Are there common techniques in openstreetmap to manually control which elements are to be displayed at a specific zoom level?
For instance: motorways usually consist of two ways, one for each direction. At lower zoom levels displaying only one of these ways is sufficient. So in a local copy of OSM I marked one of these ways with a tag like this
In general adding tags to direct rendering behaviour is not a good idea (although it has a long history on OSM).
What you are talking about is geographic generalisation. There are a host of techniques for this sort of thing, some tools such as Qgis may do simple generalisations, but for many areas this is quite complex & often requires quite fancy coding.
Turning dual carriageways into a single line string at low zooms is one such issue. Some people do appear to have solved it. It’s potentially easier if the two carriageways can clearly be linked & then one can do something like draw the shortest line between the closest nodes on each carriageway & bisect those lines to provide a series of points to build a median line string. There are numerous other ways (e.g., buffer each carriageway & merge & find the medial axis) etc.,
For your purposes any query which reliably selects one of two carriageways will work without the need for tags (it could be direction, or some are tagged explicitly).