@bresch Yes, I get it.
The question is, we cannot simply replace the v with the velocity components at each axis since the drag is proportional to v^2. Or else the product of the drag force at each axis will not consistent with the total drag force applied to the inertia frame.
Note: Because this topic is already solved, I create a new topic for this separated issue. We can have further discussion here:
We can have further discussion here: