ldac 2018 paper - linked datalinkedbuildingdata.net/ldac2018/files/presentations... · checking...

Toward French Semantic Regulationchecking based on semantic rules

LDAC 2018 – Paper

CSTB/TI - Nicolas BUS; Guillaume PICINBONO

2

Context

« Digitizing French Building Regulation »

=> Methods and POC

Sponsor : French Ministry for Housing

18 Month => July 2018

Involving regulation experts and knowledge enginneers

Lead by CSTB

3

Checker

Building Model

Set of

Contraints

(requirements)

List of

non-compliant

elements

Recap : principle a compliance checker

4

Objectives

1. Build an extensible and white box compliance

checker

2. Bridge the gap between human readable texts and

IFC concepts

5

What does white box means ?

- Easy to extend and to add new constraints

- Constraint are decoupled from checker implementation

- Constraints and are auditable

- Constraints use a standard language

6

How large is the gap ?

- Regulation and BIM use non-bijective vocabularies

- Regulation use a lot of synonyms

- The regulation is ambiguous

- The regulatory text is not autonomous (lots of references)

- IFC is verbose and low level

- IFC geometry is polymorph with high level of detail

7

Bringing regulation closer to BIM

Regulation

text

Semi-formal

contraints

Construction experts

8

Bringing regulation closer to BIM

Regulation

text

Semi-formal

contraints

Controled vocabulary

Controled syntax

Editing tools

Construction experts

9

Bringing regulation closer to BIM

Regulation

text

Controled vocabulary

Controled syntax

Construction experts

Regulation

Ontology

+

RulesKnowledge engineer

Semi-formal

contraints

10

Contraints examples

SEMI-FORMAL CONSTRAINT

IF structure_element in latest_storey no fireproof THEN_NON_COMPLIANT

element fireproof element fire_rating > 1h

Always begins with« IF »Always ends with

« THEN_NON_COMPLIANT »

Use controled vocabulary

REGULATION TEXT

All elements that takes part to the structure, under the roof, slab must be resistant enough in

case of fire (see fire safety standard…).

Equivalence

11

Regulation Ontology

IFCOWL

Simplification

Alignment

High level concept

Fre

nch

Re

gu

latio

n O

nto

log

y

Properties,classifications, types,

Geometry, topology

Regulation vocabulary built on

IFC concepts

Regulation concepts built on

other regulation concepts

12

Regulation Ontology

IFCOWL

Simplification

Alignment

High level concept

:RelDefinedByProperties

Fre

nch

Re

gu

latio

n O

nto

log

y

:FireRating

:FireProof

:ProtectedSpace

13

Simplification example : properties

IFC-OWL way to express an object property

Simplified way

Equivalence

14

Digitizing regulation contraints

Regulation

text

Semi-formal

contraints Semantic

constraints

Controled vocabulary

Controled syntax

Construction experts Knowledge engineer

Regulation

Ontology

+

RulesKnowledge engineer

15

Triple

Store

IFC RDF

Geometrical

Semantic

Preprocessor

SPARQL-

constraints

BCFzip

Regulation

ontology

+

SPARQL-rules

Semantic Compliance Checker

16

Integration with a BIM platform (KROQI)

Watching report as

BCF annotations

17

Improving geometry level of detail

Improving topological inferences

Improving rules classification

Applying constraints corresponding to building type

Improving Ontology and rule maintenance

Automating Ontology documentation including Rules

What next ?

Thanks your attention…

Toward French Semantic Regulationchecking based on semantic rules

LDAC 2018 – Paper

CSTB/TI - Nicolas BUS; Guillaume PICINBONO

nicolas.bus@cstb.fr

Appendices

20

Contraints examples

SEMI-FORMAL CONSTRAINTIF structure_element in latest_storey no fireproof THEN NON-COMPLIANT

SPARQL-RULES + ONTOLOGY?e :LoadBearing true

=>

?e :StructureElement

…

SPARQL CONSTRAINTSELECT ?e

?building :LatestStorey ?storey .

NOT EXISTS (?e a :FireProof)

?e a :StructureElement; :In ?storey .

21

Geometrical

preprocessor

Simplification examples : geometry

Bounding boxes

Polymorph high LOD geometry

Simplified geometry

22

WC : <wc1>

Clear space <cs1_1>

Clear space <cs1_2>

WC : <wc2>

Clear space : <cs2_1>

Clear space : <cs2_2>

Hand rail

23

KROQI plateform

28/09/2015 - Auteur Nom

IFC file uploaded

24

Contraints examples

SEMI-FORMAL CONSTRAINTIF no clear_space of width 0.8m, depth 1m beside a WC_seat THEN NON-COMPLIANT

RULE + ONTOLOGY?wc a :WCSeat

=>

?wc :NeedClearSpace ?cs1 ; :NeedClearSpace ?cs2 .

?cs1 :Width 0.8 ; :Depth 1 .

…?cs1 :X (?WC_X + ?dX) .

?cs1 :Y (?WC_Y) .

CONSTRAINTSELECT ?wc

…?wc a :WC ; ?wc :NeedClearSpace ?cs

?e :Intersects ?cs

...

HAVING COUNT(DISTINCT ?cs)=2

?cs1Clear space

?cs2Clear space

?wcWC Seat

25

IFC

SPARQL-

constraints

BCFzipChecker Webservice

Contraints examples

26

KROQI Platform – Launching service

Executing

checker service

27

KROQI Platform – Choosing a set of constraints

Selecting a

Contraints set to apply

28

KROQI Platform – Displaying results

Reading the report

29

KROQI Platform

Watching report as

BCF annotations