My theory is that they implemented an equivalent to traction control, but with no sensors at the wheel. It's all internal to the KERS, and probably a software hack. For example, if a drop in current from the KERS indicates the motor/generator needs less power even though the throttle is open, you might be able to infer wheel-spin without measuring it directly, and momentarily shut down the KERS to reduce power to the wheels.
Even if they audited the code, it might look like code that protects that motor/generator from over-revving than a software traction control.
I've been lucky enough to go to MC for several years and usually get at least one session on the outside of Rascasse. Everyone has problems accelerating out of that corner, especially the first 10-15 laps. They all get on the power, and oversteer. You watch their rear tires get shiny, the rear slips out of line, and you hear them back off the throttle.
Except Vettel. Constant power, a strange exhaust, and never ever oversteering. I remember jumping up and down the first time I saw it: "THAT'S why he's so fast--RB have different technology. No one else can go though the turn like that. No one."
I never did figure out what it could be, but your theory makes perfect sense. Thanks.
How it would work is to use the KERS to charge during engine acceleration. A brilliant idea, and it wouldn't have to be plumbed into the standard McLaren ECU all the teams have to use.
Lewis Hamilton stated that Vettel's RBR car was getting on full throttle 65 feet(!) sooner than all the other cars at a corner in Singapore...which means that there's clearly something going on that isn't just Newey's aero.
I'd be surprised if they hadn't looked at something like this, but it wouldn't explain the suspicious sound. The electric motor that KERS controls would be relatively silent compared to the V-8... (Oh how I miss the V-12's...)
Even if they audited the code, it might look like code that protects that motor/generator from over-revving than a software traction control.