RW - 7 Margarita Belen Ianne E - 20230203

Name : Margarita Belen Ianne E.
Vito Date: February 3, 2023
Course/Section: CEM221_D01_2Q2223
Research Work 7
Title of Journal: BIM-based graph data model for automatic generative design of
modular buildings
Author/s: Vincent J.L. Gan
Objectives:
The study's goal was to provide a general overview of the application of BIM for
optimizing buildings and the benefit of quick design generation for prefabrication.
Methodology:
The investigation was carried out in three stages. First, information requirements for
BIM, including IFC MVD, were identified. Next, the data model for modular buildings was
graphed. Finally, graph-structured data was transformed for autonomous generative design.
By consulting both public building regulations and contacts with professionals, the
necessary data for the design was determined. The spatial, topological, geometric, and semantic
information needs were pertinent to the design and prefabrication of modular buildings.
To explicitly articulate the spatial properties for volumetric modules as well as the
topological relationships, geometries, and semantics for various precast components in
1
modular construction, the graph data model was created after the requirements were gathered.
In this way, the spatial arrangement of modular buildings was described and altered using
spatial features. For the convenience of information querying, graph-structured data were
stored in matrices or relational databases.
Results and Discussion:
To organize the work that has to be done by the many stakeholders, a map of the process
has been created. This map links a sequence of products, resources, and data exchange
requirements throughout the various stages of design. The process map serves as a reference
for communications with professionals working in the industry, as well as published
guidebooks and construction laws. The procedure flowchart contained a total of six stages:
inception, conceptualization, detailing, pre-construction, and construction.
The Dynamo Studio and Python programming language combination is used to create
the BIM-based algorithms. The methods begin by mapping the three-dimensional spaces of a
modular building into a collection of vertices in the adjacency and connectivity graphs. After
that, the algorithms automatically derive the spatial relationships between the vertices using a
set of edges. To be more specific, the algorithms consider the space position of each volumetric
module, also known as the centroid coordinates (xMk, yMk), in order to add the relevant vertex
to the adjacency graph A(G) and the connection graph C. (G). After then, the algorithms
compute the adjacency matrix, which is denoted by R, and generate edges, denoted by Eij,
between adjacent modules in A. (G). Also, the algorithms compute weigh tij in the connectivity
matrix ('R) for neighboring modules that have interior openings, and then add the relevant
edges ('Eij) to the connectivity matrix (C) (G). Each volumetric module, denoted by the symbol
2
Mk, is made up of precast components in addition to its spatial characteristics (i.e., framing
elements, joints, and planar elements). The topological relationships between precast
components, as well as their geometries and semantics, are extracted separately and associated
with corresponding graph vertexes. This results in a comprehensive description of BIM as a
relational graph that can be used for a variety of information queries.
Conclusion:
The purpose of this study is to introduce a unique graph data model that is based on BIM
and has the potential to automate the generative design of modular buildings. To construct
modular buildings, it is necessary to be able to describe the spatial properties associated with
volumetric modules, as well as the topological relationships, geometries, and semantics
associated with precast components. The IFC MVD standard was developed to meet these
needs.
These findings contribute to the standardization of design information in precast regions by
identifying the information needs for developing modular buildings. To convert the required
information in BIM to the graph data model and to compute the necessary model parameters
in compliance with the proposed MVD structure, data translation algorithms are being built.
These algorithms are being developed by the data transformation team. In addition to the
algorithms that read and transform information from BIM, we are developing a prototype
system that will be interfaced in BIM-authoring software to make it possible to generate the
design of modular buildings. The prototype system has the capability to parametrically edit the
spatial properties of volumetric modules, build the 3D model in BIM, and evaluate the fitness
of each alternate in the process of discovering a set of optimal solutions. The applicability of
3
the approach for planning the spatial organization of modular off-site construction is evaluated
by testing it on the optimization of a modularized apartment.
This study adds to the standardization of design information for modular buildings, which
is in line with the precast industry's IFC4Precast. The graph formulation gave us a new way to
describe building features, making it easier to use computational methods (like metaheuristic
algorithms and machine learning) to find the best design solution. As part of the next steps for
this study, the graph data model was recommended to be combined with deep learning
algorithms (like graph convolutional net and generative adversarial net) to improve intelligent
design in precast areas. Also, the graph data model can be made bigger by adding the
computational power of evolutionary algorithms to make design optimization more reliable
and effective.
Reference/s:
Gan, V. J. L. (2022). BIM-based graph data model for automatic generative design of
modular buildings. Automation in Construction, 134(November 2021), 104062.
https://doi.org/10.1016/j.autcon.2021.104062

RW - 7 Margarita Belen Ianne E - 20230203

Uploaded by

Copyright:

Available Formats

RW - 7 Margarita Belen Ianne E - 20230203

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

RW - 7 Margarita Belen Ianne E - 20230203

Uploaded by

Copyright:

Available Formats

Name : Margarita Belen Ianne E.

Vito Date: February 3, 2023

Author/s: Vincent J.L. Gan

topological relationships, geometries, and semantics for various precast components in

stored in matrices or relational databases.

Results and Discussion:

for communications with professionals working in the industry, as well as published

inception, conceptualization, detailing, pre-construction, and construction.

with corresponding graph vertexes. This results in a comprehensive description of BIM as a

relational graph that can be used for a variety of information queries.

volumetric modules, as well as the topological relationships, geometries, and semantics

These findings contribute to the standardization of design information in precast regions by

by testing it on the optimization of a modularized apartment.

computational power of evolutionary algorithms to make design optimization more reliable

You might also like