Automation in Construction 128 (2021) 103743

Contents lists available at ScienceDirect

Automation in Construction
journal homepage: www.elsevier.com/locate/autcon

Loose coupling of GIS and BIM data models for automated compliance
checking against zoning codes
Yelin Demir Altıntaş *, Mustafa Emre Ilal
Department of Architecture, Faculty of Architecture, İzmir Institute of Technology, Gulbahce Campus, Urla, İzmir, Turkey

A R T I C L E I N F O A B S T R A C T

Keywords: Modeling urban data is crucial for supporting automated code checking processes. Ideally, digitally modelled
Automated compliance checking building codes and urban data should be retrieved from municipalities, and the digital building model should be
BIM checked automatically based on the collected information. However, BIM tools do not contain and do not allow
Data coupling
managing geographical information at a neighborhood scale. Current GIS applications also do not store all of the
GIS
Interoperability
information required by building codes. Even if they did, interoperability between GIS and BIM environments are
problematic.
This paper describes the development of a zoning domain model for automated compliance checking of
building projects. The proposed model is illustrated through a proof-of-concept GIS application, where geometric
and semantic data are stored, queried and exported as a GML file. Use of this data model for automated code
checking is an example for how GIS data can seamlessly complement BIM data making expansion of BIM schemas
unnecessary.

1. Introduction divisions and each develops its own tools while a common data standard
is missing.
During architectural design processes, architects need data on the Automated compliance checking against zoning codes is one of the
surrounding context. While utilization of some contextual data is at the areas that suffer from shortcomings in BIM-GIS integration. Code
discretion of the architect, some contextual data represent constraints checking requires neighborhood scale data integrated with the building
that all designs should meet and their utilization is mandatory. Local data in a common environment for a complete checking process. In ideal
building regulations define and reference such data. Two examples are code checking workflows, sets of rules from all applicable building codes
setbacks, which are minimum distances from the building to the parcel will be collected directly from authorities, required zoning information
borders, and grade elevation of main access roads for parcels that in GIS environment will be retrieved from the local municipality, and
determine the reference ground level for projects. Unfortunately, in compliance checking will be conducted on submitted building infor­
existing design environments, it is not possible to access such data on the mation models.
surroundings. Building Information Modeling (BIM) tools integrate and Industry Foundation Classes (IFC) data model acts as an international
manage information from various disciplines including entire design, BIM standard for interoperability and is accepted as the standard most
construction and maintenance processes, but they are limited in storing likely to succeed [5]. Yet, it does not represent enough entities or at­
and managing geographical information on the surrounding landscape tributes which are necessary for code checking. Many of the concepts
of buildings [1,2]. Furthermore, BIM models are not able to query to­ mentioned in building codes, such as block, parcel, setback distance,
pological relationships and store topological data which is in fact garden, slope, land use, roads, streets and topological relationships be­
required by automated code compliance checking processes [3,4]. It is tween geometric entities etc. are lacking in IFC. As a result, IFC, on its
Geographic Information Systems (GIS) that have capabilities for inte­ own, is not adequate for automated compliance checking processes.
grating and managing geographical information. Yet, interoperability in CityGML is the commonly used open data format for representation,
geographical information exchange between GIS and BIM environments storage and exchange of city and landscape models. It includes many
is problematic. These two distinctive domains have clear professional classes mentioned in building code texts [6], but not all of the required

* Corresponding author.
E-mail addresses: dmryelin@gmail.com (Y. Demir Altıntaş), emreilal@iyte.edu.tr (M.E. Ilal).

https://doi.org/10.1016/j.autcon.2021.103743
Received 19 January 2021; Received in revised form 29 March 2021; Accepted 25 April 2021
Available online 11 May 2021
0926-5805/© 2021 Elsevier B.V. All rights reserved.
Y. Demir Altıntaş and M.E. Ilal Automation in Construction 128 (2021) 103743

classes, attributes and relations exist in this data model. The “con­ development of BIM and GIS tools, and professional divisions were the
struction order” of the parcel, if the parcel has a “front garden” or not, if primary causes for the differences between BIM and GIS. Although BIM
the parcel is a “corner parcel” or not, the “buildable area” of the parcel and GIS differ from each other, both can benefit from each other if they
are all attributes and relations of the parcel class that do not exist in could exchange data effectively. Recent studies on BIM-GIS interoper­
CityGML. Some of the concepts mentioned in building codes are not ability commonly focus on developing new ways of integrating BIM data
represented anywhere. For example, consider the following rule. into GIS environment and vice versa, developing new extensions to data
models such as CityGML and IFC, developing new methodologies for
“Setback distances where there is a front yard and setback distances from
making data conversion and translation between data models, and
roads, green areas and parking lots should be at least 5.00 m.” (Izmir
developing independent domain models.
Municipality Housing and Zoning Code; Clause 27).
IFC and CityGML are the most comprehensive and representative
Properly implementing this rule in an automated code checking data models in BIM and GIS domains, respectively. Thus, many studies
system, requires buildings, roads, green areas, parking areas, adja­ aimed to convert and translate them to be compatible with each other
cencies and distance relationships between these objects to be modelled [7–14]. Some studies integrated BIM data into the GIS environment and
as part of the data set. vice versa [2,3,15–20].
Modeling the parcel and its relations to its neighbors is also not Data model extension studies add concepts that do not already exist
enough for a complete code checking process. For example, consider the in the IFC model and thus, expand the scope of the model by integrating
following rule; various domain data [4,21–24]. There are also data model extension
studies for geographical data involving adding building related data that
“In attached ordered blocks, if there is an existing building in any of the do not already exist and thus, expand the scope of the geographical
parcels within the same block, setback distances are the same as setback model [25].
distances used in that parcel.” (Izmir Municipality Housing and Zoning Even though many researchers have studied how to integrate BIM
Code; Clause 27). and GIS data and how to address all the differences, many challenges are
still being encountered and it is still very hard to integrate data
For implementing this rule in an automated code checking system,
the construction order and setback distance information of all parcels throughout the design and construction processes. Data access and
conversion/translation problems between BIM and GIS cause data loss,
that face the same road in the same block should be modelled.
The goal of the research presented in this paper is to 1) determine the incomplete and unreliable transformations of information, incorrect
GIS data on the surrounding urban context required for zoning regula­ mapping, and manual data re-entry. IFC and GML data standards are
tions in the city of İzmir and 2) propose a solution for integrating this incompatible as they are different in terms of representation, scale and
data with BIM data in order to facilitate automated code compliance aim. They are developed for different professionals and processes. For
checking against zoning codes. In this context, the study aims to show example, IFC data models contain much more detailed information from
that all geographical data related to a neighborhood and required for the various disciplines than GML. Some concepts in GML have no corre­
automated zoning code checking process can be modelled in GIS envi­ sponding entity definitions in IFC data model and vice versa. Thus, data
ronment and this geographical data can be coupled with BIM data, loss from IFC to GML is inevitable.
constituting a complete data set for automated code checking. To ach­ Data extension studies are also problematic since, the aim of BIM is
ieve this aim, a methodology is described in this study that follows an to support lifecycle of a building from design to construction and the aim
approach based on the development of an independent third-party of GIS is to store, analyze and manage geographically referenced spatial
platform. The developed methodology has the following stages: 1) data for urban scale studies. Thus, it is impractical to expect BIM to
Identifying the required data in the area of study and identifying how explicitly define geographical information; and GIS to manage detailed
much of the data can be modelled with BIM model or GIS model; 2) building data, and extend information storage capability of the two
Constructing BIM and GIS domain models by using the identified data in environments except their scope.
the previous step; 3) Developing an application that integrates the two Currently, there are various interpretations of how to best transform
a BIM file into GIS data and there is no accepted standard. One approach
domain models and manages them as a whole.
This research is limited to two-dimensional data and topological that is proving to be effective for many fields is developing independent
applications that draw data from both BIM and GIS environments and
relationships of two-dimensional objects. The complex situations
including the analysis of topological relations in 3D and usage of 3D GIS combining them in a unified domain model appropriate to the field
[7,26–31]. The domain model manages both the surrounding data
objects are left for future research. Additionally, the proposed domain
model is limited to neighborhood scale information including roads, represented in GML and the building data represented in IFC and be­
comes the base for exchanging information. This approach does not
sidewalks, blocks, parcels, buildings and their boundaries leaving the
building related properties such as door, window, wall, stair out of require BIM or GIS systems to extend their data models but depends on
new domain-specific models linking the two to be defined. In this study,
scope.
In short, this study pursues the process of integration between BIM an independent application is developed that draws data from both BIM
and GIS environments and combines them in one unified domain model
and GIS environments by bringing BIM and GIS data to a third platform
that is independent of BIM and GIS data models. The developed meth­ appropriate to the field. The merging of BIM and GIS creates a powerful
tool, which is then applied to an automated compliance checking
odology is evaluated through a prototype implementation. Such a loose
coupling approach has not been used for automated compliance process.
checking previously. The literature is summarized in the next section.
2.2. Automated code compliance checking
2. Related work
Automated compliance checking is an area of research that aims to
2.1. BIM-GIS integration provide computational support for compliance checking of building
projects against building codes developed by government agencies.
BIM and GIS integration enables effective management of design Building codes consist of zoning information that include rules, stan­
data in various stages of a project’s lifecycle, from planning and design dards, and specifications. They are legal documents for the evaluation of
to construction, operation, and maintenance. While BIM and GIS tech­ building projects. The regulatory knowledge specifies technical and
nologies are closely related, they remain largely isolated. Distinctive functional criteria that buildings and their settlements must meet
throughout buildings’ intended lives [32]. Many countries have national

