Last week, one of the students in my course on BIM tools for sustainable design created a model that had an excess of shading surfaces and as such was rejected by Green Building Studio as invalid geometry. In order to remedy that I suggested that the same model could be processed within Revit Architecture 2009 since, as I have mentioned before, it has the ability to strip the model of excess shading surfaces. At that point this seemed like a good idea, right? Wrong. The model, which had been 4 MB in size, grew to be 48 MB in size and needless to say it was not applicable for simulation. We submitted the XML file to GBS for their analysis, and this is what they found:
| RAC 2008 | RAC 2009 | |
| int.floor | 232 | 3965 |
| int.wall | 232 | 305 |
| ext.wall | 376 | 8830 |
| roof | 3 | 665 |
| raised floor | 183 | 650 |
| ceiling | 0 | 2 |
| shade | 924 | 863 |
| und.wall | 40 | 2068 |
| und.slab | 33 | 1066 |
| opening | 1311 | 1457 |
| spaces | 76 | 86 |
This morning I decided to put a simple model created in RAC 2008 to the test and, since I suspected that having the structural elements might have something to do with XML output, I placed a simple round (pay attention to the word “round”) concrete column in the center of this test structure. Analysis of the 2008 XML file did not show anything unusual, so I proceeded and opened the same file in RAC 2009. Not to my surprise the newly generated 2009 file quadrupled in size and these are the findings:
| RAC 2008 | RAC 2009 | |
| slab.grade | 1 | 32 |
| roof | 1 | 32 |
| ext.wall | 4 | 36 |
The round column is being defined with 32 segments, and just for the future reference when modeling curvilinear geometry think of the polygon count, as even in these BIM days it still matters.
The conclusion is that the structural element is subtracted from theroom volume, the void is therefore tessellated, and surface type enumerators are assigned to the newly generated surfaces. Interestingly the enumerator for the vertical faces is the exterior wall, and this poses an alternate problem related to the overall calculated mass of the analytical model. Now, although in both versions of RAC the structural element is designated as Room Bounding, somehow RAC 2008 was semi-smart about it and it was ignoring them, where RAC 2009, possibly due to the enhancements with the room object, is quite happy in exporting those structure bound faces as well as exponentially increasing the size of the gbXML file.
In order to make previously created RAC models applicable, one should declare all of the structural elements as non Room Bounding and by simply doing this the gbXML output from RAC 2009 will be almost identical to the one from RAC 2008. This will increase the usable room area within the project, but we can all agree that this is the small price to pay for being green, or maybe not?
The only slight difference in the file size will be caused by the changed nomenclature within the 2009 generated XML file that is used for more eloquent description of exported surfaces.You can download both the 2008 and 2009 samples and their corresponding XML files by following this link.
In order to make previously created RAC models applicable, one should declare all of the structural elements as non Room Bounding and by simply doing this the gbXML output from RAC 2009 will be almost identical to the one from RAC 2008. This will increase the usable room area within the project, but we can all agree that this is the small price to pay for being green, or maybe not?
The only slight difference in the file size will be caused by the changed nomenclature within the 2009 generated XML file that is used for more eloquent description of exported surfaces.You can download both the 2008 and 2009 samples and their corresponding XML files by following this link.
