Lost foam casting (LFC) is a near-net-shape process where expanded-polystyrene (EPS) patterns are coated, embedded in unbonded sand under vacuum, and replaced by molten metal — a 2026 OEM capability summary from Tsuchiyoshi ACTY confirms LFC's role as a mainstream low-defect route for steel, iron, and bronze-steel bimetallic parts [S1].
Capacity bands used by Chinese LFC suppliers in 2026 cluster at 1,000-10,000 t/yr for foundry job shops and 10,000-30,000 t/yr for industrial lines; for example, Qingdao Shengmei Machinery lists LFC lines alongside no-bake sand reclamation and V-process molding equipment on its 2026-06 catalog page, serving North American and European foundries [S2]. Specification begins with four numbers: annual tonnage, maximum pattern envelope (L x W x H), flask internal dimensions, and the coating-station throughput that keeps the line fed.
Process Map and Why Vacuum Defines the Cell
LFC differs from conventional green-sand and no-bake molding lines in that the pattern is not withdrawn — it is vaporised in situ by the incoming melt, with decomposition products drawn out through the coated sand by negative pressure. The 2018 China Foundry investigation of carbon contamination in low-carbon-steel LFC parts confirms that pattern type (EPS, STMMA, copolymer blends), coating permeability, and applied vacuum together govern carbon pickup in the cast skin [S6].
Vacuum level is the single highest-leverage setpoint on the cell. Operating windows on industrial LFC installations typically sit at 0.03-0.06 MPa gauge (≈ 300-600 mbar absolute depression); the Qingdao Shengmei equipment range publicly advertised for 2026 specifies 0.04-0.05 MPa working vacuum with a 0.1 MPa design vacuum on the holding tank [S2]. Go too low and gas-defect and lustrous-carbon defect rates climb; push past the design ceiling and sand fluidisation collapses, defeating the compaction step that supports the foam against metallostatic head.
The Four Sizing Levers Engineers Should Lock First
Selection is driven by a small set of physical constraints, not by marketing brochures. For a new LFC cell in 2026, the four numbers to freeze before talking to a vendor are: (1) annual output in tonnes, (2) the L x W x H envelope of the largest pattern the foundry will run, (3) flask size that fits that pattern with 50-80 mm sand clearance on each face, and (4) the dry-cycle time of the coating booth that determines whether the line is coating-bound or pouring-bound. Reference production-capacity listings from Chinese LFC foundries in 2026 confirm the 1,000-30,000 t/yr envelope and 10,000+ t/yr single-line output is achievable for EPS-patterned steel and iron parts [S4].
Once those four numbers are set, every other spec follows: sand tank volume scales with flask footprint, vacuum pump displacement scales with total flask surface area, coating booth length scales with pattern count per shift, and pouring ladle capacity scales with the heaviest pour. Where the line is to be installed on an existing foundry floor, retrofit constraints typically cap flask size at 1,200 x 1,000 x 800 mm; greenfield cells at integrated steel and iron plants extend to 2,000 x 1,500 x 1,200 mm per published 2026 LFC supplier data [S2].
Pattern Foam, Coating and Carbon Control
Pattern density is the most common hidden source of variation. The China Foundry LFC study found that EPS pattern densities of 0.018-0.025 g/cm³ gave the cleanest low-carbon-steel castings; heavier 0.030-0.040 g/cm³ patterns (often used for large iron pours) raised surface carbon by a measurable margin and required either longer vacuum hold or a more permeable coating [S6]. Foundries running mixed steel and iron on one LFC cell therefore keep at least two pattern-density SKUs in stock and switch coating recipes between heats.
Refractory coating rheology is the second carbon-control lever. The 2026 Tsuchiyoshi ACTY foundry-materials catalogue documents alcohol- and water-based LFC coatings with solids loadings of 35-50 wt%, Brookfield viscosity bands of 800-2,500 mPa·s at 25 °C, and dry-film thickness targets of 0.5-1.5 mm depending on alloy and pour weight [S1]. A bimetallic bronze-steel LFC study in Chemical and Petroleum Engineering (2021) further shows that pre-placed bronze inserts survive the EPS decomposition step only when coating permeability is high enough to evacuate styrene monomer residue before molten steel arrival, anchoring permeability at the top of the spec band for insert-bearing parts [S3].
Sand, Reclamation and the Auxiliary Loop
Unbonded silica or chromite sand is the third LFC consumable and the one most often undersized. A 5,000 t/yr LFC line burning through 25-40 kg of compacted sand per flask, 200-400 flasks per shift, needs a 4-8 t/hr sand cooling and reclamation loop to stay clean; without it, fines and pyrolysis residue accumulate, coating plugging accelerates, and defect rates rise within weeks. Qingdao Shengmei's 2026 product line groups no-bake sand reclamation with LFC and V-process cells because the reclamation plant is shared infrastructure, not a bolt-on [S2].