2
Y. Demir Altıntaş and M.E. Ilal Automation in Construction 128 (2021) 103743

building codes that are valid for all building design and construction asserted that a BIM model is not enough to represent the data required
projects. In Turkey, two documents that regulate construction are for compliance checking against laws and regulations. Hence, they
Standardized Building and Zoning Regulation for Planned Areas, and proposed an approach that included the extension of BIM model to add
Construction Law No.3194. Each municipality develops its own local the missing information specifically for the purpose of construction
regulations within the framework set by these documents. operations and then checking the model’s compliance with regulations.
Automated compliance checking studies for building models have However, the number of studies that study BIM-GIS interoperability in
been continuing for the last three decades. Singapore Building and the area of automated compliance checking is still limited.
Construction Authority’s CORENET project in 1995 was the first auto­ In summary, previous efforts mainly focused on integrating GIS data
mated code compliance checking system used in AEC industry. The into BIM environments or vice versa using data conversion/translation
project aimed to provide a web based electronic system to do code and model extension approaches and had to deal with information loss
checking over submitted building plans in the areas of building control, and mismatch. The current study proposes a loose coupling approach
fire code, environmental health, public housing and vehicle parking where GIS based zoning data and BIM based project data is taken into a
[33]. Norwegian efforts for rule checking aimed to check BIM projects separate application with its own independent domain model that
for evaluating spatial program requirements, and building accessibility combines GIS, BIM, and building code data to facilitate automated
[34]. The Cooperative Research Center for Construction Innovation in zoning code checking.
Australia funded the Design Check project in 2006 which enabled
automated checking of Australia’s requirements for disabled accessi­ 2.3. Current use of GIS in Izmir Municipalities
bility [35]. SMARTcodes project by International Code Council (ICC)
was also initiated in 2006, aimed mapping the paper based building In recent years, digital city models started playing a central role for
codes into computer interpretable code sets and, automate online code- zoning information storage and communication. Yet, manual checking
compliance checking of building projects [36]. General Services of building designs against zoning codes is still the current practice.
Administration (GSA) in United States funded development of a rule Zoning information is generally kept as written texts and scanned plan
checking system in the areas of circulation and security validation of documents. Before a building permit is issued, a skilled person checks
buildings [35]. Korean SEUMTER electronic system enabled the auto­ the building design for code compliance manually following regulations
mation of building code compliance checking against fire prevention and guidelines and this manual process takes time, is very labor inten­
[37]. Malsane et al. [24] and Dimyadi et al. [32] also developed a sive, and is error prone due to its highly repetitive nature. This manual
building regulation-specific object model for automated compliance code checking process is limited - by law - to thirty days. However,
checking against fire safety. Choi, Choi, and Kim [37] developed an delays are common depending on project size, project site and the
automated BIM-based system for checking high-rise and complex workload of the municipality [44]. Furthermore, projects usually
buildings against the evacuation regulation compliance. Yang and Xu require multiple revisions.
[38] studied the development of an online code checking process. These Some of the zoning information is held digitally in isolated GIS en­
studies represent a significant progress in this research area, yet they vironments. Neighborhood names, locations of authorized buildings,
mostly deal with regulations of specific disciplines and focused on locations of landmarks, numbering of buildings, names of roads and
checking individual building designs isolated from their context. They streets are some of the information that are recorded digitally. However,
do not provide a generic solution, which would be applicable to all types there is no standard data model employed by all municipalities and
of clauses. Also building regulations vary between countries further existing digital information is inadequate for code checking. In addition,
limiting the applicability of proposed models. because of dependence on traditional methods of using information
There is an increasing demand for using BIM tools in code checking from printouts and archive documents and steep learning curves of
processes and building permit approval processes [35,39]. Even though, newer systems; city information database systems are currently not
present BIM tools used in these code-checking processes are expected to being used.
hold regulation specific data, they are not capable of containing all the Lack of collaboration between local governments is another reason
required building code data for a fully automated code checking [5]. For for the lack of a common city information database system. Usage of
example, for this purpose, Malsane et al. [5] developed an IFC-compliant both digital and manual systems, and thus, the existence of various
England and Wales building regulation specific object model focusing on different systems that municipalities utilize, cause communication is­
fire safety and extended the building model as current IFC files are not sues. There is no common system and every municipality organizes and
rich enough for use in the automated compliance checking process. systematizes its own data. There is a need for the development and
BIM-GIS interoperability issue in the context of automated code adoption of standards for effective management and sharing of the ever-
compliance checking has been studied in recent years. As an example, growing data on our cities. The proposed integration methodology aims
Olsson et al. [40] proposed integration of BIM and GIS data for auto­ to meet these requirements and manage BIM and GIS data in a unified
mating building permission process and checking building models domain model to be used in an automated code checking process.
against Swedish building permission regulations. Their approach
included importing a BIM model into a geospatial environment as mu­ 3. Neighborhood data analysis and modeling
nicipalities who conduct the checking process are more familiar to
geospatial environments rather than BIM environments. Benner, Geiger In the first phase of methodology development, the building code of
and Häfele [41] studied building information and geo information İzmir is analyzed to define what should be represented explicitly for the
integration for building licensing processes by transforming BIM data in purposes of automated zoning compliance checking. Then, an analysis of
IFC format into a CityGML city model where the rule checking will be the building code document is done to define the missing concepts in
performed. They integrated geo information with building data, and BIM that should be represented explicitly for the purposes of automated
enabled automated check of a building for its consistency against legal zoning compliance checking. In the third phase, a representation for
regulations. In their study, Van Berlo, Dijkmans and Stoter [42] used building code concepts that need to be managed by GIS is designed and a
data import and data transformation approaches to create an integrated UML model is developed. Finally, the representation is implemented
model for checking building permits. Additionally, this model enabled using the open-source geographical information system, QGIS.
architects to use geo information in early design phases. In their
approach, firstly they converted geo information objects as IFC and then 3.1. Decomposition of building codes
they imported these objects into BIM environment where the rule
checking process will be conducted. Salama and El-Gohary [43] also Being a metropolis, İzmir’s housing and zoning code is

