatory pressure has a very practical consequence: the compressor selection decision — reciprocating, screw, or centrifugal — is being made earlier, scrutinized harder, and tied more tightly to skid integration than it was five years ago. We have built propylene BOG packages for petrochemical tank farms, ammonia compression units for chemical plants, fuel-gas booster skids, and marine BOG packages for a 14,000 TEU LNG-fuelled container ship. The pattern is consistent: teams that map flow and pressure ratio to the right machine family up front avoid the expensive re-spins that show up four months into fabrication.
This guide gives you the engineering basis for that decision — the flow-versus-pressure-ratio maps, the API 617/618/619 framing, the BOG-specific reasoning, the skid-integration scope, and a decision tree you can actually use. Where it helps, we map the logic to real packages we have delivered.
Bottom line up front: there is no universally "best" compressor. There is the machine that matches your duty envelope. Reciprocating wins on high pressure ratio and low flow. Screw owns the messy middle — variable composition, wet gas, wide turndown, moderate pressure. Centrifugal takes over when flow is large, steady, and clean. Most of the hard calls live on the boundaries, and most of those boundaries land squarely on the screw machine.
Table of Contents
- Why Compressor Selection Decides the Project, Not Just the Equipment
- The Three Families: A Working Mental Model
- Reciprocating Compressors (API 618): High Pressure Ratio, Low Flow
- Screw Compressors (API 619): The Versatile Middle
- Centrifugal Compressors (API 617): Large, Steady, Clean Flow
- Head-to-Head: The Master Comparison Table
- BOG Compressors: A Service Class of Their Own
- The Compressor Package / Skid: What Actually Ships
- Drivers, Variable-Speed Control, and the Electric-Drive Shift
- Maintenance, Reliability, and Total Cost of Ownership
- Common EPC Procurement Pitfalls (and How to Avoid Them)
- The Selection Decision Tree
- Real Project Mapping: Which Case Used Which Machine, and Why
- Standards, Certification & Documentation
- Lmart's Gas Compression Capability
- Conclusion
- FAQ
- Related Reading
Why Compressor Selection Decides the Project, Not Just the Equipment
Engineers tend to treat compressor selection as an equipment-datasheet exercise — pick the machine, fill in the curve, move on. In EPC and shipbuilding reality, the choice cascades into almost everything downstream: skid footprint and weight, electrical load, vibration and pulsation mitigation, cooling-water and lube-oil utilities, control-system complexity, maintenance access, spare-parts strategy, and the inspection plan that governs how long fabrication takes.
Get the machine family wrong and you do not just swap a compressor. You re-do the skid GA, re-size the buffer drum, re-rate the motor, re-route the piping, and re-open the ITP. We have watched projects lose months exactly this way — not because the compressor failed, but because it was selected against the wrong duty assumption.
Three forces make 2026 a harder selection year than most:
- Methane-slip and efficiency rules. MEPC 80's 3 g/kWh methane cap and EEDI Phase 3 push marine BOG handling toward recovery and reliquefaction rather than venting or gas-combustion-unit burning. That changes the duty: instead of an intermittent "burn the excess" service, you now have a continuous recompression-and-return service with a wider turndown (DNV).
- Electric-drive migration. The market's structural drivers now explicitly include "the rapid shift toward electric-drive compression packages" (Mordor Intelligence). Variable-frequency drives change the turndown story for every machine family and tilt some borderline calls.
- Petrochemical capacity buildout. Propylene, polypropylene, and LPG tank farms keep coming online across Asia-Pacific, which leads the BOG compressor market at 36.2% share. Tank-farm vapor recovery is a different duty from marine cargo BOG, and it favors different machines.
The job of this guide is to make the family-level call defensible before the datasheet stage — so the rest of the package falls into place instead of fighting you.
The Three Families: A Working Mental Model
Before the standards and the curves, here is the mental model we use internally when a new RFQ lands.
Every positive-pressure machine does one of two things to raise pressure. Positive-displacement machines (reciprocating and screw) trap a fixed volume of gas and squeeze it — so they deliver pressure almost regardless of flow, which makes them good at high pressure ratios and tolerant of variable conditions. Dynamic machines (centrifugal) accelerate gas with an impeller and convert velocity to pressure — so they love large, steady, unbroken flow and lose their footing when flow swings or gets dirty.
That single distinction explains most of the selection logic:
- If you need a lot of pressure rise out of a modest flow, you want positive displacement. Reciprocating first, screw second.
- If you have huge, stable flow and a moderate pressure ratio, you want dynamic. Centrifugal.
- If your gas is wet, dirty, variable in composition, or swings in flow, you lean toward screw, which shrugs off conditions that would foul a centrifugal or hammer a reciprocating valve set.
The API standards formalize this. As the industry literature puts it plainly: API 617 covers centrifugal and axial machines, API 618 covers reciprocating machines, and API 619 covers rotary/screw machines (Inspenet; Power Magazine).
A useful rule-of-thumb for per-stage pressure ratio, from the same body of literature: roughly 3.5 for centrifugal, 4.5 for reciprocating, and 6.0 for screw per stage (Power Magazine). That number alone resolves a surprising fraction of selection arguments — when someone insists a single-casing centrifugal can hit a 10:1 ratio, the per-stage limit tells you it cannot, not without multiple casings or intercooling that blow up the footprint.
Keep this model in your head as we go machine by machine.
Reciprocating Compressors (API 618): High Pressure Ratio, Low Flow
How it works
A reciprocating compressor uses a piston driven by a crankshaft to compress gas in discrete strokes inside a cylinder. Suction and discharge valves open and close with each stroke. Because the piston physically displaces a fixed swept volume, the machine builds whatever pressure the system demands — within mechanical limits — and the flow it delivers is set by speed and cylinder geometry rather than by the downstream pressure.
That is its defining strength and its defining weakness.
Where it wins
- High pressure ratio. Reciprocating machines deliver ultra-high pressure and can compress nearly anything that finds its way through the pipes (Projectmaterials). When the duty calls for a large pressure rise from a relatively small flow — hydrogen makeup, high-pressure injection, certain BOG recompression-and-return duties — reciprocating is often the only machine that fits in a single, sane footprint.
