01Answer in brief
A traditional steam engine cannot simply be filled with CO₂. Its cylinders, valve chest and rod packing are cast iron, built for 8–17 bar (125–250 psi). A CO₂ cycle that condenses on 5 °C water already sits at 46 bar (666 psia) on the low side and 200 bar (2,900 psia) on the high side. Even the exhaust pressure is three times the old engine's rating.
CO₂ also produces about 12–15 times more work per litre of swept volume than steam. Its mean effective pressure is about 58 bar, against 3–5 bar in a steam engine. That is good for size, but it means piston and rod loads 12–17 times what the old crank, crosshead and bearings were built for.
There are three ways to make an old engine run on CO₂. The best gives about 13–16 kW (18–21 hp) net, which is less than the same engine produces on steam (16 kW condensing) and half of what a purpose-built CO₂ piston expander delivers (32.2 kW, 43 hp). The old engine is not the way to go. But several of its design features are well worth carrying into the purpose-built CO₂ machine (Section 06).
Net hydraulic power from 500,000 BTU/h of logs
Net = oil power at 207 bar after driving the CO₂ feed pump. Colours: ■ purpose-built CO₂ piston, ■ old engine rebuilt for CO₂, ■ old engine in pressurised crankcase, ■ the same old engine on steam (from the previous study). The low-pressure CO₂ gas cycle produces no net power and is not shown.
02The core problem: pressure and force
Example: a typical old 40 hp horizontal mill engine
| Item | On steam (as built) | Same bore on CO₂ |
|---|---|---|
| Bore × stroke | 9 × 12 in · 229 × 305 mm, double-acting, ~200 rpm | |
| Working pressure | 8.6 bar · 125 psi | 196 bar · 2,840 psi |
| Peak pressure difference | ≈ 8.6 bar | 150 bar (196 → 46) |
| Peak rod load | 35 kN · 7,950 lbf | 617 kN · 138,700 lbf (17×) |
| Mean effective pressure | 3–5 bar | 45–58 bar |
| Indicated power | ≈ 29 kW · 39 hp | would be ~10× (but the engine breaks) |
Representative dimensions for an engine of this class; individual engines vary.
What fails first
- Cast-iron cylinders and valve chest: a pressure (burst) hazard at 46–200 bar. Brittle failure throws fragments.
- Slide (D) valve: pressure holds it against its seat. At 150 bar across a ~100 cm² valve face, the seating force is about 150 kN (≈ 34,000 lbf), 17 times the steam load, so the valve cannot be moved. It must be replaced with balanced piston valves or poppets.
- Rod gland packing: fibre or metallic packing built for 15 bar steam will not seal 46–200 bar CO₂. High-pressure chevron PTFE stacks, as in CO₂ plunger pumps, are needed.
- Crank, crosshead, bearings: rated for about 35 kN peak, against 617 kN on CO₂ at the original bore.
- Cylinder oil: CO₂ dissolves into oil and thins it, washes it off the walls, and carries it into the recuperator and the 500 °C+ heater coil, where it cokes.
- Clearance volume of 6–12 % (normal for steam) wastes dense CO₂ unless the engine is built to recompress it.
03Three ways to run an old steam engine on CO₂
| Route | What is done | CO₂ in | Expander kW | CO₂ pump kW | Net hyd. kW / hp | Verdict |
|---|---|---|---|---|---|---|
| A. Keep the engine's 15 bar rating; closed CO₂ gas loop, 5 → 15 bar, with a piston compressor on the same crank | CO₂ cannot condense at 15 bar with 5 °C water (it would need −28 °C), so it must stay a gas and be compressed | 300 °C, 15 bar | 92 | 129 (compressor) | negative | Does not run — compressing the gas takes more work than the engine produces |
| B. Pressurised crankcase: put the unmodified engine inside a 70 bar pressure vessel at exhaust pressure | Cylinder parts see only Δp ≈ 14 bar; CO₂ boils subcritically at 60 bar | 300 °C, 60 bar | 8.6 | 1.4 | 5.2 / 7.0 | Not worth it — 1.3 : 1 pressure ratio, and a multi-tonne coded pressure vessel around the engine |
| C. Rebuild: new forged-steel HP cylinder sleeved to Ø55 mm (2.2 in), poppet valves, high-pressure gland; old frame, crank and flywheel kept | Bore shrunk until the rod load is back to the original 35 kN rating; oil-lubricated, so the CO₂ is limited to 300–350 °C | 300–350 °C, 200 bar | 31–34 | 9.9 | 13.2–15.7 / 17.6–21.0 | Works, poorly — it is really a new expander on an old frame |
Route C also gives up efficiency: oil-lubricated rings and an old frame limit the CO₂ to about 300–350 °C instead of 500 °C. The rebuilt valve gear and leaky, wide-clearance cylinders reach an assumed 65 % expander efficiency (80 % for the purpose-built machine). And driving a rotary hydraulic pump from the crank costs 88 % mechanical × 90 % pump efficiency, against 92 % for a direct plunger.
