To go from point a to point b at a cruising speed of v, you need to speed up to v, maintain v, and brake to a complete stop. Analyze each separately:
1. Speeding up. Since this takes so little distance (compared to the total length), let's ignore resistance. This speedup takes KE = ½·m·v² of energy to do.
2. Maintain speed. This means spending energy to resist air friction (drag) and ground friction. Drag is ½·ρ·v²·Cd·A [0]; friction is (at most) μ·m·g. Energy is force × distance, and since the distance at top speed is effectively constant, it's going to be O(m + v²) in this case. [1]
3. Braking. You can theoretically recover -½·m·v² here, but most of the energy is going to be lost to heat.
So, when you sum these up, the energy it takes to drive a vehicle a certain distance is linear in mass and quadratic in speed.
[0] The drag coefficient, Cd, is really weird. At low speeds, it's actually proportional to 1/v instead of a constant (i.e., drag is linear in velocity, not quadratic).
[1] Yeah, the article says that energy is proportional to v³. It's actually wrong--the power is proportional to v³. Whether or not the total energy is proportional to v³ or v² depends on what you hold constant: if you hold total time constant, it's v³, but if you hold distance, it's v².
This doesn't change the conclusion; you still end up with a low spot on the graph. It is just that the graph looks more like a parabola.
Does anyone know of a similarly simple explanation of how much energy a vehicle must use relative to it's weight?