the cause is the desire to get a system that can be described in only two vectorial parameters that are perpendicular to each other and not time dependant.
The effect is, that instead of dealing with three time-variant equations, you only have to deal with those two parameters do describe a system.
Knowing the point of operation, re-transforming dq-values allows you to easily determine the necessary switching state of the inverter, something that will be a highly complex thing if you try without the transformation.
I know the benefit of dq-transform but i want to know what are effects of coupling which force us to decople the dq variables? for reference, i attached one paper you can see it
basically, if you transform the real world system into dq-components, you get access to the torque producing component and the flux - or speed -generating cpomponent.
As these components are perpendicular to each other the do not interact which means, you can create any point of operation by defining speed and torque independently.
This behaviour is the key to control an electric machine precisely -->Y Field oreinted control.
the paper describes just what I expected, though it takes the look backwards.
The AC-conversion referred to makes use of the dq-transformation to consider the grid as a rotating machine, described in dq-components, thus allowing to make use of all the well known math to control the DC-side.
The effect simply is, that well established methods and models can be adapted to the DC-conversion.
I'm sorry, but maybe I still didn't get the core of your question.