Sand temperature at the flask is a hard operating limit: above 60-70 °C, EPS distortion begins and dimensional accuracy on the casting degrades. Industrial LFC cells in 2026 pair the sand tank with either a fluidised-bed cooler or a chilled-air loop returning sand to 25-35 °C before re-compaction; a 1,500 m³/hr chilled-air unit is typical for a 10,000 t/yr line per Chinese supplier reference data [S2]. The same logic applies whether the LFC cell is being installed standalone or as the casting side of a static-pressure molding machine on a hybrid floor.
Throughput Math: Coating Booth vs Pouring Station
The rate-limiting step on most LFC lines in 2026 is not the pour — it is the coating and drying booth. A single-station dip-and-dry booth handles 30-60 patterns per shift for medium patterns (300-600 mm), falling to 12-20 patterns per shift for large patterns above 1,000 mm because of the 4-8 hr drying window for water-based coatings. For a 5,000 t/yr foundry running 200 mm-thick iron castings at roughly 20-30 kg per pour, the booth must sustain 20-25 patterns/hr, which is unreachable with a single booth — a two-booth parallel layout is the standard 2026 answer [S1].
Pouring station capacity is the inverse constraint for foundries making heavy iron or steel parts. A 500 kg pour into a 1,200 x 1,000 x 800 mm flask takes 60-90 s of ladle time plus 8-12 min of solidification under vacuum, which sets the floor on flask turnaround. Suppliers cluster the 10,000 t/yr industrial LFC cell around a single 2-3 t pouring ladle with a 12-15 min cycle, or a twin-ladle layout for 20,000-30,000 t/yr cells [S2]. For foundries mixing lost-foam with conventional sand work, a separate automatic molding line running in parallel absorbs the small-pattern volume that does not need LFC's near-net-shape advantage.
Low-Pressure and Vacuum-Assisted Variants
Low-pressure lost foam casting (LP-LFC) pressurises the crucible at 0.02-0.10 MPa to push melt up a refractory stalk into the evacuated flask, rather than relying on gravity pour.
For non-ferrous work below 1,000 t/yr, LP-LFC competes with conventional low-pressure die casting; for ferrous work above 5,000 t/yr, gravity LFC under vacuum remains the 2026 default. The lost-foam casting line spec, regardless of variant, always comes back to: flask size, vacuum envelope, coating-booth throughput, and the sand-reclamation loop that ties it all together.
Decision Matrix: When LFC Is the Right Answer
Choose LFC when the part has internal cavities that would otherwise require cores (engine blocks, valve bodies, pump housings, pipe fittings), when the geometry is too complex for green-sand parting, and when annual volume justifies the foam-pattern tooling. Reconsider when pattern tooling cost cannot be amortised below roughly 500 pieces, when the alloy is highly reactive with carbon (some stainless grades where lustrous carbon defects are unacceptable), or when the casting weight exceeds the practical flask limit of about 5 t per pour. A 2026 LFC process overview from Tsuchiyoshi ACTY lists carbon-steel, grey iron, ductile iron, and bronze-steel bimetallics as the routine alloy range, with austenitic stainless flagged as requiring coated-foam and tighter vacuum discipline [S1].
For a 3,000 t/yr job-shop LFC cell in 2026, expect a 1,200 x 1,000 x 800 mm flask, 0.04-0.05 MPa working vacuum, a 2 t pouring ladle, a 4-6 t/hr sand loop, and a two-booth coating station — capex of that footprint is the order of magnitude Chinese LFC equipment makers such as Qingdao Shengmei quote in their 2026 export catalogue [S2]. For 10,000+ t/yr industrial cells, plan on duplicate coating lines, a twin-ladle pouring bay, and dedicated sand-cooling towers; this is the band where the cell shifts from batch to continuous flow.
Selection Checklist Before Talking to an OEM
Before issuing an RFQ, lock the following: target annual tonnage and tonnage per alloy family, the largest pattern L x W x H, flask size derived from pattern plus 50-80 mm sand clearance, coating throughput in patterns per shift, sand-reclamation capacity in t/hr, and the floor footprint including sand cooling towers. Then ask the OEM for vacuum-pump displacement in m³/hr at working pressure (typically 2-4× flask volume per minute), coating-booth airflow for the drying step, and reference installations running the same alloy mix [S2].
Two follow-up signals to track over the next 6-12 months: (1) Chinese LFC equipment makers expanding the 10,000-30,000 t/yr single-line output band, which will pull the capex-per-tonne curve down; and (2) coating suppliers releasing lower-VOC water-based systems that pass the same permeability and dry-film spec band as alcohol-based legacy coatings, removing a regulatory friction point for European installs. Foundries specifying LFC cells alongside adjacent equipment — for example, a linear guide retrofit on a flask-handling robot — should also verify that motion-system payloads match flask-plus-pattern mass, since an LFC flask at full sand compaction is heavier than the equivalent green-sand flask by 15-25%.