3
Y. Demir Altıntaş and M.E. Ilal Automation in Construction 128 (2021) 103743

comprehensive and includes a large set of rules that have high Table 1
complexity. As all the municipalities’ housing and zoning codes consist Terms used in İzmir Building Code that are missing in BIM.
of rules that are similar, the effort for formalization of zoning informa­ Clause Statement No Object/Term Concept
tion in Izmir Municipality Housing and Zoning Code (IMHZCode) is
24 4 Width of Parcels Parcel
generalizable and applicable to a wide a range of zoning code docu­ 24 4 Zoning type Parcel/
ments. Hence, IMHZCode can be adapted to regulations of other cities of Block
Turkey. In this respect, IMHZCode has been chosen for the case study 24 4 Construction Order Parcel/
and is analyzed for comprehending the building codes. IMHZCode is the Block
24 6 Depth of Parcels Parcel
legal document that specifies minimum conditions that need to be 24 6 Order of Parcel with Front Garden Parcel
satisfied by architectural projects within their settlements. It consists of 24/42 6/7 Order of Parcel without Front Parcel
six sections. The clauses that are used for evaluating the compliance of a Garden
building project are found in section III whereas the rest of the building 24 7 Area of Parcels Parcel
25 Organization of Parcels Parcel
code is informative.
26 Integration and Separation of Parcel
IMHZCode includes information related to buildings, building parts Parcels
and the surroundings. In this stage, in order to identify the information 26 15 Face any road Parcel
that is not covered by BIM data, all clauses of IMHZCode in section III 27 1 Front Garden Distance Parcel
that are related to the surroundings are decomposed into a list of terms. 27/28/ Has Existing Building Parcel
37
These terms are then classified according to which concept they are 27/28/ Has Existing Licensed Building Block
pertinent to including parcel, building, block, road, sidewalk and ar­ 37
cade. Table 1 lists all objects and terms mentioned in Izmir Building 27 3–4-5 Front Garden Distance of Adjacent Parcel
Codes with where they are mentioned, and what they are pertinent to. Parcels
27 1 Dimension of Garden Facing Road Parcel
Each of the clauses consists of several rules and statements that define
27 1 Dimension of Garden Facing Green Parcel
conditions related to specific terms such as parcels, blocks, roads etc. Area
27 1 Dimension of Garden Facing Parcel
3.2. Concept mapping Parking Area
27/42 6 Side Garden Distance Parcel
27 9 Rear Garden Distance Parcel
After identifying the classes and their properties in IMHZCode, an 27 13–14-15 Rear Garden Distance of Adjacent Parcel
analysis is made for determining which of these code concepts and Parcels
properties are lacking in BIM. These missing entities that are the focus of 27 10 Number of Roads Facing Parcel Parcel
this study needs to be modelled and managed along with BIM data to be 27 10 Corner Parcel Parcel
27/36 11 / 3 Distance Between Two Buildings in Parcel
used by the architect during the early design phase for code compliant
Parcel
design and later for automated code compliance checking. For the 28 1–2–3-4-5-6-7-8- Depth of Building Building
analysis, IFC is selected as the BIM representation as it is the only 9-10-11
commonly used data schema aiming to define a common language to 28 3–4–5-6 Building Depth of Adjacent Parcels Building
enhance the collaboration, both for AEC/FM domain, and cross domain 28 2 Distance to Rear Neighbor Border Parcel
29 1 Maximum Building Façade Length Building
and cross-disciplines. The analysis determined whether each a domain 30/31/ 1–4 / 3–6 / 4 Maximum Height of Buildings Building
model concept maps to an entity, a property from a property set (Pset) or 35
a quantity from a quantity set (Qto) in IFC. For example, zoningType 30/31 1–4 / 3 Number of Storeys Building
property of block class has a correspondence in IFC as Pset_BuildingUse; 31 3–6 Allowable Construction Area Parcel
32 Dead End Streets and Closed Roads Road
while constructionOrder property do not have a correspondence in IFC
33 Disaster Area Block
[45]. 34 Non-Settlement Area Block
The analysis results showed that %70 of the data (48 of 69 terms) on 34 3 Building Façade Length Coincident Building
the surroundings that was identified with IMHZ code analysis, are to Road
lacking representation in BIM, but can be modelled using GIS. As it is 34 6–7 Floor Area Ratio Parcel
34/37 6/4 Distance of Building to Road (Front Parcel
impractical to expect BIM to integrate geographical information in its Garden Distance)
domain, all the identified code data on the surroundings regardless of 34 6 Distance of Building to Parcel Parcel
whether the data has a correspondence in BIM or not is modelled in GIS. Border
35 Cadastral Parcels Parcel
36 Multiple Construction Permits in a Parcel
3.3. Constructing GIS domain model
Parcel
37 1 Grade Elevation of Road Facing the Road
In this step, terms that were identified in IMHZCode are brought Parcel
together in an object hierarchy. Classes are identified by analyzing if 37 Grade Elevation Parcel
they contain members having certain attributes; and properties are 37/38/ 1 / 6–7 /3 Sidewalk Top Level Sidewalk
40
identified by analyzing if they contain data related to a specific element. 37 7 Given Building Height for Road Road
Properties are shared by all objects of a class. For example, “construction 37 11 Road Level Road
order” is a property of the “parcel” class mentioned in IMHZCode. There 37 18 Front Garden Level Parcel
are also properties such as area, width, distance etc., which are mostly 37 18 Slope Road
37 19 Rear Garden Level Parcel
used to define requirements. This study focuses on entities that are
37 20 Side Garden Level Parcel
missing in BIM models, namely the classes related to zoning, including 40 3 Sidewalk Width Sidewalk
block, road, sidewalk, parcel, building and arcade. Entities of the 42 2 Base Area Building
building and building parts are already considered and modelled by BIM 42 2–8–11-12 Width of Cantilevers Building
models and are not the major focus of the study. This partial domain 42 4 Adjacent Parcel’s Cantilever Side Building
Offset
model that is constituted by classes and relationships between them is 42 6–9 Adjacent Parcel’s Cantilever Width Building
described using Unified Modeling Language (UML) class diagrams in 42 7 Road Width Road
Fig. 1. (continued on next page)
In the zoning domain model, “surrounded by” association is used for

