Very nice - would be good to use this for the MG4 if we could get some info on things like rolling resistance, drivetrain efficiency etc. Some of these could be estimated if someone could perform some rather tedious tests, I suppose.
Tedious is the right word.
Roll down tests (i.e. at zero power) could be done, just need a good data logger for speed readings. With enough variance in starting speeds it is possible to tease out CdA and Crr with some regression analysis.
Alternatively it's also possible to do independent testing of Crr, and then assume it is a fixed value (which if using the same test venue is not unreasonable), and assign all other energy consumption to the remaining factors. It's just balance sheet accounting for energy.
The hard part is accounting for uncontrolled variables (wind being the obvious one). With the right venue that can be minimised, or measured with suitable data loggers.
Best however is to eliminate/minimise/control or tightly manage such variables.
There is a long (formerly abandoned) tunnel in the UK set up for just this sort of thing:
Catesby Tunnel is a 2.7km underground test facility for aerodynamics and vehicle performance testing located in the heart of the uk.
catesbytunnel.com
Alternatively, if there was a way with OBD to capture/log the car's power and speed data at high frequency, then testing on a circuit (e.g. with lap times order of magnitude 1-minute) by performing multiple laps enables use of
a really neat visual diagnostic technique (originally described by Prof. Robert Chung) as virtual elevation testing.
The circuit needs to be "not twisty" like a race track - more like a smooth oval - to reduce variance with tyre scrub (edit: and avoid the need for any braking*). Some elevation change on the circuit can be advantageous but it doesn't need much - it's pretty interesting how good of a diagnostic tool this is. I used it a lot with cycling, eventually morphing into a real time data capture/modelling system at the velodrome giving CdA outcomes instantly (rather than doing post-hoc processing of logged data).
At motorway speeds, aero drag probably dominates by some reasonable margin, I would think.
90+% I would expect when on flat road. Once you add in a gradient though then increasing gravitational potential energy can become the dominant factor.
* this could be problematic with no way to turn off regen braking