Best practice modeling a house base

[Rastisan]

So, I'm modeling an ordinary house struggling a bit with the changing composites of the perimeter walls at the base of the house:
1. the typical perimeter wall of the building is made from cavity brick, insulated with mineral wool finished with plaster.
2. 60 cm above the external terrain I want the a different plaster finish, insulation switch to XPS and have the wall hydro insulated with bitumen
3. Below the terrain I want to keep the same XPS insulation but change the facade finish to drainage foil and cavity brick to cast concrete.

Now, I have achieved this with 3 different composites and modeled 3 different walls that are stacked on top of each-other. I assume all would be fine if the terrain were flat and consequently the walls would stack on top of each-other in flat layers. But since the terrain is terraced we have an external staircase on one end of the building and an external supporting wall on the courtyard. As you can imagine the three different composites "follow" the inclination by the staircase and a "drop" by the wall. This means that by the staircase we have the #3 composite under the staircase ramp that meets the #2 composite above the stair ramp and #1 that meets with the #2 at a inclined line parallel to the stair ramp.

I have modeled this with boolean operations with no problems. The issue is in the floor plan which has problems with displaying 3 different composites in the same story. The walls overlap or are displayed on story "foundations" even though they are on the correct level in 3D.

The other problem are the building corner where we also have the external support wall where in 3D and 2D i get some "runaway" geometry. I think this is so because at this (and other) corners the wall elements do not meet in 1-1, 2-2 or 3-3 fashion but different wall composites met in a 3-2 or 2-1 fashion and AC does not know precisely how to join them.

Another problem is the door. Since it belongs to the #1 composite which ends 60 cm above ground the door also ends there. Or more precisely: Even though the wall doesn't reach the ground the door does (bravo), but then the #2 composite "runs over" the door at the base.

Based on all of the above, what would be the best practice to model perimiter walls with above characteristics but without issues?

If possible I would avoid separating the wall core from the rest (modeling just the wall and the using 3 different facades). Separating the core from the rest of the layers makes sense with slabs but I would like to keep the walls in one composite if possible.

Should I redefine my 3 composites in such a way that the #1 stays the same . The #2 and #3 receive a removal of the core layer. Then I model the #1 all the way from the roof to the foundations. At the base (60 cm above the terrain) I then boolean subtract the facade layers and apply the #2 composite. For underground I apply the #3 and also model a concrete wall "inside" the brick wall (Archicad will automaticaly adjust the brick geometry as I have noticed).

Opinions and best practice principles greatly appreciated!

Rastisan

[Barry Kelly]

Rather than having 3 separate walls that you stack on top of each other (which is perfectly fine to do) you could also look at using a single complex profile wall.

You can even add modifiers to the fills in that complex profile so that you can adjust heights and thicknesses for the various components in each wall you place in the model.

i.e. you split the length of the wall when ever you need different heights compared to the adjacent wall.

It is the same complex profile but you can now control the heights in each wall individually.

Barry.

[Rastisan]

I understand the principle, thank you for this recommendation. As I'm familiar with complex profiles I think this method would work fine in straight situations (where the outside terrain is flat and the 3 different composite run in a straight fashion ), probably also in the around the support wall (achieved with dividing the wall as you suggest), but I don't know if you can set up the composite profile in such a manner that the changes of wall composite would follow the geometry of the staircase (meaning starting at the lower level and then gradually rise to the higher level in a diagonal fashion). This would mean that you can adjust the modifier height parameter for both the start and finish of this particular division of the wall adjacent to the stair ramp. is this possible?

Thanks again,

Rastisan

[Barry Kelly]

Walls can not rake and the same goes for the modifier heights if you are looking at a gradual sloping change in height.

You have to split the wall and adjust the heights for each section so it steps up the rake.

Barry.

[Rastisan]

No, I want the this kind of geometry:

[Rastisan]

Goin back to my initial post: Here are screenshots of the issues I get with "open corners" and overlaps in 2D when modeling the house base using 3 different composites:

As you can see, there are overlaps, discontinuous skins (insulation, hydro insulation) and other weird stuff.

I will now try to remodel these walls as  I described in my initial post and report the results.

[Barry Kelly]

If you want the walls to rake at the stairs then you will have no choice, you will have to use separate walls stacked on top of each other.

Trim (Solid Element Operation) the top of the lower wall to a roof that is in a layer that can be turned off.

SEO the bottom of the upper wall to that same roof (zero thickness roof).

When stacking separate walls on top of each other, you may need to place them in separate layers with different intersection group numbers, so you can turn some off in plan, those you don't want to see.

In elevation and 3D you will have all the wall layers on.

I am not sure that you will get the results you want if you have the walls set to 'Projected with Overhead' and you are relying on a Floor Plan Cut Plane height.

All of the walls will be fighting each other if they are all in one layer even if there heights are set correctly.

Barry.

[Rastisan]

I manged to remove weird connections (runaway corners, overlaps and such) with changing the values of layer intersection priority groups. However I discovered, that you have to do that for 3D view and plans separately. I find this strange.

To explain: I have no problem with layer visibility being separate for 2D and 3D since it doesn't concern physical properties of structures, only their visibility. But in the case of intersection priorities they relate to physical properties by my understanding, Isn't it weird that you can have different layer intersection priority groups for various views? Am I missing something here or not understanding correctly something?

I apologise if this kind of questions sound silly but I want to avoid potential problems downstream that would result from errors in this stage of developement of the project.

[Barry Kelly]

In plan you may not want to see all of your overlapping walls.

That is why I suggested putting some on hidden layers.

Also in plan, you may not want all of these walls to trim to each other, hence the different intersection group numbers for the layers.

In 3D/elevation/section, you will want to see all of the walls (layers) and you will probably want them all to trim with each other, hence all layers having the same intersection group number.

The view in plan is different to the 3D/section/elevation views.

So you may want the elements to interact differently with each other.

Barry.