4
Y. Demir Altıntaş and M.E. Ilal Automation in Construction 128 (2021) 103743

Table 1 (continued ) vehicles and sidewalks for pedestrians. A sidewalk (pavement) is a

Clause Statement No Object/Term Concept paved and elevated path by the side of a roadway for the use of pedes­
trians. Sidewalks are part of the road and sidewalk width is included in
42 8 Height of Bottom of Cantilever from Building
Sidewalk Top Level
the road width. In QGIS, roads and sidewalks are linear features. Roads
55 2–4 Evened Earth Level Parcel are modelled using lines and nodes going through road centerlines as
55 4 Top Level of Ground Floor Slab Building shown in Fig. 4(a). Sidewalks are vector-based lines automatically
63 1 Height of Arcades Arcade derived from polygon layer of block as shown in Fig. 4(b). Each edge of
63 1 Depth of Arcades Arcade
sidewalk along the block can be associated with the parcel it is adjacent
64 1–2-6 Required Area for Locating Trees Parcel
64 4–5 Number of Existing Trees Parcel to. When needed, thickness can be given from the block boundary to­
wards the road. However, representing road with centerline and side­
walk with the edge facing parcel are adequate for automated code
block-road/sidewalk relationships as roads surround blocks. “Contain” checking based on IMHZCode regulations. Building code requires in­
association is used for block-parcel, parcel-building, block-arcade and formation only on how roads and sidewalks are related to surrounding
road-sidewalk relationships. A block contains one or more parcels and features, area information and geometry is not necessary for roads and
zero or more arcades. A parcel may contain a number of buildings, and a sidewalks. Arcades are not common and the case study does not include
road may contain more than one sidewalk. By removing the sidewalk one but arcades should be modelled using polygons.
width from the road width, the value of the vehicle road can be ob­ As a next step, an attribute table is organized for each layer including
tained. “Has” association is used for defining the boundaries of parcel the attributes defined in the domain model. The attribute tables contain
and building geometries as İzmir Building Codes consider lengths, re­ semantic data on geographical features. Each feature is represented as a
lations and distances between parcel and building objects. Generaliza­ different record in the tables.
tions are “is” relationships in the UML model providing specializations
for the parcel and boundary line classes. Parcel parent class has two 3.4.2. Spatial queries and topological relationships
child classes: “zoning parcel” and “non-standard parcel.” “Non-standard Building codes have rules, which query spatial relations between
parcel” is defined as a specialized parcel class for representing areas such building, parcel, block and road. For example, consider the rule:
as green-area, parking area, disaster area, market-place, recreation area
and gas station area. Construction in these areas are generally not “Generally, buildings’ maximum elevations are determined based on the
allowed and rather than İzmir Building Code document, specialized road’s grade elevation that the parcel, where the building is located,
documents are valid for the organization of these kind of areas. faces.” (Izmir Municipality Housing and Zoning Code; Clause 37).
Boundary line is the other parent class from which “parcel boundary” The parcel id where building is located and the road’s id that the
and “building boundary” classes are derived. parcel faces should be known for determining the grade elevation.
Hence, three basic spatial relations including containment, adjacency
3.4. Case study: implementation in QGIS and intersection are structured between zoning objects in IMHZCode as
shown in Fig. 5. Containment relationship is structured for block-parcel,
In order to provide a proof-of-concept implementation for the new parcel-building, road-sidewalk and adjacency relationship is structured
zoning domain model required by automated code checking, an appli­ for block-road and parcel-sidewalk. In some queries, intersection rela­
cation is developed using the open source Quantum GIS (QGIS) soft­ tionship is also used as in the parcel-block relation, the shared geometry
ware. Çamiçi neighborhood of the Bornova district in Izmir was selected is the entire geometry of the parcel object. The queries are constructed
for the case study. Its 2D map was taken from the municipality. The CAD using SQL-like expression syntax of QGIS.
file was imported into QGIS environment and transformed into shape­ During the code checking process, the design information is sub­
files. Fig. 2 shows the modelled vector data structure of the mitted as a BIM file that includes the block and parcel number of the
neighborhood. building, the neighborhood name and the building id. The GIS should
First, for each class in the domain model a layer has been defined in complement this data with information only on the relevant neighbor­
QGIS. Features (instances of a class) are drawn using vector-based tools hood and not the whole city. This is be done by filtering the GIS data
on the appropriate layer and their associated semantic data is entered based on the parcel id. The block containing the parcel as well as a
into the attribute table of the layer. Then, spatial analysis queries have number of neighboring blocks and surrounding roads are filtered and
been developed for automated creation of topological relationships be­ can be saved to a GML file for use during code checking.
tween features.
4. Developing the application
3.4.1. Modeling classes and properties
Each class including block, road, parcel, sidewalk, building and ar­ For the test case of the study, urban data from QGIS should be
cade is handled as individual layers. The geographical features (objects) retrieved, the BIM data should be retrieved and the checking should be
belonging to each layer are modelled using vector data tools. conducted based on the computationally modelled Izmir Municipality
A block is a piece of land, which is already built-up or open to new Housing and Zoning Code regulations and the retrieved GIS data. Thus,
developments, surrounded by roads. The block, and the parcels located before developing a proof of concept prototype, Izmir Municipality
in each block, are assigned numbers and letters as identifiers. A parcel Housing and Zoning Code regulations and a BIM project are modelled in
(or lot) refers to a building site in this study. It is a subdivided property this research.
that can be purchased and developed by the property owner within
boundaries defined by local regulations. Each building located in a
parcel has both a parcel Id and a block Id for identifying its location. For 4.1. Building data modeling
modeling parcel, block, and buildings in GIS, polygons are used as
shown in Fig. 3. Parcel and building feature boundaries are explicitly For the study, a BIM project is modelled to conduct BIM-GIS data
represented with boundary line classes. İzmir Building Codes query coupling and create the dataset required for a compliance checking
adjacency relationships and differentiate between front, side, and rear process. As automated code checking can work if the building elements
borders of buildings and parcels. Each boundary line can be associated and their relations are modelled in an intelligent format, creating a
with the feature it is adjacent to and have its own setback distance. building model in BIM application is fundamental. A building object
Road is a route for accessibility and travel including roadways for with its surroundings is modelled using Graphisoft Archicad v21 as

5
Y. Demir Altıntaş and M.E. Ilal Automation in Construction 128 (2021) 103743

Fig. 1. Zoning domain model classes missing in BIM.

shown in Fig. 6. As explained in the study, a BIM model does not include and not modelled in detail as compliance checking of a detailed building
information related to the surroundings for a complete code compliance model was conducted and studied in the study of Macit [46] already. In
checking. Hence, only a parcel object could be modelled as the data in addition, modeling the building’s borders is adequate for the study that
the surroundings and then exported together with the building object studies the queries that relate building to its surrounding.
using IFC data model.
Parcel is modelled using the “mesh” tool in Archicad and recorded as
ifcSite entity in the IFC data model. Building is modelled using “zone” 4.2. Building code modeling
tool of Archicad and recorded as ifcSpace entity in the IFC data model.
For demonstrative purposes, the building is modelled as a simple object 4.2.1. Constructing rule database
Automated compliance checking systems require rules in