04The purpose-built CO₂ piston expander
- Small, forged cylinders rated 250–300 bar: HP Ø63 mm / LP Ø86 mm × 100 mm stroke at about 400 rpm (crank version), or HP Ø82 / LP Ø111 × 400 mm at 30 cycles/min (free-piston tandem driving hydraulic plungers). Peak gas loads are 28–32 kN, which a small, light running gear can carry.
- Compound expansion 196 → 95 → 46 bar, giving 2.1 : 1 per stage and keeping force variation modest.
- Hydraulically actuated poppet valves, timed by the PLC, so cutoff can be varied while running.
- Oil-free hot end: carbon-graphite or PTFE-bronze rings, so no oil reaches the recuperator or heater. This allows CO₂ at 500 °C (932 °F), and the hotter CO₂ alone is worth +10 kW over 300 °C.
- Cooled distance piece between the hot cylinders and the hydraulic or crank end.
- Low clearance (≈ 2–3 %) with exhaust-closure recompression ("cushioning").
| At 500 °C / 200 bar | Metric | Imperial |
|---|---|---|
| CO₂ flow | 0.301 kg/s | 2,390 lb/h |
| Expander work | 45.1 kW | 60.5 hp |
| Gross hydraulic | 41.5 kW | 55.7 hp |
| CO₂ pump | 7.9 kW | 10.6 hp |
| Net hydraulic | 32.2 kW | 43.2 hp |
| Oil at 207 bar | 120 L/min | 31.8 gpm |
| Mean effective pressure | 58 bar | 845 psi |
| Swept volume flow (LP) | 7.75 L/s | 16.4 cfm |
| Estimated expander mass | 250–450 kg | 550–1,000 lb |
At 300 °C (the same temperature limit as the converted engine) the purpose-built machine still delivers 22.4 kW. At an expander efficiency of 70 % instead of 80 % it delivers 28.3 kW at 500 °C.
05Side-by-side comparison
| Criterion | Purpose-built CO₂ piston expander | Old steam engine rebuilt for CO₂ (route C) |
|---|---|---|
| Net hydraulic power | 32.2 kW · 43 hp | 13–16 kW · 18–21 hp |
| Wood → net hydraulic | 22.0 % | 9.0–10.7 % |
| Max CO₂ temperature | 500 °C (to 700 °C with Ni-alloy) | 300–350 °C (oil limit) |
| Pressures | 200–300 bar, designed for it | 200 bar in new HP parts; the old frame is the weak link |
| Cylinders / valves | Forged steel, poppet valves, variable cutoff | New HP cylinder + poppets; the original valve gear is mostly discarded |
| Lubrication | Oil-free hot end | Cylinder oil → oil separator needed; coking risk in the heater |
| Rod seal | Cooled distance piece, high-pressure PTFE stack | Retrofit high-pressure gland in an old stuffing box |
| Drive to hydraulics | Direct plungers (92 %) or crank + pump | Crank → belt/coupling → rotary pump (≈ 79 %) |
| Size / mass | ≈ 1 × 0.6 m, 250–450 kg | Engine + flywheel ≈ 2–3 t (4,400–6,600 lb) |
| Transport to a Yukon claim | Pickup or skid | Heavy-haul, crane |
| Cost driver | New design and certification | Antique engine purchase + almost all hot parts new anyway |
| Pressure-safety case | Clean, coded from the start | Mixed old cast iron + new high-pressure parts; hard to certify |
| Field repair | Moderate (rings, valve seats are consumables) | Good on the running gear; HP parts as specialised as the new machine |
06What the old steam engine does teach the CO₂ design
Ideas to carry over
- Variable cutoff for load control, as with link motion, Corliss trip gear or a governor. Here it is done with PLC-timed poppets, so output follows the wash-plant load without throttling.
