I can't explain it any better, engine to wheels efficiency in the 90s (high 90s with some optimizing) for direct drive, with experienced hypermilers showing us how to get stellar mileage out of even a basic 5 speed taking full advantage of changes in situation. I can easily imagine a pusher trailer that can also be hypermiled or "right sized".
Show me the efficient generator trailer, not the pipe dream.
here is a modern tdi:
1. motors are not always at peak efficiency.
A. generally not efficient at slow speed, starting at about 0% efficient at very low rpm. No gear box means more time spent at low rpm, so lots of time at slow speeds the generator is at more of disadvantage since that is how the power flows to the wheels.
B. efficiency is still largely load dependent. Peaking around 90-95%
2. Controllers are not %100 efficient, say %90 (large heat sinks)
3. The generator has losses going from mechanical to electrical, like %80 efficient.
4. driveline losses, say %95 efficient.
These are just swags, but let me total them up for this example so we have at least a starting point. Lets say demand is optimal.
.95 * .90 * .80 * .95 = 0.65 efficient in steady state mode with optimum load.
compared to a tdi that is %95 efficient (assuming some effort in driveline optimization) at driving the wheels under optimum conditions.
Now if you want to use "peak bsfc" as your argument, we are assuming peak bsfc in both "pusher" and "generator" here.
But if the tdi direct drive has to "compete" with the generator, the "at the wheels" bsfc for the tdi is like 206 gm/kwh. Whereas the same efficient engine running a generator is like 301 gm/kwh.
So lets play the load the battery for peak bsfc game now, but first note that pretty much any operating condition above 3 bar will be to the advantage of the direct drive in this graph already.
5. lithium batteries are 80%-90% in charge/discharge efficient under optimal conditions.
6. Say you have a %90 efficient charger and bms, which is also dependent on state of charge and current.
So aside from the fact that you can't ever touch most of the bsfc graph with a generator in front of this tdi driving the wheels, you are looking at a generator efficiency of
.65 * .90 *.90 = 53% efficient when loading the generator for bsfc (assuming all the electrical are in a fairly optimal operating range). That is a wheel bsfc of 369, below the very bottom line on the bsfc chart.
Now you may take issue with the specific numbers here, but the point is clear, there are a lot of losses in the series proposal.
I have heard speculation about higher peak BSFC with rpm targeting but I have never seen any actual data, nor do I know that a cvt wouldn't still be more efficient than an "electronic torque converter" in maintaining target rpm.
Sending all your ICE power through a generator to drive your EV wheels certainly seems like an inefficient solution all around, and requires %54 more engine in steady state for the same power at the wheels in this example.