- Composition flexibility. It tolerates a wide range of molecular weights and gas compositions without re-wheeling, which a centrifugal cannot match.
- Efficiency at the design point. For the right service, isentropic efficiency is high.
Where it costs you
- Pulsation. Discrete strokes create pressure pulsation in the suction and discharge piping. API 618 (and its companion API 688) drive a formal pulsation and mechanical-response analysis to keep that pulsation from cracking piping or exciting resonances. The Southwest Research Institute's API 618/688 pulsation analysis exists precisely because untreated reciprocating pulsation is a real failure mode (SwRI). The 6th edition of API 618 and 2nd edition of API 688 were under industry review through 2025 (EFRC/recip.org).
- Maintenance load. Valves, rings, and packing wear. Plan for it.
- Temperature ceiling. API 618 reciprocating machines carry a discharge-temperature limit around 350°F (≈177°C), driven by valve life (Power Magazine).
- Footprint and vibration. Reciprocating mass means foundation and vibration considerations — a real constraint on a vibration-sensitive marine skid.
Capacity control: how reciprocating machines turn down
One reason reciprocating compressors stay relevant in variable-flow service is the richness of their capacity-control toolkit. You can unload cylinder ends with valve unloaders, add clearance pockets that change the effective swept volume, run stepped (multi-step) unloading, recycle gas, or — increasingly — drive the whole machine with a variable-frequency drive. In BOG service where flow swings hard, a combination of step unloading plus VFD lets a reciprocating package follow a wide turndown without the surge worry a centrifugal carries. The trade-off is mechanical complexity: every unloading device is one more thing to maintain and one more failure mode to instrument.
Single-acting vs. double-acting, and stage count
Process reciprocating machines are usually double-acting (compressing on both piston strokes) and frequently multi-stage with intercooling, because intercooling between stages both controls discharge temperature and improves efficiency by moving compression closer to isothermal. Each added stage means another cylinder, another intercooler, another separator, and more skid real estate — which is exactly why the per-stage ratio rule matters at the selection stage. A duty that needs three reciprocating stages is a very different skid (and price, and footprint) from one that needs two.
The honest read
Reciprocating is the high-pressure specialist. If your pressure ratio is genuinely high and your flow is modest, nothing else competes on a single skid. The price of admission is a disciplined pulsation study, a maintenance plan that respects valve life, and a clear-eyed view of how many stages the duty really needs. When EPC teams get burned by reciprocating machines, it is almost never the compression principle that failed — it is an under-scoped pulsation study, an optimistic single-stage assumption, or a maintenance plan that ignored valve and packing replacement intervals.
Screw Compressors (API 619): The Versatile Middle
How it works
A screw compressor uses two intermeshing helical rotors. Gas drawn into the lobe spaces is progressively reduced in volume as the rotors turn and the gas moves axially toward discharge. Oil-injected (oil-flooded) screws inject lubricant that seals the rotor clearances, removes heat of compression, and lubricates — which is why they handle the heat and the sealing so gracefully.
Where it wins
- The messy middle. API 619 machines deliver constant flow regardless of pressure, which makes them ideal for low-to-medium pressure services, waste-gas recovery, process gas with high viscosity, and gas that is wet or carries entrained liquid (Inspenet). This is exactly the duty most petrochemical BOG and tank-farm vapor recovery services live in.
- Highest per-stage pressure ratio. At roughly 6.0 per stage, screw machines can hit a high ratio in a single casing where a centrifugal would need multiple stages (Power Magazine). Our Hengli ST-PP propylene BOG package achieved a 300:1 overall compression ratio — taking propylene from near-atmospheric cryogenic conditions up to 1.8 MPa(g) discharge — using a single-casing two-stage screw compressor. A centrifugal could not have done that on one skid; a reciprocating would have struggled with the wet, transient suction.
- Turndown and stability. Screws tolerate suction-pressure fluctuation and flow swings far better than centrifugals. On the Hengli package we held ±0.5 kPa suction-pressure fluctuation tolerance with adaptive speed control across a 2,800–4,500 RPM range. That kind of turndown is precisely what BOG service — with its voyage-profile or tank-pressure-driven flow swings — demands.
- Compact and economical. The design is simple, compact, and economical compared with the alternatives (Inspenet) — a real advantage on a space-constrained skid.
Where it costs you
- Oil carryover. Oil-injected screws put lubricant into the gas stream. You manage it with multi-stage coalescing oil separation, but for ultra-clean process gas you must engineer the separation train carefully.
- Temperature ceiling. API 619 oil-flooded machines carry a discharge-temperature limit around 300°F (≈150°C), set by the lubricant (Power Magazine).
- Efficiency vs. centrifugal at large clean flow. When flow is large, steady, and clean, a centrifugal will usually beat a screw on efficiency.
Oil-injected vs. oil-free screw
Two screw families exist and the choice is consequential. Oil-injected (oil-flooded) screws — the API 619 workhorse — inject lubricant for sealing, cooling, and lubrication, which is what gives them their wide-turndown, wet-gas tolerance. They put oil into the gas and need a separation train. Oil-free (dry) screws keep lubricant out of the compression chamber using timing gears and tighter clearances; they protect process purity but run hotter, cost more, and tolerate wet/dirty gas less gracefully. For petrochemical and tank-farm BOG — where the gas is going back into a process that already contains hydrocarbons and trace oil is acceptable — oil-injected is almost always the right and most economical answer. Where the recovered gas feeds a purity-critical downstream, you either accept an engineered multi-stage coalescing separation train on an oil-injected machine or pay for oil-free.
