Fakultät Bauingenieurwesen Institut für Bauinformatik Prof. Dr.-Ing. Raimar J.

Scherer

Application of graph databases and graph theory

concepts for advanced analysing of BIM models based
on IFC standard

Ali Ismail MS.c.

Ahmed Nahar MS.c.
Prof. Dr.-Ing. Raimar J. Scherer

EG-ICE 2017 Nottingham, 10.07.2017

 Motivation and objectives

 IFC to labeled property graph workflow:

 IFC Meta Graph(IMG) model

 IFC Object Graph(IOG) models

 Example of filters and queries based on the graph DB

 Conclusion and outlook

 Motivation & Objectives

Motivation

BIM models contain huge amount of information

and complex relationships between their elements.
This information could remain inaccessible due to the
absent of suitable data management tools.
Converting BIM models into an effective
information retrievable model based on Labeld
Property Graph (LPG) databases could significantly
facilitate exploring and analyzing the BIM highly
connected data
 Motivation & Objectives

Objectives
 Workflow for automatic transformation of IFC
EXPRESS schema and IFC models into IFC Meta
and Object Graph databases
 Utilizing graph theory concepts to explore, manage
and analyse all information inside BIM models by:

 Running queries for information retrieval.

 Topology analysis of the model.

 Integrated formwork for BIM models and linking

the with other project

EG-ICE 2017
information
Präsentationsname XYZ
 Workflow and tools

( Cypher Query Language )

Demo:

http://ifcwebserver.org/ifc2neo4j
Neo4j features
 Cypher Graph Query Language

Cypher is a declarative graph query language that allows for expressive 
and efficient querying and updating of a property graph. Cypher is a 
relatively simple but still very powerful language. Very complicated 
database queries can easily be expressed through Cypher
 Labeled Property Graphs

The labeled property graph model consists of a set of nodes and edges
(relationships) . An edge is always related to exactly two nodes with a fixed
direction from a start to an end node, defining the property graph as a directed
graph. Apart from that, two nodes can be connected by multiple edges at the
same time, thus being a multigraph. Both, nodes and edges, can store a set of
key-value pairs, called properties and nodes can be tagged with labels additionally.
 IFC Meta Graph (IMG) model

EXPRESS -> EXPRESS Parser -> Cypher -> Graph DB

 IFC classes

 Attributes

 Relationships
 IFC Meta Graph (IMG) model

Mapping mechanism

(1) Mapping of classes and

attributes
 IFC Meta Graph (IMG) model

Mapping mechanism

(2) Mapping of the relationships

Optional= true
 IFC Meta Graph (IMG) model

Path analysis Cypher Query

MATCH (a:IfcRoot{Version:'IFC2x3'})-[r*]-(b:IfcSlab)
RETURN r

Outcomes
 IFC Object Graph (IOG) models

BIM Software -> IFC -> IFCWebserver -> Cypher -> Graph DB

(1) Preparing IFC data

(2) Generate DB
 IFC Object Graph (IOG) models

Enhancement of Graph Database

MATCH (s:IfcSlab) SET
s:IfcBuildingElement:IfcElement:IfcProduct:IfcObject:IfcObjec
tDefinition:IfcObject:IfcRoot

Multi-models graph DB

Multi-labelled nodes
 IFC Object Graph (IOG) models

Filtration based on Multi-labeled

MATCH (n: IfcBuildingElement)RETURN n

 Filters and Queries

Verification of graph
database

Evaluation of graph
capabilities

Topology analysis
 Filters and Queries

Verification of Graph DB
 IFC class types and entity list
MATCH (n {Model: 'Muster003.ifc'})
WITH DISTINCT LABELS (n) AS LABELS, COUNT (n) AS
temp
Cypher command UNWIND LABELS AS Class_name
RETURN DISTINCT Class_name, SUM (temp) AS
Number_of_entities
ORDER BY Class_name

Class_name Number_of_entities
IfcApplication 1

IfcArbitraryOpenProfileDef 57

IfcAxis2Placement2D 33
Outcomes
IfcAxis2Placement3D 153

IfcBuilding 1

IfcBuildingStorey 1
IfcCartesianPoint 2534
…………… ………
 Filters and Queries

 Unconnected nodes

MATCH (n {Model: 'Muster003.ifc'})

