I think you're making the same common mistake most people make, and
omitting the things that will kill your idea dead.
a/ raw materials to work ready alloys and castings
This is a bitch, you need a geologist, a miner, a metallurgist, a
chemist, and a handful of labourers, and a bunch of time....
b/ lathe bed, piece of cast iron with parallel sides, brunel made one
by rubbing three sheets of iron together with grinding compound, took
This is a bitch, those labourers are in for more back breaking work
c/ precision screw threads, make and repair taps and dies.
This is a bitch, any fule can cut a thread of dimensions and
proportions of their own choosing, making one that works for different
materials and is good enough so __any__ size x nut will fit ___any___
size x bolt and stays torqued until you put a spanner on it is tough
e/ ball bearings
This is a bitch on heat, you want a REAL modelling challenge make two
identical roller bearings from stock metal that have 10% of the
performance envelope of shop bought ones
This is tougher than it sounds, 1920s lube oil won't keep a modern high
performance engine alive
My answer to you question is this.
3 x identical 40 containers, each a complete identical shop, each
powered by an identical petter CS type compression ignition motor
powering identical ac alternators, dc dynamo and compressors, with a
backup prime moves as steam engine
3 x more identical containers containing every conceivable spare for
(don't forget mundane stuff like gasket paper)
3 x more identical containers containing a variety of fuels and lubes
3 x more identical containers containing identical smelting and foundry
30 x more identical containers containing raw materials
yeah, what you want can be done, but think coastal shipping size
"base", not something you can load in the back of a pick up.