The slide valve and built-in volume ratio
A detail that separates a good screw selection from a mediocre one is the built-in volume ratio (Vi) and whether the machine has a slide valve. The Vi is fixed by the rotor geometry and sets the pressure ratio at which the machine is most efficient. Run a fixed-Vi machine far from its design pressure ratio and you pay an efficiency penalty through over- or under-compression. A slide valve lets the machine vary capacity (and, on variable-Vi designs, the internal volume ratio) so it stays efficient across a range of conditions — which is exactly why screws hold efficiency across the wide turndown that BOG demands. On the Hengli propylene package, the combination of adaptive speed control across 2,800–4,500 RPM and the screw's inherent turndown is what delivered 82.5% isentropic efficiency against a ~78% benchmark.
The honest read
If your service does not clearly belong to reciprocating (very high ratio, low flow) or centrifugal (very large, clean, steady flow), it almost certainly belongs to screw. The screw machine owns the boundary cases — wet gas, variable composition, wide turndown, moderate-to-high pressure, compact footprint. That is why so much petrochemical BOG and tank-farm vapor recovery is screw-based, and why Lmart's API 619 oil-injected screw package is our most-requested gas-compression configuration. The two things that most often go wrong with screw selection are an under-sized oil-separation train (oil carryover into a downstream that could not tolerate it) and a fixed-Vi machine specified for a duty whose pressure ratio drifts far from the design point.
Centrifugal Compressors (API 617): Large, Steady, Clean Flow
How it works
A centrifugal compressor spins an impeller at high speed, throwing gas radially outward and converting that velocity into pressure in a diffuser. It is a dynamic machine: it does not trap a fixed volume, so its pressure rise depends on flow. The performance curve, the surge line, and the choke point all matter in ways they simply do not for positive-displacement machines.
Where it wins
- Large, steady flow. API 617 applies to centrifugal, axial, and mixed-flow machines in continuous service across refineries, petrochemical plants, gas pipelines, and gas processing, handling large gas flows at moderate-to-high pressures (API 617 8th ed.). Centrifugals love large, steady, unbroken flow and run reliably under those conditions (Projectmaterials).
- Oil-free gas path. No lubricant contacts the process gas in a typical centrifugal — important when product purity is non-negotiable.
- Low maintenance, high availability. Few wearing parts in the gas path means long run times between overhauls.
- Market momentum in BOG. Centrifugal units already control 51.5% of the BOG compressor market and are growing — driven heavily by large onshore LNG terminals and big marine reliquefaction trains (Mordor Intelligence).
Where it costs you
- Sensitive to flow swings. Drop below the surge line and the machine surges — a destructive instability. BOG service with wide flow turndown is exactly where centrifugals get nervous, which is why marine reliquefaction trains pair them with careful anti-surge control and recycle.
- Composition sensitivity. Change the molecular weight materially and the curve shifts; centrifugals are not as forgiving of variable composition as positive-displacement machines.
- Per-stage ratio limit. At roughly 3.5 per stage, high overall ratios require multiple stages/casings (Power Magazine).
- Temperature headroom. The flip side of the ratio limit is the highest temperature ceiling of the three, around 450°F (≈232°C), set by rotor metallurgy (Power Magazine).
Surge, choke, and why the operating window matters
The single most important concept in centrifugal selection is the operating window between surge and choke. Surge is a flow reversal — the gas momentarily flows backward through the machine — that occurs when flow drops below the surge line; it is violent, damaging, and must be prevented. Choke (stonewall) is the opposite limit, where flow can rise no further. Everything useful happens between them. A centrifugal selected for BOG must either operate comfortably inside that window across its whole duty range, or carry an anti-surge control system with a recycle/hot-gas-bypass loop that opens to keep flow above the surge line during turndown. That recycle loop is not free: it costs energy whenever it is open, which is precisely the penalty that makes centrifugals less attractive for wide-turndown tank-farm BOG and more attractive for steady large-terminal duty.
Single-shaft vs. integrally geared
For BOG and process duty you will most often see single-shaft (beam-style) centrifugals, but integrally geared machines — multiple pinions on a central bull gear, each stage at its optimal speed with intercooling between stages — are common where high efficiency across several stages matters. Integrally geared machines pack a high overall ratio into a compact footprint at strong efficiency, which is why they show up on larger LNG and air-separation duties. The trade-off is mechanical complexity and a tighter dependence on clean, well-conditioned gas.
The honest read
Centrifugal is the large-flow, clean-gas, steady-service machine. On a big onshore LNG terminal BOG train or a large marine reliquefaction plant, it is often the right and most efficient answer. On a petrochemical tank farm with wide flow swings and variable composition, it usually is not — that is screw or reciprocating territory. The most common centrifugal selection error in BOG work is ignoring how much time the machine will actually spend in turndown: a centrifugal that looks efficient at the design point can spend half its life on recycle, burning energy to avoid surge, when a screw would have followed the flow without complaint.
Head-to-Head: The Master Comparison Table
| Attribute | Reciprocating (API 618) | Screw (API 619) | Centrifugal (API 617) |
|---|---|---|---|
| Working principle | Positive displacement, piston | Positive displacement, helical rotors | Dynamic, impeller |
| Flow regime | Low to moderate | Low to medium-high | Large, steady |
| Per-stage pressure ratio (rule of thumb) | ~4.5 | ~6.0 (highest) | ~3.5 |
| Pressure capability | Highest | Moderate to high | Moderate to high (multi-stage) |
| Discharge temperature limit | ~350°F / ~177°C (valve life) | ~300°F / ~150°C (lubricant) | ~450°F / ~232°C (rotor metallurgy) |
| Variable composition tolerance | Excellent | Excellent | Poor |
| Wet / dirty gas tolerance | Moderate (valve risk) | Excellent (oil-flooded) | Poor |
| Flow turndown | Good (step + speed) | Excellent (speed) | Limited (surge risk) |
| Pulsation | Significant — API 618/688 study required | Low | Negligible |
| Vibration / footprint | Heaviest, reciprocating mass | Compact | Compact for the flow |
| Oil in gas path | Crankcase, generally separated | Oil-injected — coalescing separation needed | Oil-free path |
| Maintenance load | Higher (valves, rings, packing) | Moderate | Lowest |
| Typical BOG fit | High-ratio recompression/return, low flow | Petrochem & tank-farm BOG, wide turndown | Large LNG terminal / marine reliquefaction trains |
| Lmart project example | (high-ratio injection duties) | Hengli ST-PP propylene BOG (300:1, 630 kW) | (large terminal trains) |
Quotable takeaway: Choose the machine by the duty envelope, not by habit. Reciprocating for high ratio and low flow; centrifugal for large, clean, steady flow; screw for everything wet, variable, or in between — which in petrochemical and marine BOG is most of it.
