Revit coordinates workflow - separate building and terrain

[amarkovi]

I would appreciate if someone could explain what would be the best way to model a project consisting of separate models for the terrain and a building, concerning the coordinates - in terms of setting everything up and using shared coordinates to link the models.

This is our workflow:

1) DWG site survey, with real world cooridnates (x,y,z) - usually it is a 2D DWG file which is used as an underlay in the Site view and from which we acquire the coordinates in Revit. We then use a CSV file with the 3D information of each survey point and create a topography from it. So this first file has a DWG layout of the site and a 3D model of the terrain, placed according to the coordinates, both in longitude/latitude as well as elevation (from sea level).

This file containing the topography has two levels: Sea and Building. The Sea level is at elevation 0, the building level is where the "0.00" of the building will be.

2) We model the building in a separate file.

3) The building is inserted in the terrain file as a Revit link and placed where it needs to be: both in plan view and in elevation.

4) The terrain is linked into the building file via shared coordinates. Now the terrain can be modelled more in detail around the building, and the building can also be adjusted in relation to the terrain - if needed.

5) Once all the modelling is finished we produce the drawings from the building model (plans, sections, elevations, site plan...).

Some of the documentation needs to show the elevation relative to the 0.00 of the building (finished floor of the ground floor level, for example), other views need to show the elevation relative to sea level. 

What should be done with the project base point and survey point in the terrain file and what in the building file, to avoid problems? What would be the best way to setup everything to avoid problems?

[SteveKStafford]

I do what you describe routinely. I don't differentiate between Project North and True North in the Site model (though it is possible). I regard it as the real world as if there is no Project North, that's something that is more meaningful or useful in the building model.

Most the projects I support are hosted by ACC so using Publish Coordinates is not an option. My approach focuses on using Acquire Coordinates (AC). When I link the Building model to site I place a marker family that identifies a prime grid location (Grid A/1) so when I link the Site model to the building I can use the marker to orient the site model to the building oriented to a Project North (easy to draw).  The marker is also raised up to ground floor elevation so it can be used to align the building vertically. I store the grid intersection info and the elevation in the properties of the marker family so I can query it from within the building model. Then I use AC and select the site model.

To recap: I use AC on the DWG source in the site model (keep the source near Revit's origin) and then acquire coordinates from the site (RVT) model in the building(s). Other disciplines use AC on the building (architecture) model. Revit building models use by Internal Origin to Internal Origin for linking.

If the building must move at some point in the design process I do that in the site model and place a new marker and repeat the process in the building model to resolve it's position, using Reset Coordinates and then Acquire Coordinates again. In some cases I will use Specify Coordinates at Point to update the building model with it's new position. Either way works.