LARGE LUNAR DRI FURNACES
DRI furnaces 200 ft. high or 61 meters with  10m x 10m x 10m hearths for 7700 tons of iron per run will be necessary to produce the 100 million tons or more of steel needed for 1000 SPS.  Pipes, heat exchangers, RWGS reactors, pumping systems, "lump" or briquette factories, gas storage facilities will be comparably large.  Decades of lunar development will be necessary.

If 7700 tons of iron is made daily for 150 days, then one furnace can produce 1.15 million tons per year.  If power is available by night from remote solar power plants, 2.3 million tons per year.  In ten years, ten of these giant DRI furnaces could make 115 to 230 million tons of steel. 

Power towers could supply electricity to induction heating systems and process heat from molten salt to super heat CO reductant gas.  Combined with heat exchangers this will be very efficient.

Large open hearth and/or basic oxygen furnaces heated by molten salt pipelines from power towers at 1000 C. + and electricity could convert iron to steel.

Ceramic blocks several meters thick will line the shaft.  The blocks will be interlocking and welded with microwaves or electron beams.  Using low hills instead of building pyramids of blocks will save blocks and save time for construction. 

Vast areas of mare must be mined for carbon.  If pockets of volcanic gas exist beneath low volcanic domes (200 + volcanic domes in the Marius Hills) that contain CO gas then we must drill for that gas.  Ground penetrating radars and siesmic probes could reveal the existence of these gas pockets, should they exist.

Enormous amounts of flux will be needed.  It will be necessary to erect huge solar dishes made of magnesium foil perhaps, since magnesium production may outstrip aluminum production, to roast regolith and boil away iron oxide at 1200 C and silica at 1500 C.  The FeO will be condensed and used as ore.  The calcium oxide and aluminum oxide enriched material will be used as flux.  MgO that is broiled out of regolith will also be used as flux.
This operation will be as big a job as iron smelting and steel making.  Limestone deposits do not exist on the Moon.  Slag will be used for concrete structures in lunar lava tubes.
David A. Dietzler, 2007
On the flux page I sketched block structures for roasting minerals with solar heat.  Perhaps it would be easier to dig trenches and section them off with block walls, cover with cast basalt slabs with manhole covers in them to lower robots in to dig up materials, rather than build big welded interlocking block structures.  I call this a "trench furnace."  Orange volcanic glass is rich in iron.  FeO content 20% in the form of silicate minerals.  Green volcanic glass is rich in magnesium bearing olivine.  Volcanic glass will be a good "ore."  FeO, the mineral wustite when found on Earth, will not need flux as it contains no silica, unlike ferrosilte (Fe2Si2O6) and fayalite (Fe2Si2O4).  The trench furnace can also broil anorthite to drive off SiO2 and leave calcium aluminate and/or CaO.  Trench furnaces and massive parabolic foil reflectors to concentrate solar energy on to the minerals to obtain FeO and flux might be the way to produce enough ore and flux to make large scale iron and steel production on the Moon possible.  Flux production is the "rate limiting step" because it involves more than just mining mare regolith and magnetically extracting iron bearing minerals. 


Energy for Very Large Lunar DRI Furnaces

On the page Lunar  Forge Reactions I determined that about 700 MWhrs would be needed to smelt 100 tons of iron from 236 tons of ferrosilite.  If we apply this energy over 12 hours we'd need a 58 MWe power source.  If we want to smelt 7700 tons of iron in 12 hours a 450 MWe power source would be needed.  If heat leakage from the furnace was low enough perhaps we could apply the energy over 24 hours and tap the furnace for iron every few hours.  A 225 MWe power source would then be needed.  A 10% efficient solar power panel farm 1825 meters square for the 12 hour smelt and a 1290 m square for the 24 hour. 

These furnaces built into hills or crater rims would need large compressors, CO storage tanks, wide pipes, etc.

After many decades of development on the Moon it seems plausible that these giant furnaces could be built.  Even so, if we create these replicas of Pittsburg on the Moon we will be able to smelt  7700*14*12 = 1,293,600 tons of iron/steel per year and twice that if we have nuclear reactors operating on the Moon to power the furnaces constantly.  It would still take one of these monsters 50 to 100 years to produce enough steel to build 1000 SPSs and ten of them to produce that much metal (about 125 million tons) in 5 to 10 years.  However, how long would it take to reach this stage of industrial development on the Moon? 30, 40 or 50 years? Or longer?  Can we produce hundreds of millions of tons of FLUX too? Aluminum and titanium are more difficult to produce than steel, so I doubt that production of those metals will exceed steel production on the Moon.