6
Y. Demir Altıntaş and M.E. Ilal Automation in Construction 128 (2021) 103743

models are some of them. However, it is hard to model comprehensive

building code models with decision tables. On the other hand, rule-
based models include hard-coding that makes it difficult to update and
revise the model due to the building code amendments (change of
setback distances or increase/decrease of building height limits). In
addition, the codes are modelled by system programmers and hard to
comprehend by the building code authors, which can cause mis­
understandings between system programmers and the building code
authors. Logic-based models and object-oriented models require code
modelers to have specialized programming knowledge and expertise in
case of modeling the rules [46].
Maintainability is a crucial criteria in case of modeling building
codes because the rules are updated and changed continuously. Another
important criteria is that building code modeling should not require any
programming knowledge and expertise. The rules should be adaptable
with no further expertise in programming. Semantic modeling is the
approach that meets all the criteria for efficient modeling of building
codes. It is easy to understand, build and maintain building code rep­
resentations for non-programmers using semantic modeling. By using a
semantic approach, an AEC domain professional can identify and
implement computable rules for model checking [47].
In this research, semantic modeling approach is used for developing
Fig. 2. Çamiçi neighborhood of Bornova.
a representation for the Izmir Municipality Housing and Zoning Code
regulations. Rules are modelled using the four level representation
paradigm [48], and modelled independent of the checking engine and
stored in Microsoft’s relational database application Access. Whenever
IMHZCode is updated, the rules in the database can be modified inde­
pendently without requiring any changes to the checking engine.
The rules are constructed using the concepts and properties in the
unified domain model. Thus, while some of the concepts belong to BIM
data, some of them belong to GIS data. Then, relationship hierarchy of
rules are modelled with rule-sets. The rules are connected to each other
with logical conjunctions such as OR and AND. With OR conjunction,
only one rule should be satisfied. AND conjunction is used for combining
the rules indicating different conditions for the same property. With
AND conjunction, all rules that indicate different requirements should
be satisfied by the property of an object. At the end of this step, rule-sets
are grouped according to the concepts they are related to and this
grouping enabled to access all the rules that apply to a determined
Fig. 3. (a) Block layer modelled using polygons (b) Parcel layer modelled
concept (Fig. 7).
using polygons.

4.2.2. Analyzing rule structure

IMHZCode includes various type of rule statements including value
checking rules, attribute checking rules, applicability clarifying rules,
descriptive rules and qualitative rules. Analyzing the structure of the
rules and identifying the type of each statement is needed for deter­
mining which of the rules can be and should be modelled with the four
levelled methodology.
Value checking rule statements check correctness based on value and
implies strict constraints. They indicate how things should be. The rules
related to the distance, height, storey number, length are examples of
this type. Such as the following:

“IMHZCode, Clause 63: Height of arcades are 3.50meters.”

Attribute checking rule statements check correctness based on an

attribute. The attribute can both be a value or an enumeration. Such as
the following:
Fig. 4. (a) Road layer modelled using lines and points (b) Sidewalk layer
modelled using lines. “IMHZCode, Clause 28: In semi-detached construction order, build­
ing depth is the depth of semi-detached building.”
computable format for checking of projects that are also in digital Third type of rule statements clarify applicability conditions and
format. There are various methodologies in literature that are being used identifies what to do based on specified conditions. Such as the
for rule representation and modeling in computable format for the following:
purpose of automated compliance checking. Decision tables, rule-based
models, logic-based models, object-oriented models and semantic “IMHZCode, Clause 24: Parcel areas, where number of building
storeys are ten or more, cannot be less than 2000 m2.”

7
Y. Demir Altıntaş and M.E. Ilal Automation in Construction 128 (2021) 103743

Fig. 5. Containment, adjacency and intersection relations.

Fifth type of rule statements include qualitative evaluations, which

are open to interpretation and require human judgement. They are
mostly under the initiatives of local authority. Such as the following:

“IMHZCode, Clause 37: In exceptional situations, where parcel has a

high slope from road or there are different elevations in the same
parcel, technical committee’s decisions based on their surveys in the
area are valid for defining the elevations in site.”

Fourth and fifth type of rule statements are informative clauses.

These types of rules are subjective rules and direct meaning cannot al­
ways be taken from them. They require human judgement for under­
standing the meaning. Thus, they are not checkable and cannot be
transferred into computer processable format easily. These statements
are outside the scope of the study and are not modelled since they cannot
be integrated into such a system. Since only value checking rules,
attribute checking rules, applicability clarifying rules can be represented
in digital format, they are identified and extracted from the Izmir Mu­
nicipality Housing and Zoning Code.
There are 43 clauses in section III of IMHZCode document. This is the
section which contains the rules for evaluating the compliance of a
building project. When the clauses are decomposed into a list of rules,
443 rules are identified. After the wording of the rules are analyzed, 288
rules are determined to be formalizable. 155 rules necessitate human
judgement. Hence, 65% of the IMHZCode rules can be represented using
Fig. 6. BIM model. the semantic modeling approach. Compliance checking for these rules
can be automated without requiring any user intervention.
The first three type of statements are declarative statements, which
declare information unambiguously. They are measurable and clear in
their meaning. These kind of statements can be transferred into com­ 4.3. The prototype
puter processable format easily [5]. Fourth type of rule statements are
descriptive rules. They give general information and do not indicate The prototype system has been developed and implemented using
acceptable and threshold values, clarify calculation methods or identify the Java language. It consists of four components: Building Model
obligations. They just identify probabilistic situations and are mostly Parser, Geographical Information Parser, Building Code Parser and
related to aesthetic values such as the following examples; Building Code Checker as shown in Fig. 8.
Building Model Parser accesses and reads the IFC data. In this study,
“IMHZCode, Clause 40: Entrance canopies can be compatible with BIM data exported as IFC2x3 is parsed using the JSDAI library and IFC
the character of the building and the streetscape.” objects are created in Java environment.

Fig. 7. Rule-set group objects constituting rules.