BOG Compressors: A Service Class of Their Own
BOG — boil-off gas — is what evaporates off a cryogenic liquid (LNG, propylene, propane, LPG, ethylene) because no insulation is perfect and ambient heat always leaks in. Left alone, BOG raises tank pressure until you either vent it (a loss, and now an emissions problem) or relieve it through a safety valve (worse). The job of a BOG compressor is to take that low-pressure, often-cold vapor and do something useful with it: re-inject it into the process, feed it to a fuel system, or compress and cool it for reliquefaction back into the tank.
Why BOG is hard
BOG service punishes a poorly chosen compressor because the duty is rarely steady:
- Flow swings constantly. On an LNG-fuelled vessel, BOG flow varies with voyage profile, tank-pressure-management strategy, ambient conditions, and operating mode. The package is less about a single design point and more about stable performance across a duty envelope with frequent transients — exactly the framing we applied on the 14,000 TEU container-ship BOG package.
- Suction is cold and sometimes wet. Cryogenic suction (the Hengli propylene BOG ran at −43°C service) demands materials and design rated for it — ASME B31.3 piping compliance for low-temperature service was a hard requirement there.
- Composition can drift. Mixed hydrocarbons, varying molecular weight.
- Turndown is wide. You must run gracefully from near-zero boil-off to peak, without surging (centrifugal) or hammering valves (reciprocating).
Why petrochemical & marine BOG leans positive-displacement
Put those four together — variable flow, cold/wet suction, drifting composition, wide turndown — and you have a near-perfect description of where dynamic centrifugal machines struggle and positive-displacement screws excel. That is why so much petrochemical tank-farm BOG and a large share of marine cargo BOG runs on screw or reciprocating machines rather than centrifugal, even though centrifugal dominates the large-terminal end of the same market.
The split is real and worth stating clearly:
- Large onshore LNG terminals and big marine reliquefaction trains → often centrifugal (large, relatively steady flow, efficiency-driven). This is the 51.5% centrifugal market share at work (Mordor Intelligence).
- Petrochemical tank-farm vapor recovery (propylene, propane, LPG, ammonia) → predominantly screw, sometimes reciprocating (wide turndown, wet/cold suction, compact skids). This is the fastest-growing segment at 8.2% CAGR (Mordor Intelligence).
Reliquefaction is reshaping the marine duty
The regulatory squeeze — MEPC 80's methane cap, EEDI Phase 3, the DNV guidance on counting BOG-as-fuel emissions — is steering marine owners away from venting and toward recovery and reliquefaction (DNV). Wärtsilä alone reports more than 75 BOG reliquefaction plants in the global LNG carrier fleet (Wärtsilä). Reliquefaction turns an intermittent "deal-with-the-excess" service into a continuous recompress-cool-return service, which generally means more running hours, tighter turndown control, and higher availability requirements for the compressor package — a shift that favors machines with graceful turndown and robust control logic.
For project teams this has a concrete implication: the marine BOG package you specify in 2026 should assume continuous reliquefaction-style duty and a wide turndown, not the older intermittent-burn assumption. Specify the compressor — and the anti-surge or speed-control system — accordingly.
The Compressor Package / Skid: What Actually Ships
A bare compressor is not what an EPC buyer or a shipyard receives. What ships is a compressor package — a skid-mounted, pre-assembled, factory-tested module that integrates the machine with everything it needs to run safely and hand off cleanly. Treating the package, not the compressor, as the unit of selection is what separates a smooth integration from a four-month interface fight.
What a complete package integrates
A typical Lmart gas compressor package skid includes:
- The compressor and driver — electric motor (increasingly VFD), with explosion-proof rating where required (the Hengli propylene package was Ex-certified; the Huachang ammonia unit was ExdIIBT4).
- Suction scrubber / separator (buffer drum) — removes entrained liquid before the gas hits the machine; on reciprocating skids the buffer drum also helps damp pulsation.
- Intercooler and aftercooler — control discharge temperature between and after stages; on the BOG booster skids these regulate gas temperature ahead of reinjection.
- Lube-oil station and, for screws, oil-separation train — multi-stage coalescing separation to control oil carryover.
- Pulsation dampeners (reciprocating) — sized through the API 618/688 study.
- Skid frame and piping — offshore-rated and vibration-resistant where the service demands it.
- Instrumentation and control system — PLC or DCS with anti-surge/speed control, protection logic, and emergency shutdown (ESD).
- Interface terminals — clearly demarcated electrical/control I/O, utilities (cooling water, instrument air), drains/vents philosophy, and safety signals to the plant or vessel.
Why skid integration controls schedule
The single most underrated source of EPC schedule risk in compression projects is the interface boundary — the "what the vendor provides vs. what the site/shipyard provides" line. A package built by a vendor who maps that boundary early ships as a plug-and-play module; one that doesn't generates RFIs and rework on site.
We have proven the schedule upside repeatedly. The BOG Booster Skid for Eni's NGUYA FLNG platform (Congo LNG project) was delivered in 35 days — pre-assembled, tested, ABS-certified, ready to integrate — because of pre-material procurement, parallel assembly and welding, and continuous in-process QA/QC. That delivery speed is not a compressor property; it is a packaging-discipline property. The same logic carried the Wison (Nantong) / Burckhardt Compression BOG booster skid (25 tons, ABS-certified, 35-day build).
For marine work, packaging discipline is also quality preservation: corrosion protection, tagged lifting points, and shipment documentation aligned to shipyard receiving procedures — so the as-built condition survives the trip to the yard. We treated this as part of the deliverable on the 14,000 TEU container-ship BOG package.
