Since last night's stream explained that FTB's update will take another day or two in testing to properly balance the new worldgen settings, I went ahead and made a 1.4.6 testbed install with only Redpower, TE and NEI.
I tested with a powered furnace, since that one will tell me in detail how much MJ it is consuming at any given time. I was able to confirm Zjarek_S' findings. 1000 W = 1000 J/s = 1 MJ/t = 20 MJ/s. By extension, 50 J = 1 MJ.
And, to finally answer Tylor's question: The windmill in DW20's video, running at just over 50% speed, was producing 2.5 kW and should therefore generate about 2.5 MJ/t when feeding a blulectric engine. If the scale is linear, this should mean that at 20% speed it produces 1 kW = 1 MJ/t and at 100% speed it produces 5 kW = 5 MJ/t.
I also confirmed that, at least when directly hooked up to a powered furnace, the blulectric engine will throttle itself to the required amount of power. Once the furnace's internal storage is full, the engine will not consume more than roughly 2 kW (= 2 MJ/t), no matter how many solar panels you add.
I also observed that voltage indeed plays a role in power output, although it seems that there's some handwaving going on in a few cases. For example, if you place an empty machine (such as a blulectric furnace) next to a solar panel, the solar panel will charge it from 0V to 100V at a pretty much constant speed, seemingly suggesting that voltage is irrelevant. However, if the same machine is actively doing something, then that doesn't apply - if the voltage drops, then power generation will drop as well, meaning the machine drains its reserves faster and faster the lower its reserves drop (remember, in Redpower, voltage is equivalent to the charge status of the network).
How to confirm this in experiment: take a blulectric furnace and hook it up to six solar panels, let it charge to full. Then give the furnace some cobblestone to smelt. Since the six solar panels generate 1.2 kW, and the furnace only requires 1 kW at peak efficiency, it will stay fully charged even while working. Now, break one solar panel. The five remaining ones still produce 1 kW, just as much as the furnace consumes. But due to miniscule distance-based losses and the brief dips directly after the breaking where the system restabilizes itself, voltage is ever so slightly below 100V and begins to decline very, very slowly. Millivolt by millivolt first, but the lower the voltage goes, the less power the panels produce, and therefore the voltage drops faster and faster until it collapses down to about 600W. The furnace won't consume any power below 60V, so the voltage will not drop lower than that; but, in this state, the furnace will only run at 60% speed. Now, add in the sixth panel again. Previously, at 100V, this was much more than enough to keep the furnace topped off with power at all times; now however, it barely manages to bump up efficiency by some 10%-12%, still a ways off from charging back up. In fact, you can add a seventh and an eigth solar panel, and the voltage will still remain stuck at 60V because total output is still only 0.96 kW at this point.