An investigation on effort and applicability of a computation of a space partitioning from boundary representations of building components and rooms
Final Report Abstract
The project focusses on an algorithm to derive topological information from a given boundary representation of a set of objects. The field of application is digital building models which can consist of thousands of objects describing either components and built-in components or rooms and spaces. Relevant are: • intersections of objects • gaps between objects and voids in the interior of a building which are locations where no object is specified • the exterior • all topological neighboring relations. The research focused on the development of an algorithm to handle realistic models. The algorithm was implemented in Java. The challenge in this field is the handling of inexact computation caused by the restricted storage place for variables in the computer. The approach in this project bases on integer variables to store the coordinates of points. Given points of given objects are rounded to the nearest integer and then inserted into a mesh of tetrahedrons. The surface of each object is stepwise reconstructed in this mesh. Intersection points are computed. A local refinement can be necessary to achieve a valid mesh. The coordinate system is refined for this purpose. Results are convincing. Present modeling tools used in practice are able to compute clashes. However, they are not able to compute gaps and voids in any cases. In addition, they have inaccuracies and faults in the identification of the exterior. The research conducted in this project bases on findings in the field of mesh generation. However, research in mesh generation focuses typically on specific properties of a generated mesh, e.g., to ensure good results using the mesh for a numerical analysis of the physical behavior of objects. They do not focus preliminary on the above-mentioned topological information that are specifically relevant for lots of applications in architecture and civil engineering.
Publications
-
Accuracy Aspects when Transforming a Boundary Representation of Solids into a Tetrahedral Space Partition, in Abualdenien, J. et al. (Ed.): Workshop on Intelligent Computing in Engineering (EG-ICE 2021), pp. 320 – 329, ISBN: 978-1-7138- 5547-7, 2021
Vetter, J. & Huhnt, W.
-
Space Partitioning asaHolistic Alternative toTraditional Geometric Modeling Workflows intheAEC Industry. Lecture Notes in Civil Engineering, 415-426. Springer Nature Switzerland.
Huhnt, Wolfgang; Vetter, Joanna Zarah & Sternal, Maximilian
-
Transforming Building Components intoaSpace Partition toIdentify Indoor andOutdoor Spaces inDigital Building Models. Lecture Notes in Civil Engineering, 403-413. Springer Nature Switzerland.
Vetter, Joanna Zarah & Huhnt, Wolfgang