- Compounding (HP → receiver → LP) to keep the pressure ratio and force variation per cylinder moderate.
- Uniflow exhaust (exhaust ports uncovered by the piston at the end of the stroke) for low exhaust losses, plus deliberate cushioning recompression.
- Poppet valves (Lentz/Stumpf practice) instead of slide valves. They are the only valves that survive high pressure difference.
- Crosshead plus distance piece to keep the hot cylinder away from the cold running gear and hydraulics.
- Low speed, flywheel and belt drive to a standard hydraulic pump. That gives a robust, field-serviceable drive at 200–400 rpm.
Where an old engine still has a role
- On steam, as intended: 16 kW (21 hp) condensing, or about 10 kW plus 150 kg/h of thaw steam non-condensing. That beats the same engine converted to CO₂.
- As a test-bed frame for the first CO₂ cylinder and valve trials at low power. The crank, flywheel and crosshead are proven, and the CO₂ cylinder can be sized to the frame's 35 kN rating.
- For marketing and heritage in the Yukon: a visible link to Klondike-era steam, with a modern CO₂ power unit doing the real work.
07Conclusions
Build a purpose-built CO₂ piston expander; do not convert an old steam engine
- An unmodified old steam engine cannot run a CO₂ cycle with 5 °C cooling. Its 8–17 bar rating is below even the 46 bar CO₂ exhaust pressure. Running it on a low-pressure CO₂ gas loop produces no net power.
- Enclosing the old engine in a pressurised crankcase works only at a 1.3 : 1 pressure ratio: 5.2 kW net, and a large pressure vessel.
- Rebuilding the engine with new high-pressure cylinders on the old frame gives 13–16 kW net, limited by rod load, oil temperature and conversion losses. That is less than the same engine produces on steam.
- A purpose-built CO₂ piston expander gives 32.2 kW (43 hp) net at 500 °C / 200 bar, or 39.2 kW at 700 °C. That is 2–2.5× the best conversion, at roughly 1/5–1/10 of the mass.
- Carry over the steam-engine lessons that suit CO₂: compounding, variable cutoff, poppet valves, uniflow exhaust, a crosshead with a cooled distance piece, and a low-speed flywheel drive.
08Basis and assumptions
| Parameter | Value used |
|---|---|
| Heat source | 500,000 BTU/h (146.5 kW) logs, 800 °C flue, 150 °C stack with air preheater; CO₂ heater exit = CO₂ inlet + 30 K. Gives 110.8–114.5 kW to CO₂ |
| Cooling / condensing | 5 °C water/glycol; CO₂ condensing 10 °C at 45.0 bar, subcooled to 8 °C |
| Recuperator | Counterflow, 10 K minimum pinch |
| Purpose-built expander | 80 % isentropic (70 % sensitivity); plunger transfer 92 % |
| Converted engine (route C) | 65 % isentropic (70 % sensitivity); mechanical 88 % × hydraulic pump 90 %; CO₂ limited to 300–350 °C by cylinder oil |
| Pressurised crankcase (route B) | High side 61 bar (Δp ≈ 14 bar within old cylinder rating); 65 % expander |
| Low-pressure gas loop (route A) | 5 → 15 bar, compressor 70 %, expander 65 %, 300 °C, recuperated |
| CO₂ feed pump | 65 % isentropic, driven by hydraulic motor (85 %) |
| Reference steam engine | 9 × 12 in horizontal, 125 psi, ~200 rpm, ≈ 39 ihp — representative, not a specific engine |
| Properties | CoolProp 8 (Span–Wagner CO₂) |
Mass and size figures are order-of-magnitude estimates. This is a preliminary feasibility comparison, not a certified design.