8
Y. Demir Altıntaş and M.E. Ilal Automation in Construction 128 (2021) 103743

Geographical Information Parser reads the GML data exported from During preprocessing, the prototype calculates setback distances using
GIS environment, for accessing and extracting GIS data, the Document coordinates and dimensions of building and parcel objects and assigns
Object Model (DOM) Parser is used. the calculated values to domain objects’ transient properties.
Building Code Parser reads the building code data from the MS Ac­ The unified domain model contains all the classes and properties
cess database and instantiates rule objects for the checking process. from both BIM and GIS domains as shown in Fig. 12. The automated
Building Code Checker checks the building data extracted from BIM code checking application then has access to process this unified dataset
against building code data, using the data on the surroundings extracted - the coupled domain model.
from GIS. Building Code Checker ensures that the design complies with
the regulations. At the end of the process, it presents a report. The 4.5. Testing of prototype
checker firstly accesses the rule-set group objects and makes use of the
rule-sets grouped according to the class they are related to. For each The prototype is tested in a use-case scenario demonstrating how
class, it applies all rules in the related rule-set group. automated compliance checking processes are carried out with various
levels of complexity. The use case scenario is specifically designed for
evaluation and validation of the BIM-GIS unified model. The use-case
4.4. Coupling of BIM and GIS data
scenario looks for correctness in checking results. Correctness is tested
by varying the conditions within the context gradually. The goal is to
The main purpose of this research is to use the building information
eliminate false negative and false positive results and ensure that the
provided by IFC, integrate it with required data on the surroundings
unified domain objects are able to represent the missing data on the
from GML and to create a dataset that will be used in an automated code
surroundings correctly. The correctness of the coupling has been eval­
checking process. Thus, data coupling is needed between BIM and GIS
uated by using the prototype to carry out testing and validation. Results
datasets to create a unified model. For creating a unified model, BIM and
indicate correct integration of BIM and GIS data.
GIS data models are analyzed and compared in detail, and overlapping
Considering the data on the surroundings during architectural design
parts are coupled. The new integrated semantic model merges BIM and
process is crucial as it effects both the checking results and the checking
GIS data into one unique model.
process itself. In some cases, there can be existing buildings within the
For data coupling, firstly BIM and GIS data models are analyzed to
block where the building site is located. In those cases, exception rules
reveal the classes that are similar in semantics and contain similar
are valid and the prototype identifies the existing buildings within the
properties. This has been done using UML diagrams of BIM and GIS data
block for getting the required information for the code checking process.
models where the classes and properties are matched. The classes that
For demonstration, clause 27 (setbacks) have been selected. It is one of
exist in both UML diagrams and reference the same object have been
the most complex clauses in İzmir Building Code in addition to its
considered “coupling elements.” For semantic interoperability, the
requirement of data on the surroundings from GIS.
common concepts having the same meaning from BIM and GIS are
Clause 27 requires allowed front, side and rear setback information
coupled by using the common properties, which are IDs. As the site/
for the parcel from the GIS file. If there are existing buildings within the
parcel object is at the overlap of BIM and GIS, parcel objects from BIM
block, the prototype also identifies the setback distances of the adjacent
and GIS are matched/mapped using their IDs for unique identification as
parcels within the same block to be used during the checking process as
shown in Fig. 9. For the GML file, ID of parcel object is used for unique
references. Ruleset RS.27 includes three sections that indicate the con­
identification; for the IFC file, the name of the IfcSite element is used.
ditions for determining setback distances for a parcel. First section is for
At the end of the coupling process, parcel object in BIM and parcel
front setback distance and it indicates that front setback distances
object in GIS become a single parcel object that holds all the information
should be at least 5.00 m. If there is an existing building in any of the
recorded separately in IFC and GML data models as shown in Fig. 10.
parcels within the same block, setback distances should be the same as
Secondly, geometries are coupled. Geometric reasoning is used for
setback distances used in that parcel. Second section is for side setback
making sure that the two parcels are identical geometrically. A Java
distance and it indicates that side setback distances should be 3.00 m.
algorithm is developed that compares the two parcel geometries
Third section is for rear setback distance and it indicates that rear
extracted from BIM and GIS to identify if the corresponding objects have
setback distances should be half of the building height. If there is an
the same geometry and location with each other.
existing building in any of the parcels within the same block, setback
After the coupling process, the building in IFC is parsed and added to
distances should be the same as setback distances used in that parcel.
the buildings list in the Java application where all the other existing
Thus, 27th clause requires setback distances defined for that parcel and
buildings on the surrounding parcels are also stored as shown in Fig. 11.
existing setback distances applied for the surrounding parcels from GIS
Geometric reasoning is also used for populating properties that do
for the checking process.
not exist in BIM but are needed during checking. These properties are
For the case, the building is located in a parcel whose construction
calculated based on dimensions, locations, and relationships of existing
order is attached. The architect planned the front setback distance as
physical elements. For example, setback distance property does not exist
3.00 m since there are existing buildings in the surrounding parcels
in BIM, but is required by Izmir Municipality Housing and Zoning Code.
having front setback distances of 3.00 m and the new building is subject
to the existing established building line (Fig. 13(a)). Thus based on the
rule, “if there is an existing building in any of the parcels within the same
block, setback distances should be the same as setback distances used in
that parcel,” the project passes RS.27 with the defined dataset.
However, if there are no existing buildings in the surrounding parcels
as shown in Fig. 13(b), the condition defined for setbacks changes.
Ruleset RS.27 indicates that if there are no existing buildings in the
block, front setback distances should be at least 5.00 m. Thus, the 27th
clause fails in second condition since the design does not meet the
setback distance requirements.
Ruleset 28 (depth of buildings) is also considered for the scenario as
building depth is related to the allowable setbacks defined for that
parcel. As the building depth is defined as 22.00 m for the scenario, the
Fig. 8. The structure of the prototype. dataset passes the check on Ruleset 28 that indicates building depths

9
Y. Demir Altıntaş and M.E. Ilal Automation in Construction 128 (2021) 103743

Fig. 9. Coupling BIM and GIS data.

should not exceed 22.00 m. The checking results of the scenario is given checking of building projects, resulting in a list of zoning concepts
in Fig. 14. mentioned in building regulations. Then a mapping was defined be­
tween the IFC model entities and the identified concepts. The analysis
5. Conclusion results showed that %70 of the neighborhood data (48 of 69 terms) that
was identified with IMHZ code analysis, is not represented in the IFC
The research presented in this paper was aimed at developing a data model. A follow-up analysis showed that they can be modelled
methodology for the loose coupling of BIM and GIS data in an automated using GIS. Thereafter, the crux of the study focused on developing a
zoning code checking context. The integration of BIM and GIS data is domain model for the representation of the identified missing zoning
necessary for facilitating automated compliance checking against zoning data required for automated code checking processes. Here, first, an
codes, since BIM model does not contain the required neighborhood object-oriented model for the zoning concepts was developed that could
data. This study firstly identified the neighborhood data referenced in be implemented using GIS constructs. Then, this domain model was
building regulations and required for automated code compliance implemented in QGIS. The classes and attributes became layers and

10
Y. Demir Altıntaş and M.E. Ilal Automation in Construction 128 (2021) 103743

Fig. 10. Coupling of BIM and GIS parcel objects.

combined with BIM data using a semantic mapping approach in a uni­

fied environment independent of any BIM-GIS platform or compliance
checking system. The prototype populates the domain model by only
extracting what is necessary from both the IFC data from BIM systems,
and the GML data from the QGIS application. The rules that also exist
independently, are then used to check the project’s compliance with
IMHZCode.
The loose coupling of BIM and GIS data in an independent third
platform is a break from past efforts that relied on data translation or
model extension approaches. Processing BIM and GIS in the same
environment independently avoids data translations between the
models, incorrect mapping and thus, loss of information. The indepen­
dent nature of the coupling environment also does not increase data sets
Fig. 11. Integration of ifcBuilding with GML dataset.
and workloads of either BIM or GIS environments. After all, it is
impractical to expect BIM to explicitly define geographical information;
properties, while topological relationships were structured using spatial
and GIS to manage detailed building data. The integrated environment
queries. It was shown that GIS is able to adequately represent zoning
developed in this study can manage data coming from BIM and GIS
concepts relating to the surroundings of a building project within the
worlds in the same structure and with the same tools. The study focuses
scope of IMHZCode. Finally, a software prototype was developed that
on the integration of concepts defined in IFC and GML standards within
conducts automated code compliance checking. The GIS data was
a new domain-specific model. The domain is automated zoning code

11
Y. Demir Altıntaş and M.E. Ilal Automation in Construction 128 (2021) 103743

Fig. 12. Unified zoning domain model.

12
Y. Demir Altıntaş and M.E. Ilal Automation in Construction 128 (2021) 103743

Fig. 13. (a) First condition where there are existing buildings on surroundings (b) Second condition where there are no existing buildings on surroundings.