WHERE NOT ((n)--())
WITH DISTINCT LABELS (n) AS LABELS, COUNT (n) AS temp
Cypher command UNWIND LABELS AS Class_name
RETURN DISTINCT Class_name,SUM(temp) AS
Number_of_unconnected_entities
ORDER BY Class_name

Number_of_unconnected_enti
Class_name
ties
IfcDerivedUnit 2

IfcMaterial 3
Outcomes
IfcPresentationLayerAssignment 4

IfcPresentationStyleAssignment 16

IfcProductRepresentation 9
 Filters and Queries

Evaluation of graph capabilities

Filter objects based on property values

MATCH (wall: IfcWallStandardCase

{Model:'Muster003.ifc',IFCID:'2091'})-[rel]-
(property:IfcPropertySet)
RETURN DISTINCT wall,rel,property

Outcomes
 Filters and Queries

Evaluation of graph capabilities

Get name of assigned materials of certain object(s)

MATCH
(d: IfcDoor {Model: 'Muster003.ifc‘, globalId:xyz})-[*1..5]-(m: IfcMaterial)
RETURN DISTINCT (m.IFCID) AS IFCID, (m.name) AS material_name
 Filters and Queries

Simplified analysis of model topology for emergency routes

Construction of possible paths

Navigation routes through

accesses

MATCH p=(s1:IfcSpace)-
[r1:BoundedBy]-
(d:IfcDoor)-
[r2:BoundedBy]-
(s2:IfcSpace)
WHERE s1.IFCID >
s2.IFCID
RETURN p
 Filters and Queries

Navigation routes through

space boundaries

MATCH p=(s1:IfcSpace)
<-[r1:RelatingSpace]-
(b:IfcRelSpaceBoundary)
-[r2:RelatingSpace]-
>(s2:IfcSpace)
WHERE s1.IFCID >
s2.IFCID
RETURN p
 Filters and Queries

Retrieval query for emergency

routes:
Emergency exit door.
Accesses navigation routes.
Space boundary routes.
MATCH
p=(d:IfcDoor{IFCID:'9824',Model:'Muster003.i
fc'})-[r:BoundedBy]->(s:IfcSpace)
RETURN p
UNION
MATCH
p=(s1:IfcSpace{Model:'Muster003.ifc'})-
[r1:BoundedBy]-(d1:IfcDoor)-[r2:BoundedBy]-
(s2:IfcSpace)
WHERE s1.IFCID > s2.IFCID
RETURN p
UNION
MATCH p=(s1:IfcSpace
{Model:'Muster003.ifc'})
<-[r1:RelatingSpace]-(r:IfcRelSpaceBoundary)
-[r2:RelatingSpace]->(s2:IfcSpace)
WHERE s1.IFCID > s2.IFCID RETURN p

Slide 24 of 40
 Filters and Queries

‘Office_A’ model- First floor layout

Slide 25 of 40
 Filters and Queries

‘Office_A’ model-Neo4j

Slide 26 of 40
 Filters and Queries

Compare 2 Versions of IFC models

Week 26 Week 27
Week 28

Week 29 Week 30 Week 37

https://github.com/openBIMstandards/DataSetSchependomlaan

Find out the new windows installed in the Week37 compared with the
Week30
Filters and Queries
 Discussion and conclusion

Achievements

Create IFC Meta Graph (IMG) model

Generic approaches

Create IFC Object Graph (IOG) models

Support understanding of IFC data schema

Cypher queries

Analyse building information inside BIMs

Provide realistic answers for typical queries

Pre-defined queries

Run advanced analysis of BIM models

 Discussion and conclusion

Outlook

 Improve the mapping and converting process to

cover 100% of all relationships and inverse
relationships.
 Integration with a geometry engine and include
geometry information in the graph DB (positions,
BBOX, volumes, etc.)
 User friendly web application for accessing and
managing the graph database
 Performance and benchmark tests
Thank you for your Attention

Questions ?

e-mail: ali.ismail@tu-dresden.de