Drivers, Variable-Speed Control, and the Electric-Drive Shift
The compressor is only half the machine. The driver and the control system decide how the package behaves across its duty range — and in 2026 this is where the most interesting selection shifts are happening.
Why the market is moving to electric VFD drives
The structural drivers of the BOG compressor market now explicitly include "the rapid shift toward electric-drive compression packages," and EEDI Phase 3 "raises the premium on variable-speed electric drives that trim auxiliary load at partial operating points" (Mordor Intelligence; LNG Industry). The reason is straightforward physics: in BOG service the machine spends most of its life at partial load, and a fixed-speed motor controlled by recycle or throttling wastes energy at every off-design point. A variable-frequency drive lets the compressor slow down to match the actual boil-off, cutting power consumption where it matters most — at the partial-load points where the machine actually lives.
For a marine package, that auxiliary-load reduction feeds directly into the vessel's EEDI calculation, which is why owners increasingly specify VFD. For a petrochemical tank farm, it is simpler: the electricity bill is lower and the machine follows the flow without burning energy on recycle.
How VFD changes the family-level decision
VFD shifts some borderline calls:
- It widens screw turndown even further. A screw on VFD can follow flow over a very wide band at good efficiency — reinforcing screw as the BOG default. The Hengli package's 2,800–4,500 RPM adaptive speed range is exactly this in action.
- It softens the centrifugal turndown problem — somewhat. Variable speed moves the surge line and extends the usable window, making centrifugals viable over a wider range than fixed-speed. It does not eliminate surge; you still need anti-surge control.
- It gives reciprocating machines a continuous turndown lever on top of step unloading, smoothing what used to be a stepped capacity curve.
Control-system scope is part of selection
A capable package control system — PLC or DCS — does far more than start and stop the machine. On our screw packages the distributed control system handled real-time suction-pressure compensation, adaptive speed control, and predictive-maintenance algorithms. On the offshore BOG booster skids, PLC-based control integrated emergency-shutdown logic and protection interlocks. The control philosophy (anti-surge for centrifugal, capacity/speed for positive displacement, ESD for everyone) must be settled at selection, not bolted on later, because it drives instrument count, I/O, and the interface to the plant or vessel safety system.
Quotable takeaway: In BOG service the machine lives at partial load. A variable-frequency drive is not a luxury — it is how you stop paying an energy penalty for every hour the boil-off isn't at the design point.
Maintenance, Reliability, and Total Cost of Ownership
Selection does not end at the purchase order. The three machine families have very different lifecycle profiles, and a buyer who optimizes only on capital cost frequently overpays across the life of the asset.
Maintenance load by family
- Reciprocating carries the highest routine maintenance load. Valves, piston rings, and packing are consumables with defined replacement intervals; a serious reciprocating program plans valve overhauls and tracks valve temperature as a leading indicator. The upside is that almost every part is field-serviceable, and the machine tolerates a long service life if maintained.
- Screw sits in the middle. The rotors themselves are robust, and the main attention items are the lube-oil system, the oil-separation elements (coalescers age and must be changed), bearings, and the slide-valve mechanism. Oil quality monitoring is the single highest-value maintenance habit for a screw package.
- Centrifugal has the lowest routine maintenance load — few wearing parts in the gas path mean long intervals between overhauls — but the overhauls, when they come, are specialist work, and condition monitoring (vibration, bearing temperature, performance drift toward surge) is essential because the failure modes are less forgiving.
Reliability and run-time evidence
Reliability is best judged on demonstrated run-time, not brochure claims. The Hengli propylene package reached an 8,400-hour annual operation capacity and processed over 42,000 tons of propylene vapor since 2023 — the kind of operating evidence that should anchor a reliability discussion. We design QA routines to be consistent across projects precisely so that this kind of run-time performance is repeatable rather than lucky.
Total cost of ownership, honestly
When you total capital cost, energy, maintenance, spares, and downtime over a 15–20 year life, the picture often inverts the capital-cost ranking:
- A cheaper fixed-speed machine that spends its life on recycle can cost more in energy than a VFD machine that cost more up front.
- A reciprocating machine with low capital cost can carry higher lifetime maintenance than a screw in the same duty.
- A centrifugal with high availability and low maintenance can win on TCO in a steady large-flow duty even if it costs more to buy.
The discipline is to run the TCO comparison against the actual duty cycle — how many hours at what load — not against the design point. This is the same total-cost-of-ownership logic we apply to industrial refrigeration packages, and it consistently changes which machine looks "cheapest."
Spare-parts strategy and obsolescence
A package is supportable only if its spares are. Map critical spares at the design stage — valves and rings (reciprocating), oil-separation elements and bearings (screw), seals and instrumentation (all) — and confirm lead times and obsolescence risk before signing. On marine packages we build spare-part mapping into the documentation set so the shipowner is not stranded mid-ocean by an unplanned consumable.
Common EPC Procurement Pitfalls (and How to Avoid Them)
After enough projects, the same mistakes recur. Here are the ones that cost the most, and the cheap fixes.
1. Specifying a single design point instead of a duty envelope. This is the root cause of most wrong selections. BOG and process flows swing; specify minimum/normal/maximum flow and the transients, and state the turndown explicitly. Fix: make the duty envelope a mandatory RFQ field.
2. Underestimating the pulsation study on reciprocating machines. Skipping or under-scoping the API 618/688 pulsation and mechanical-response analysis leads to piping vibration, cracked welds, and field rework. Fix: scope the pulsation study into the package contract from day one (SwRI).
3. Forcing a centrifugal into a wide-turndown service. It looks efficient on the datasheet and then surges or burns energy on recycle in practice. Fix: if the turndown is wide, default to screw unless the flow is genuinely large and steady.
4. Leaving the interface boundary undefined. The vendor/site demarcation — utilities, drains/vents, electrical and control I/O, safety signals — is the single biggest controllable schedule risk. Fix: freeze the boundary in the technical agreement, with a clear "vendor provides vs. site provides" list, before fabrication.