Fig. 14. The checking results for the scenario (a) There are existing buildings in block (b) There are no existing buildings in block.

13
Y. Demir Altıntaş and M.E. Ilal Automation in Construction 128 (2021) 103743

compliance checking. [13] R. Sebastian, M. Böhms, P. Van Den Helm, BIM and GIS for low-disturbance
construction, in: Proc. 13th Int. Conf. Constr. Appl. Virtual Real., London, UK,
Modeling zoning data in GIS, integrating it with BIM data, and
2013, pp. 469–479. http://itc.scix.net/data/works/att/convr-2013-47.pdf.
combining it with building codes in digital format, brings many ad­ [14] I.C. Wu, S.H. Hsieh, Transformation from IFC data model to GML data model:
vantages in the area of code checking processes in municipalities. Digital methodology and tool development, J. Chin. Inst. Eng. 30 (2007) 1085–1090,
checking of projects, even with partial automation, simplifies the work https://doi.org/10.1080/02533839.2007.9671335.
[15] I. Hijazi, M. Ehlers, S. Zlatanova, U. Isikdag, IFC to CityGML transformation
of approving authorities, leads to faster managing and processing. The framework for geo-analysis: a water utility network case, in: 4th Int. Work. 3D Geo-
ability to superimpose GIS data and BIM data in an application provides Information, Ghent, Belgium, 2009, pp. 123–127, https://doi.org/978-90-
automated visual control, faster turnaround in feedback, faster ap­ 9024820-2.
[16] E. Elbeltagi, M. Dawood, Integrated visualized time control system for repetitive
provals for buildings; while preventing errors due to manual checking construction projects, Autom. Constr. 20 (2011) 940–953, https://doi.org/
and inconsistencies in the interpretation of codes. It is foreseen that the 10.1016/j.autcon.2011.03.012.
developed checking system can be integrated as a plug-in into a BIM [17] J. Irizarry, E.P. Karan, Optimizing location of tower cranes on construction sites
through Gis and Bim integration, J. Inf. Technol. Constr. 17 (2012) 351–366.
platform in future studies and will be able to provide architects imme­ [18] J. Irizarry, E.P. Karan, F. Jalaei, Integrating BIM and GIS to improve the visual
diate feedback with regard to code compliance, even in early design monitoring of construction supply chain management, Autom. Constr. 31 (2013)
phases. 241–254, https://doi.org/10.1016/j.autcon.2012.12.005.
[19] J. Shi, P. Liu, An agent-based evacuation model to support fire safety design based
Other challenging directions for future research include: 1) Applying on an integrated 3D GIS and BIM platform, in: Comput. Civ. Build. Eng, 2014,
the loose coupling methodology developed in the research to other fields pp. 1893–1900, https://doi.org/10.1061/9780784413616.053.
that require management of BIM and GIS data together. (E.g., energy [20] T. Park, T. Kang, Y. Lee, K. Seo, Project cost estimation of national road in
preliminary feasibility stage using BIM/GIS platform, in: Sixth Annu. Int. Conf.
monitoring and planning applications in future smart cities where GIS
Comput. Civ. Build. Eng, 2014, pp. 423–430, https://doi.org/10.1061/
and BIM data need to be combined with neighborhood scale energy 9780784413616.053.
simulations), 2) Extending the unified model to enable reasoning in 3D, [21] A. Borrmann, T.H. Kolbe, A. Donaubauer, H. Steuer, J.R. Jubierre, M. Flurl, Multi-
using data from 3D BIM and 3D geospatial environments, and 3) scale geometric-semantic modeling of shield tunnels for GIS and BIM applications,
Comput. Civ. Infrastruct. Eng. 30 (2015) 263–281, https://doi.org/10.1111/
Investigating if concept mapping can be an automated process. mice.12090.
[22] M. Zhiliang, W. Zhenhua, S. Wu, L. Zhe, Application and extension of the IFC
Funding standard in construction cost estimating for tendering in China, Autom. Constr. 20
(2011) 196–204, https://doi.org/10.1016/j.autcon.2010.09.017.
[23] S.-H. Lee, B.-G. Kim, IFC extension for road structures and digital modeling,
This research did not receive any specific grant from funding Procedia Eng. 14 (2011) 1037–1042, https://doi.org/10.1016/j.
agencies in the public, commercial, or not-for-profit sectors. proeng.2011.07.130.
[24] S. Malsane, J. Matthews, S. Lockley, P.E.D. Love, D. Greenwood, Development of
an object model for automated compliance checking, Autom. Constr. 49 (2015)
Declaration of Competing Interest 51–58, https://doi.org/10.1016/j.autcon.2014.10.004.
[25] R. de Laat, L. van Berlo, Integration of BIM and GIS: the development of the
CityGML GeoBIM extension, in: C. Kolbe, T.H. König, G. Nagel (Eds.), Adv. 3D Geo-
The authors declare that they have no known competing financial Information Sci, Springer, Berlin, Heidelberg, 2011, pp. 211–225, https://doi.org/
interests or personal relationships that could have appeared to influence 10.1007/978-3-642-12670-3_13.
[26] M. El-Mekawy, A. Östman, I. Hijazi, A unified building model for 3D urban GIS,
the work reported in this paper. ISPRS Int. J. Geo-Information 1 (2012) 120–145, https://doi.org/10.3390/
ijgi1020120.
References [27] T. Gilbert, S. Barr, P. James, J. Morley, Q. Ji, Software systems approach to multi-
scale GIS-BIM utility infrastructure network integration and resource flow
simulation, ISPRS Int. J. Geo-Information 7 (2018) 310, https://doi.org/10.3390/
[1] C. Mignard, C. Nicolle, Merging BIM and GIS using ontologies application to urban
ijgi7080310.
facility management in ACTIVe3D, Comput. Ind. 65 (2014) 1276–1290, https://
[28] A.H. Hor, A. Jadidi, G. Sohn, BIM-GIS integrated geospatial information model
doi.org/10.1016/j.compind.2014.07.008.
using semantic web and RDF graphs, in: ISPRS Ann. Photogramm. Remote Sens.
[2] A. Rafiee, E. Dias, S. Fruijtier, H. Scholten, From BIM to geo-analysis: view
Spat. Inf. Sci., Prague, Czech Republic, 2016, pp. 73–79, https://doi.org/10.5194/
coverage and shadow analysis by BIM/GIS integration, Procedia Environ. Sci. 22
isprs-annals-III-4-73-2016.
(2014) 397–402, https://doi.org/10.1016/j.proenv.2014.11.037.
[29] H. Kim, Z. Chen, C.S. Cho, H. Moon, K. Ju, W. Choi, Integration of BIM and GIS:
[3] U. Isikdag, J. Underwood, G. Aouad, An investigation into the applicability of
highway cut and fill earthwork balancing, Comput. Civ. Eng. 2015 (2015)
building information models in geospatial environment in support of site selection
468–474, https://doi.org/10.1061/9780784479247.058.
and fire response management processes, Adv. Eng. Inform. 22 (2008) 504–519,
[30] L. Knoth, J. Scholz, J. Strobl, M. Mittlböck, B. Vockner, C. Atzl, A. Rajabifard,
https://doi.org/10.1016/j.aei.2008.06.001.
B. Atazadeh, Cross-domain building models—a step towards interoperability,
[4] E.P. Karan, J. Irizarry, Extending BIM interoperability to preconstruction
ISPRS Int. J. Geo-Information 7 (2018) 363, https://doi.org/10.3390/ijgi7090363.
operations using geospatial analyses and semantic web services, Autom. Constr. 53
[31] Y. Song, X. Wang, Y. Tan, P. Wu, M. Sutrisna, J.C.P. Cheng, K. Hampson, Trends
(2015) 1–12, https://doi.org/10.1016/j.autcon.2015.02.012.
and opportunities of BIM-GIS integration in the architecture, engineering and
[5] S. Malsane, The Application of Automated Rule Checking to Existing Uk Building
construction industry: a review from a spatio-temporal statistical perspective,
Regulations Using BIM Technologies, University of Northumbria at Newcastle,
ISPRS Int. J. Geo-Information 6 (2017) 1–32, https://doi.org/10.3390/
2015.
ijgi6120397.
[6] G. Groger, T.H. Kolbe, C. Nagel, K.-H. Haefele, Open Geospatial Consortium OGC
[32] J. Dimyadi, C. Clifton, M. Spearpoint, R. Amor, Regulatory knowledge encoding
City Geography Markup Language (CityGML) Encoding Standard, 2012.
guidelines for automated compliance audit of building engineering design,
[7] S. Amirebrahimi, A. Rajabifard, P. Mendis, T. Ngo, A data model for integrating GIS
Comput. Civ. Build. Eng. (2014) 536–543, https://doi.org/10.1061/
and BIM for assessment and 3D visualisation of flood damage to building, in:
9780784413616.053.
B. Veenendaal, A. Kealy (Eds.), CEUR Workshop Proc., Brisbane, Australia, 2015,
[33] T. Liebich, J. Wix, J. Forester, Z. Qi, Speeding-up the building plan approval-the
pp. 78–89.
Singapore e-plan checking project offers automatic plan checking based on IFC, in:
[8] S. Donkers, H. Ledoux, J. Zhao, J. Stoter, Automatic conversion of IFC datasets to
Eur. Conf. Prod. Process Model. 2002 - EWork Ebus. Archit. Eng. Constr., Portoroz,
geometrically and semantically correct CityGML LOD3 buildings, Trans. GIS 20
Slovenia, 2002, pp. 467–471.
(2016) 547–569, https://doi.org/10.1111/tgis.12162.
[34] K. Lindberg, MATL, Development of a new ICT-system for registration and
[9] M. El-Mekawy, A. Östman, I. Hijazi, An evaluation of IFC-CityGML unidirectional
assessment of accessibility to public buildings, in: Prop. Manag. Statsbygg, BAS
conversion, Int. J. Adv. Comput. Sci. Appl. 3 (2012) 159–171, https://doi.org/
Conf, 2006.
10.14569/ijacsa.2012.030525.
[35] C. Eastman, J. Min Lee, Y. Suk Jeong, J. Kook Lee, Automatic rule-based checking
[10] U. Isikdag, S. Zlatanova, Towards defining a framework for automatic generation
of building designs, Autom. Constr. 18 (2009) 1011–1033. doi:https://doi.org/10
of buildings in CityGML using Building Information Models, in: 3D Geo-
.1016/j.autcon.2009.07.002.
Information Sci, Springer Verlag, Berlin, Germany, 2009, pp. 79–96, https://doi.
[36] J. Wix, D. Conover, Capturing and using knowledge with building information
org/10.1007/978-3-540-87395-2.
modelling (keynote), Inf. Knowl. Manag. - Help. Pract. Plan. Build. Proc. CIB W102
[11] T. Kang, Development of a conceptual mapping standard to link building and
3rd Int. Conf., Stuttgart (Germany) (2007) 35–48.
geospatial information, ISPRS Int. J. Geo-Information 7 (2018) 162, https://doi.
[37] J. Choi, J. Choi, I. Kim, Development of BIM-based evacuation regulation checking
org/10.3390/ijgi7050162.
system for high-rise and complex buildings, Autom. Constr. 46 (2014) 38–49,
[12] K.A. Ohori, F. Biljecki, A. Diakité, T. Krijnen, H. Ledoux, J. Stoter, Towards an
https://doi.org/10.1016/j.autcon.2013.12.005.
integration of GIS and BIM data: what are the geometric and topological issues?, in:
ISPRS Ann. Photogramm. Remote Sens. Spat. Inf. Sci., Melbourne, Australia, 2017,
pp. 1–8, https://doi.org/10.5194/isprs-annals-IV-4-W5-1-2017.

