Why mid-scale exists
Baseload plants are efficient once they run, but building them is brutal. Every train is a bespoke, stick-built job needing thousands of workers on a remote site for years. In Australia the 2010s wave of projects ran tens of billions over budget, largely on site labour and productivity. Mid-scale flips the model: make the train small enough to build in a controlled yard where labour is cheaper and quality is repeatable, then ship it. The trade is efficiency for predictability.
It also suits fields that are too small or too uncertain for a baseload commitment. You can build three trains, prove the reserves, and add three more. And because many identical units are built in sequence, each one is cheaper and faster than the last — the "learning curve" effect that never happens on one-off mega-projects.
What "modular" actually means
A module is a steel frame — a skid the size of a house or bigger — with the equipment, piping, cable trays and instruments already installed and pressure tested. Modules for one train might include a pre-treatment module, a cold-box module, a compressor module, and utility modules. They're lifted onto a ship, sailed to site, lifted onto pre-poured foundations, and connected by interconnecting pipe racks. The site work becomes civil, lifting, hook-up and commissioning rather than fabrication.
The limits are size and weight. A module has to fit on a heavy-lift vessel and be lifted by cranes on site, which caps what can be pre-built. Large spiral-wound exchangers and Frame gas turbines don't modularise well, so mid-scale designs use plate-fin cold boxes and aeroderivative turbines or motors — machinery that arrives as a package.
The process, and how it differs from baseload
The stages are identical in order — treatment, dehydration, mercury removal, heavies removal, liquefaction, flash, storage. The differences are in what's chosen at each step.
- Pre-treatment is often shared. Several trains draw from one common amine unit and dehydration unit sized for the whole site, rather than each train having its own. Only the cold end is duplicated.
- Heavies removal is simpler. Many mid-scale plants take pipeline-quality gas that's already had NGLs stripped by someone else (this is the US model — Gulf Coast pipeline gas). A single scrub column or a heavies removal unit is enough; full fractionation is skipped.
- Liquefaction uses a single or dual mixed refrigerant in a cold box. The gas and the refrigerant flow through stacked aluminium plate-fin passages. Refrigerant is compressed in one or two stages, cooled, and let down through a JT valve. A PRICO (Black & Veatch) train has one MR loop and one compressor — about as simple as an export-scale plant gets.
- Drivers are aeroderivative or electric. An LM6000 delivers about 45 MW, roughly enough for a 1–1.5 mtpa SMR train. Aeroderivatives start quickly, can be swapped as a package in days, and are widely supported. Electric motors are used where grid power is available and emissions matter.
- Storage is conventional. Full-containment tanks and a jetty are shared across all trains — the tank farm doesn't scale down with the train size.
The technology menu
| Process | Licensor | Typical range | Notes |
|---|---|---|---|
| PRICO | Black & Veatch | 0.2–1.5 mtpa | One MR loop, one compressor, one cold box. Lowest equipment count. Used at Calcasieu Pass and Plaquemines (Venture Global, US) — 18 trains of ~0.6 mtpa. |
| IPSMR | Chart Industries | 0.5–1.5 mtpa | Integrated pre-cooled SMR. Chart supplies the cold box too. Used at Corpus Christi Stage 3 (7 trains × ~1.5 mtpa, US). |
| Movable Modular Liquefaction System (MMLS) | Shell | 0.25 mtpa per unit | Elba Island (US) — 10 units. Extreme end of small trains, high unit count. |
| DMR | Shell, Air Products | 1–3 mtpa | Two MR loops. Better efficiency than SMR, no propane inventory. Bridges mid-scale into baseload and FLNG. |
| AP-SMR / AP-DMR | Air Products | 0.5–2 mtpa | Air Products' mid-scale offerings, can use small coil-wound exchangers. |
| Nitrogen expander | Various | Under 0.5 mtpa | Overlaps with small-scale. Simple and safe, but 20–30% more energy per tonne. |
Trade-offs
Strengths
- Predictable cost and schedule — the biggest reason it's chosen.
- Phased investment: produce revenue from early trains while building later ones.
- Redundancy: losing one of twelve trains costs 8% of output, not 50%.
- Fewer site labour hours, less exposure to remote-site productivity.
- Standard machinery with global spares support.
- Good turndown at site level — just run fewer trains.
Weaknesses
- Lower efficiency: SMR uses maybe 10–20% more fuel per tonne than a C3MR train at full rate.
- More equipment to maintain per tonne — twelve compressors instead of two.
- Modules cost more per tonne of steel than stick-built; the saving is in labour and schedule risk, not materials.
- Plot area can be larger, because modules need lifting clearances and repeated utilities.
- Shared tank farm and jetty still cost as much as a baseload site's.
Operating a mid-scale plant
Operators here manage a fleet rather than a single machine. The control room shows twelve or twenty near-identical trains, and the job is balancing them: which trains are running, which are in a mol-sieve regeneration step, which is down for its aeroderivative swap. Because the trains are simpler, a single trip is a smaller event and restarts are quicker — an SMR train can be back at rate in hours rather than the day-plus a C3MR train can take.
The things that catch people out are the common systems. A hiccup in the shared amine unit or fuel gas system takes every train with it, so the "boring" utilities get more attention than the cold boxes.
Refrigerant inventory per train is small, which reduces the size of any single release. But module-based layouts pack equipment densely, so congestion (which makes vapour cloud explosions worse) is a design concern. Cold boxes are perlite-filled and nitrogen-purged; a leak inside the box shows up as rising nitrogen purge pressure or hydrocarbon in the purge gas. Multiple identical trains also mean identical failure modes — a design flaw found in train 3 exists in all of them.
Where it fits in Australia
Australia's existing export fleet is almost entirely baseload, so mid-scale hasn't had a home here yet. It gets discussed for stranded onshore gas in the Northern Territory (Beetaloo), for reusing existing jetties and tanks as older plants retire, and for domestic gas supply projects on the east coast. Every US Gulf Coast export project sanctioned since about 2019 has been mid-scale or modular in some form, which is why it's now the default template internationally.
Self-check — Module 2
- What is the main reason a developer picks mid-scale over baseload, given mid-scale is less efficient?
- Why can't a spiral-wound MCHE be used in most modular designs?
- Explain why a fault in the shared amine unit matters more than a trip on one train.
- Roughly how much power does one LM6000 provide, and what size train does that support?