5. Treating the documentation set as an afterthought. EPC and classification acceptance hinge on the databook — MTCs, NDE, hydrotest, FAT, ITP. Discovering at FAT that a required record was never planned is a schedule killer. Fix: agree the ITP and documentation index at kickoff; settle hold points early because they govern fabrication duration.
6. Discovering a code requirement late. "We also need PED" or "this has to be ABS-certified" arriving months into fabrication forces expensive rework. Fix: lock all applicable codes — ASME, PED, API, classification society, ATEX — at the requisition stage, not during fabrication.
7. Ignoring the duty cycle in the cost comparison. Optimizing on capital cost alone ignores energy and maintenance over a 15–20 year life. Fix: run TCO against the real duty cycle; a VFD or a more capable machine often wins.
8. Under-sizing the oil-separation train on a screw. Oil carryover into a sensitive downstream is a common and avoidable failure. Fix: characterize the downstream's oil tolerance and size the coalescing separation accordingly.
Avoiding these eight is mostly a matter of front-loading decisions that teams are tempted to defer — duty envelope, pulsation, interface boundary, codes, documentation, and TCO. The cost of settling them at selection is hours; the cost of discovering them in fabrication is months.
The Selection Decision Tree
Here is the sequence we walk through when an RFQ arrives. It resolves the family-level decision before the datasheet stage, then flags the package-level questions that drive cost and schedule.
Step 1 — Establish the duty envelope (not a single point).
Get suction pressure/temperature, discharge pressure, flow at minimum/normal/maximum, and the expected transients. For BOG, define the turndown explicitly. A single design point is the most common cause of wrong selection.
Step 2 — Compute the overall pressure ratio and check it against per-stage limits.
- Overall ratio fits in one stage at ~3.5? Centrifugal is on the table.
- Need ~4.5–6.0 in one stage? Screw or reciprocating.
- Need a very high overall ratio (e.g., the 300:1 we hit at Hengli)? Multi-stage screw or reciprocating; centrifugal would need many casings.
Step 3 — Characterize the gas.
- Wet, dirty, or variable composition? → Screw (or reciprocating). Avoid centrifugal.
- Clean, single-composition, large steady flow? → Centrifugal.
- Purity-critical (no oil in gas)? → Centrifugal, or a carefully engineered oil-separation train on a screw.
Step 4 — Check the flow magnitude.
- Large and steady (big LNG terminal / large reliquefaction train)? → Centrifugal.
- Low flow, high ratio? → Reciprocating.
- Moderate flow with swings? → Screw.
Step 5 — Check temperature headroom against the ceiling.
Reciprocating ~350°F, screw ~300°F, centrifugal ~450°F. If the discharge temperature crowds the ceiling, add a stage with intercooling — which feeds back into Step 2.
Step 6 — Apply service constraints (often the tiebreaker).
- Marine/offshore: vibration limits, compact footprint, classification-society certification (ABS/CCS/DNV/BV), ESD interface. Vibration-sensitivity can rule out a heavy reciprocating machine.
- Hazardous area: explosion-proof rating (Ex d, ATEX).
- Continuous vs. intermittent: reliquefaction-style continuous duty rewards graceful turndown.
Step 7 — Define the package scope and interface boundary.
Separator, coolers, oil station, pulsation dampeners (recip), control system (anti-surge/speed), ESD, and the precise vendor/site demarcation. This is where schedule risk lives — settle it now, not on site.
If you work this sequence honestly, the family chooses itself, the borderline cases land on screw, and the package scope is locked before fabrication — which is the whole point.
Real Project Mapping: Which Case Used Which Machine, and Why
Theory is cheap. Here is how the logic above maps to packages Lmart has actually built. Specifics are limited to what is already public in our project references; we do not disclose contract values, margins, or confidential operating data.
Hengli ST-PP Propylene BOG → Single-Casing Two-Stage Screw
The Hengli Petrochemical 1,000 KTA ST-PP polypropylene facility in Dalian needed BOG management at −43°C cryogenic conditions: process roughly 6,208 kg/h of propylene vapor and reach 1.8 MPa(g) discharge — a 300:1 overall compression ratio that defeats conventional single-machine designs. We delivered a single-casing two-stage screw compressor package (630 kW motor) holding ±0.5 kPa suction-pressure tolerance with adaptive speed control (2,800–4,500 RPM) and ASME B31.3 compliance for the low-temperature service.
Why screw: very high overall ratio + cold/wet cryogenic suction + tight suction-pressure control + wide turndown. That is the screw machine's home turf. Reported results included 82.5% isentropic efficiency (vs an ~78% benchmark) and 2.8 mm/s vibration. The configuration's success supported replication at other large petrochemical complexes.
Satellite Petrochemical Propylene BOG (acrylic tank area, 450 kW) → Screw
A 2017 propylene-mixture BOG compression duty in an acrylic-acid tank area (engineered with Wuhuan Engineering). Tank-farm vapor recovery, variable mixture, moderate flow, 450 kW.
Why screw: classic petrochemical tank-farm BOG — variable composition, moderate flow, wide turndown, compact skid. Textbook screw service.
Satellite Petrochemical LPG BOG (propane tank area, 630 kW) → Screw
A 2018 propane-mixture LPG BOG compression duty in a propane tank area (Wuhuan Engineering), 630 kW.
Why screw: same family logic — LPG/propane tank-farm BOG with mixed composition and turndown. Positive-displacement, oil-flooded screw handles the wet, variable suction without surge concerns.
Guangxi Huayi Propane BOG (propane tank area, 630 kW) → Screw
A 2021 propane BOG compression package for a propane tank area (Wuhuan Engineering), 630 kW.
Why screw: propane tank-farm BOG recovery, the same duty class as the Satellite LPG package. Consistent reasoning, consistent machine choice — which is exactly how a coherent selection method should look across similar services.