14
Y. Demir Altıntaş and M.E. Ilal Automation in Construction 128 (2021) 103743

[38] Q.Z. Yang, X. Xu, Design knowledge modeling and software implementation for [43] D.M. Salama, N.M. El-Gohary, Semantic modeling for automated compliance
building code compliance checking, Build. Environ. 39 (2004) 689–698, https:// checking, Comput. Civ. Eng. (2011) 641–648, https://doi.org/10.1061/41182
doi.org/10.1016/j.buildenv.2003.12.004. (416)94.
[39] E. Hjelseth, Public BIM-based model checking solutions: lessons learned from [44] A. Demirciefe, Delays in Issuing of Building Permit and Occupancy Permit: An
Singapore and Norway, Build. Inf. Model. Des. Constr. Oper. 1 (2015) 421–436, Analysis of Causes and Durations, Izmir Institute of Technology, 2009. http://ope
https://doi.org/10.2495/bim150351. naccess.iyte.edu.tr/xmlui/handle/11147/3800.
[40] P.-O. Olsson, J. Axelsson, M. Hooper, L. Harrie, Automation of building permission [45] buildingSMART, b, (2016). http://www.buildingsmart-tech.org/ifc/IFC4/Add1/ht
by integration of BIM and geospatial data, ISPRS Int. J. Geo-Information 7 (2018) ml/ (accessed April 5, 2016).
307, https://doi.org/10.3390/ijgi7080307. [46] S. Macit, Computer Representation of Building Codes for Automated Compliance
[41] J. Benner, A. Geiger, K.-H. Häfele, Concept for building licensing based on Checking, Izmir Institute of Technology, 2014, https://doi.org/10.1016/j.
standardized 3d geo information, in: 5th Int. 3D GeoInfo Conf., Berlin, Germany, autcon.2017.06.018.
2010, pp. 9–12. http://www.3dgeoinfo.org/ISPRS_Conference_CD/Paper_ [47] E. Hjelseth, N. Nisbet, Exploring semantic based model checking, in: Proc. CIB W78
ISPRS/Oral/2_3DGeoInfo2010_105_Benner_BuildingLicensing.pdf. 2010 27th Int. Conf., Cairo, Egypt, 2010.
[42] L. Van Berlo, T. Dijkmans, J. Stoter, Experiment for integrating Dutch 3D spatial [48] S. Macit İlal, H.M. Günaydın, Computer representation of building codes for
planning and BIM for checking building permits, in: ISPRS Ann. Photogramm. automated compliance checking, Autom. Constr. 82 (2017) 43–58, https://doi.
Remote Sens. Spat. Inf. Sci, 2013, pp. 279–284, https://doi.org/10.5194/ org/10.1016/j.autcon.2017.06.018.
isprsannals-II-2-W1-279-2013.

15