Not worth it considering the cost of the boiler, except if you're way low on everything except iron. If you have 1 OC/recycler, consider the case where you have 4 boilers, 47 recyclers, 54 industrial steam engines, 4 aqueous accumulators, 22 magma crucibles, and 22 geothermal generators. This produces 403 eu/t, at a cost of (288 boiler + 444 recycler + 702 engine + 32 accums + 88 crucibles + 220 geos) = 1774 iron. That's 4.4 iron/eu/t. For comparison an ordinary rec/gen with 2 OCs (best for iron usage) is: 5.75 eu/t. ...I was never here.
I tend to not think about things very much in terms of resources used any more, but rather in terms of overall efficiency and usefulness. But you're right that resources is something which some would value more then I.
Also, some of your numbers are wrong, and kinda weird...
1 high-pressure steam boiler would require approximately 1 scrap per second, as I calculated earlier, which would be supplied almost exactly by 18 recyclers running full-time. (I prefer not using Overclocked machines since they're less efficient in terms of EU, but the total iron cost can be reduced considerably if you don't mind that) (That's 96 steel for the boiler, plus 189 refined iron for the recyclers, totaling 285 so far)
This boiler will provide enough steam for exactly 18 industrial steam engines, which produce 8MJ/t each, and cost 11 steel each, plus one iron for the piston, totaling 12 iron each. That's 216 iron, and it should be noted that these will produce 144MJ/t in total. (501 iron so far)
That's enough for 7.2 magma crucibles. I'll keep the decimal there so that the numbers scale up properly. (Yes, I know, you CAN use netherrack, but there's no way to really infinitely generate netherrack without devouring the nether.)
7.2 magma crucibles cost 5 iron each (36 iron) and each one will produce 1 lava every 1000 ticks, which would basically be 0.001 lava/t, totaling .0072 lava/t for them all. (537 total)
You'll also need 0.288 igneous extruders to keep those crucibles running, which is roundabout 2 iron. (539 total)
Now, the cool thing about lava is that, you don't HAVE to convert it into EU. Instead, you can centrifuge it into copper, tin, electrum, and tungsten at an industrial centrifuge, or you can turn it into obsidian at an igneous extruder, pulverize it at a pulverizer, then electrolyse it at an industrial electrolyzer into iron, (!!!) manganese, and silicon.
However, I'll assume you just wanna convert it all into EU.
For that, your best bet is the Gregtech thermal generator, which costs a geothermal, (10 iron) 5 invar, (2/3 iron each, 3 1/3 total) 2 advanced circuits, (Or was it normal? either way, not a lot to zero iron each) and 1 reinforced glass (something something small-amount-of-iron, or can even be made without any iron)
These produce 30,000 EU per unit of lava, and output at a rate of 25eu/t. This means they end up consuming .0012 lava/t, so you'll need exactly 6 to consume all the lava being produced. (80 iron, for 619 iron total so far)
Now, going back to the recycler array, while you COULD use 16 igneous extruders (112 iron) to fuel them, there are other, better sources of items, and even that cobblestone can be used in a more efficient manner. (cobblestone pipes can be recycled, so rotary macerating then induction furnacing 1/3 of the cobble, then auto-crafting it into pipes is a fast and easy way to use it more effectively) Also, it should be noted that normal cobble generators are capable of producing twice as much cobble per second as an extruder, though also are laggier in large amounts.
I'll just assume you'll use the 16 igneous extruders above, though. (731 iron)
I do believe I covered most of it... Oh! and you'll need anywhere from 1-0.5 aqueous accumulators, which is an additional 9-4.5 iron (740 total)
So, all in all, it should cost around 740 iron in total (plus a bunch of gold, silver, redstone, etc.) to produce 300 EU/t.
So no, this probably isn't the most efficient way to spend your iron to make EU, but it is reasonably space-efficient, straight-forward, can be done anywhere but the nether, and, if you happen to be acquiring 8 items per second in overflow, it can be a great way to use that for a little extra power. (And you can use it to instead make the infinite metal generation suggestions above)
Although, honestly, there are better ways to produce power from your space, especially when using steam boilers. (Shameless self-plug:
http://www.computercraft.info/forum...um-replacement-and-rp2-automated-flax-farmer/ )
Also, going back to the original topic:
Currently, my obsessive idea is the a very interesting power generation scheme I've just recently cooked up, where a large, turtle-and-combine wheat farm (it will also grow potatoes, since wheat grows faster in rows, and ignores other crops planted in those rows) will produce wheat and seeds for mycelium production in moisteners, which can be centrifuged for clay, mushrooms, and sand in an industrial centrifuge. The clay will be auto-crafted into blocks and then macerated, then the dust will be electrolyzed into aluminum, sodium, (which can actually be very effective liquid steam boiler fuel) lithium, (which can be burned or combined with wolframium in a fusion reactor for iridium) and silicon. (probably to be thrown away)
The mushrooms will be centrifuged into methane, which is also an effective boiler fuel, and the sand will finally be sent off to the real powerhouse of the system; turtle-turbary peat bogs.
peat, when cooked in a furnace, can produce 1 ash, 2 of which can be electrolyzed into a carbon cell. However, burning it in a peat engine not only produces MJ, but also produces 1.5 ash instead of 1, increasing the carbon output of the set-up, which is what I really want.
Carbon can be combined with 4 hydrogen in a chemical reactor to produce 5 methane, which is a very effective liquid boiler fuel.
Also, due to some very convenient math, 4 engines provide enough ash for 3 chemical reactors, which is 15 methane every 250 seconds, for every 4 engines.
Unfortunately, I have no idea about how many moisteners I should need, nor their output, which I'll need to be able to determine how much total yield I'll get out of this set-up.