Jiangsu Huachang Ammonia Compression Unit (1,120 kW) → Screw-class Refrigeration Compression
A December 2018 ammonia (R717) compression unit: suction quality flow 3,515–12,130 kg/h, 1,120 kW motor at 2,950 rpm, 6 kV supply, ExdIIBT4 explosion-proof, thermosiphon oil cooling, 25 bar design / 28.3 bar test, ASME U-certified, ~38.5 t empty weight.
Why this machine class: large-scale industrial refrigeration ammonia compression with a wide suction-flow band (a ~3.5:1 flow turndown) and hazardous-area duty. The wide flow band and ammonia service favor positive-displacement compression with robust explosion-proof and pressure-vessel certification — and the ASME U stamp on the package reflects the pressure-containing scope.
Zhenjiang Jiangnan Chemical Chloromethane Compression (450 kW) → Gas Compression for Reactive Service
Chloromethane compression duties (2018 and a 2021 repeat, ~450 kW), engineered with Zhejiang Titan.
Why this matters for selection: chloromethane is a corrosive, reactive process gas — material selection and sealing dominate the design as much as the flow/ratio map. The repeat order is the quiet signal: when a compression package is selected and built correctly for a difficult gas, the customer comes back.
Ningbo Huatai (Huatai Shengfu) Fuel Gas Compression (630 kW) → Fuel-Gas Booster
A 2018 fuel-gas compression duty (engineered with SEI), 630 kW.
Why fuel-gas booster: fuel-gas booster skids raise pipeline or process gas to burner/turbine inlet pressure — a moderate-ratio, moderate-flow service that fits a packaged booster skid cleanly. This is the same booster-skid discipline we applied on the FLNG BOG booster work.
14,000 TEU LNG-Fuelled Container Ship BOG Package (marine) → Package-Led, Duty-Envelope Design
Two BOG compression unit packages for a 14,000 TEU container ship's LNG fuel-system integration (for Jiangnan Shipyard / shipowner PIL), shipped September 2023. The engineering centered on stable performance across a duty envelope with frequent transients, a compact and vibration-resilient footprint, corrosion considerations, and clean interfaces to the vessel's gas-detection, ventilation, and ESD logic.
Why this is a packaging story as much as a machine story: marine BOG demands the whole package thinking — turndown, vibration, footprint, classification certification, and a crisp vendor/shipyard interface boundary. The machine selection serves the package, not the other way around.
NGUYA FLNG / Wison BOG Booster Skids (offshore) → Centrifugal/Reciprocating Booster on a Fast-Track Skid
The Eni NGUYA FLNG (Congo LNG) BOG booster skid and the Wison (Nantong) / Burckhardt Compression booster skid were both ABS-certified offshore modules delivered in 35 days, performing BOG recompression and reinjection into the main process line.
Why this is the schedule lesson: the booster duty here is offshore recompress-and-reinject. The headline isn't only the compressor — it's that disciplined modular packaging (pre-procurement, parallel assembly/welding, continuous QA, ABS certification) compressed delivery to 35 days. Selection and packaging together are what win the schedule.
The pattern across all of these: petrochemical and LPG/propylene/ammonia tank-farm BOG → screw, every time. Marine and offshore BOG → package-led design with classification certification and tight interfaces. Fuel-gas → packaged booster skid. The machine family is chosen by the duty envelope, and the borderline duties consistently land on the screw machine — which is why our API 619 oil-injected screw package is the workhorse of the portfolio.
Standards, Certification & Documentation
A gas compressor package lives or dies on the codes it is built to and the evidence that proves it. For EPC and marine buyers, the documentation set is not paperwork — it is the acceptance gate.
The governing API standards
- API 617 — Axial and Centrifugal Compressors and Expander-Compressors. Governs dynamic machines in continuous petrochemical/refinery/pipeline service (API).
- API 618 — Reciprocating Compressors for Petroleum, Chemical, and Gas Industry Services. Drives the pulsation and mechanical-response analysis (with API 688); the 6th edition was under industry review through 2025 (recip.org/EFRC).
- API 619 — Rotary-Type Positive-Displacement Compressors (screw). The basis of our oil-injected screw packages, alongside ISO 10440-1.
Pressure, piping, and process codes
- ASME Section VIII Div.1/2 — for the pressure-containing components (separators, scrubbers, coolers). The Huachang ammonia unit carried the ASME U stamp on its pressure scope.
- ASME B31.3 — process piping, including low-temperature service (the Hengli propylene BOG package complied at −43°C).
- PED 2014/68/EU (CE) — for European-market projects.
Marine classification
For shipboard and offshore packages, classification-society approval is mandatory: ABS, CCS, DNV, BV, LR, NK. The FLNG booster skids were ABS-certified; the 14,000 TEU marine BOG packages were built for shipyard/class acceptance. Lmart holds the ASME U-Stamp, PED/CE, and approvals from six classification societies (DNV, BV, CCS, ABS, LR, NK).
Hazardous-area certification
Compression of hydrocarbons happens in classified areas. Explosion-proof ratings — Ex d / ATEX 2014/34/EU — are not optional. The Hengli propylene package was Ex-certified; the Huachang ammonia unit was ExdIIBT4.
The documentation set EPC buyers expect
A compliant package ships with a structured databook: material identification and MTCs, welding records, NDE/NDT reports, hydrostatic test records, dimensional and assembly verification, instrument loop checks, and the Factory Acceptance Test (FAT) report. On the marine packages, FAT was structured around what matters to integration — functional checks, protection-logic verification, and documentation evidence that supports later onboard commissioning. The ITP (Inspection and Test Plan) defines the hold points; settle it early, because it governs how long fabrication takes.
Lmart's Gas Compression Capability
Suzhou Lmart Energy Equipment Co., Ltd. is a member of Huachang Group and a manufacturer of pressure equipment and packaged systems — ASME pressure vessels, TEMA shell-and-tube heat exchangers, industrial refrigeration packages, and gas compression units built to API 619 / ISO 10440-1, integrated as modular skids.
On gas compression specifically, our portfolio spans the full duty map covered in this guide: propylene and LPG/propane BOG packages for petrochemical tank farms (Satellite Petrochemical, Guangxi Huayi, Hengli ST-PP), ammonia compression units for industrial refrigeration (Jiangsu Huachang), chloromethane and fuel-gas compression for chemical plants, and marine/offshore BOG packages for LNG-fuelled vessels and FLNG platforms (14,000 TEU container ship; Eni NGUYA FLNG; Wison/Burckhardt booster skids).
What that experience buys a project team is selection judgment and packaging discipline — the ability to match the machine family to the real duty envelope and to ship a skid that integrates cleanly. We hold the ASME U-Stamp, PED 2014/68/EU (CE), ISO 9001/14001/45001, and approvals from six classification societies, with 600+ welding qualifications and equipment delivered to 50+ countries. Full in-house scope — design, fabrication, assembly, testing, documentation — runs in an 8,000 m² (38,000 m² total workshop area) integration facility.
We do not claim to be the largest compressor OEM, and we do not build the rotating element of every machine in-house. What we do is package, integrate, certify, and deliver gas compression skids that pass EPC and classification acceptance — and we map the selection logic honestly so the buyer gets the right machine, not the one that is easiest to sell.
Conclusion
The 2026 regulatory and market environment — methane-slip caps, EEDI Phase 3, electric-drive migration, and an Asia-Pacific-led BOG market heading past USD 2 billion — has pulled the compressor selection decision forward and raised the cost of getting it wrong. But the engineering logic underneath is stable and learnable.
Reciprocating (API 618) is the high-pressure-ratio, low-flow specialist; pay for the pulsation study and respect valve maintenance. Centrifugal (API 617) is the large-flow, clean-gas, steady-service machine; mind the surge line and the per-stage ratio limit. Screw (API 619) owns the wet, variable, wide-turndown middle — which is where most petrochemical and a large share of marine BOG actually live. Work the duty envelope first, the per-stage ratio next, then gas character, flow, temperature, service constraints, and finally the package scope and interface boundary. Do that, and the machine chooses itself.
The case mapping makes the method concrete: tank-farm BOG goes screw, every time; marine BOG is a package-led, classification-certified design; fuel gas is a booster skid. Match the machine to the duty, lock the package scope before fabrication, and the schedule stops fighting you.
FAQ
Q1: What is the single biggest factor in choosing between reciprocating, screw, and centrifugal compressors?
The duty envelope — specifically the relationship between flow and required pressure ratio. Per-stage rules of thumb are ~3.5 for centrifugal, ~4.5 for reciprocating, and ~6.0 for screw. High ratio at low flow points to reciprocating; large steady clean flow at moderate ratio points to centrifugal; everything wet, variable, or in between points to screw. Compute the overall ratio and check it against per-stage limits before anything else.
Q2: Why are most petrochemical and tank-farm BOG compressors screw machines rather than centrifugal?
Because tank-farm BOG is wet, variable in composition, swings widely in flow, and often runs from cold/cryogenic suction. Oil-injected screw machines handle wet and variable gas gracefully and offer excellent turndown via speed control, while centrifugals are prone to surge under wide flow swings and are sensitive to composition changes. Lmart's propylene and LPG BOG packages (Hengli, Satellite, Guangxi Huayi) are all screw-based for exactly these reasons.
Q3: When does a centrifugal compressor make sense for BOG service?
On large onshore LNG terminals and big marine reliquefaction trains, where flow is large and relatively steady and efficiency at scale matters most. Centrifugal units hold about 51.5% of the BOG compressor market overall, driven by these large-flow applications. The key is pairing the centrifugal with robust anti-surge control if any meaningful turndown is expected.
Q4: What are the discharge-temperature limits for each compressor type, and why do they matter?
Approximately 350°F (177°C) for reciprocating (valve life), 300°F (150°C) for oil-flooded screw (lubricant), and 450°F (232°C) for centrifugal (rotor metallurgy). They matter because a high pressure ratio raises discharge temperature; if you crowd the ceiling, you must add an intercooled stage, which changes the stage count and the package design. Always check temperature headroom during selection, not after.
Q5: What does a complete gas compressor package (skid) include beyond the compressor itself?
The compressor and driver (motor, often VFD, with Ex rating), a suction scrubber/separator (buffer drum), intercooler and aftercooler, lube-oil station and oil-separation train (for screws), pulsation dampeners (for reciprocating), an offshore-rated and vibration-resistant skid frame and piping, a PLC/DCS control system with anti-surge/speed control and ESD, and clearly demarcated interface terminals. The interface boundary — vendor scope vs. site scope — is the biggest controllable schedule risk.
Q6: How fast can a BOG compressor skid realistically be delivered?
With disciplined modular packaging it can be very fast. Lmart delivered an ABS-certified BOG booster skid for Eni's NGUYA FLNG platform in 35 days, achieved through pre-material procurement, parallel assembly and welding, and continuous in-process QA/QC. Speed is a function of packaging discipline and early interface definition, not just compressor lead time.
Q7: Which standards govern gas compressor packages, and which certifications should EPC buyers require?
The machine standards are API 617 (centrifugal), API 618 (reciprocating, with API 688 for pulsation), and API 619/ISO 10440-1 (screw). Pressure-containing components follow ASME Section VIII; piping follows ASME B31.3; European projects need PED 2014/68/EU (CE). Marine and offshore packages require classification-society approval (ABS, CCS, DNV, BV, LR, NK), and hazardous-area service requires Ex d / ATEX certification. Require the full databook — MTCs, NDE, hydrotest, FAT, and a settled ITP — as the acceptance gate.
Related Reading
- Modular Gas & Marine-Fuel Skid Design
- Ultra-High-Pressure Compressor Snubbers (252–339 bar)
- Pressure Vessel Market 2026: 4 Growth Pillars
- EPC Procurement Hub — pillar guide for buyers
By Qiangbin Chu, Suzhou Lmart Energy Equipment Co., Ltd. — gas compression packages, ASME pressure vessels, and modular skid integration for EPC and marine projects.