Ontology-based projects in industry - INF3580 guest lecture · Provided by IHS under license with...
Transcript of Ontology-based projects in industry - INF3580 guest lecture · Provided by IHS under license with...
SAFER, SMARTER, GREENER
OIL & GAS
Ontology-based projects in industryINF3580 guest lecture
Johan W. KlüwerMay 23, 2016
Security and Information Risk Management (SIRM)
Information ManagementRight information to the right people at the right time – with theright quality
Integration and InteroperabilityAccessible and consistent life cycle information
Information and Cyber SecuritySecuring business, investments, and assets
Systems and Software ReliabilityImproving performance of complex integrated software dependentsystems
2 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
O&G project: Material Master Data @ Aibel
WhoEPC contractor Aibel supported by DNV GL SIRM
WhatUniform information management across all EPC project phases:Requirements, designs, and products
Why� Better quality� Higher efficiency
HowOntology-based management of requirements, designs, and physicalassets. Semantic technology for exchange and reasoning services.
3 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Johan Sverdrup Drilling Platform (JSDP)
MMD is Aibel’s material master for piping bulk and structural steel at JSDP
4 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Health domain project: Optique pilot @ AHUSWhoAkershus University Hospital (AHUS) supported by DNV GL SIRM incollaboration with Logic and Intelligent Data (LogID), U. of Oslo
WhatOptique project commercial pilot. Integrated access to patientinformation across domains and specialist roles: diseases,diagnoses, tests, procedures
Why� Better quality� Higher efficiency
HowOntology based data access (OBDA) for medical conditions,diagnoses, and clinical processes. Integration across data sourcesand semantic document search.
5 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Relevance to industrial information in general
WhoEvery industry is subject to complex requirements
WhatSmart management of requirements, rules, solutions, andknowledge in complex scenarios. Data access, informationexchange, documentation, experience transfer.
Why� Better quality – validated basis for decisions� Higher efficiency – simplified and faster work processes
HowApply ontology based methods and semantic technologies toimplement Enterprise Semantics
6 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Computers are unsophisticated (tabloid version)
1. Everyone values precise language and consistent rules.
2. Our information systems ought to implement them.
3. No one believes 2. is possible.
7 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Timeline: SIRM in semantic technology
2000 2005 2010 2015
PoscCaesar
ISO15926
IIP IDS
DDR
IOHNRosatom
EPIM
Optique
Aibel MMD
BYTE
Ahus
20 years of innovation in enterprise semantics – sample projects
2000 2005 2010 2015
UML 1.0
XML 1.0
RDF,OW
L 1, D2RQ
BFO
LinkedDataproposal
SPARQL
R2RML,OW
L 2
Semantic technology standards – mature enough to deliver great value
8 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Enterprise semantics supporting analytics
9 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Analysis – Pulling insights from data
A basic pattern: We interpret data to obtain a suitable set of facts as a body ofevidence, then analyze for general, higher-level facts – insight.
interpretation facts insight
Our interpretation applies domain knowledge to data, mapping database recordsinto facts that are suitable for analytic methods.
Domain knowledge: Standards, regulations, best practice, experience.Databases: Relational (SQL), hierarchical (XML), other (NoSQL).Analytic methods: Statistics, machine learning.
9 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
The traditional, case-by-case approach to finding facts
Using our domain knowledge, we program transformations to produce anintermediate data set to which analytic tools can be applied.
Workflows tend to be one-off: tailor made for a limited scope, and difficult tovalidate, extend or compare with other analytic setups.
domain knowledge schema
interpretation
records analytic methods
facts insight
10 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
The enterprise semantic, uniform approachWe separate the representation of domain knowledge from any particularinterpretation challenge, producing a generic ontology that can be applied to arange of databases in the same problem domain.The ontology serves as a common language and rule base – verifiable, reusable,and extendable – but requires more work up-front.
semantic
non-semantic
domain knowledge schema
interpretation
ontology
records
interpretation
analytic methods
facts
facts
insight
insight
11 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Competing with algorithms.
Finansavisen, November 09, 2015:
12 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Competing with algorithms..
Finansavisen, November 09, 2015:
13 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Competing with algorithms…
Finansavisen, November 09, 2015:
14 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Documentation cost
Increasing complexity is driving development cost in offshore O&G.
15 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Complexity cost
Teknisk Ukeblad, March 13, 2016:
“Over to år forsinket. … Goliat har støtt på enorme problemer, med storeteknologiske utfordringer, med tre dødsfall på verftet i Sør-Korea, samt stadigeutsettelser og budsjettoverskridelser på over 16 milliarder kroner.”
16 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Understanding requirements – optimizing execution
� Domain standards� Regulations� Customer/company requirements� Best practice
Industrial ontologies provide the only candidate for a way out.
17 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Aibel’s Material Master Data (MMD) project
18 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
About Aibel
� Energy service enterprise� Oil & Gas� Renewable energy
� Process industry lifecycle� engineering� construction� upgrading� maintenance
� 5000 employees� Norway, Thailand, Singapore,UK, Denmark
18 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
The EPC project
Projectdefinition
Systemengineering
Area engineering Procurement Construction
handover handover handover handover
19 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
The need for efficient information flows
A typical EPC project at Aibel requires work in 100 different specialist ITapplications.
Every major company today relies on application mappings written on acase-by-case basis. These are typically costly to maintain, extend, or validate. Youhave a web of ad hoc interfaces that is both fragile (don’t touch, it might break) andan impediment to evolving business processes (don’t touch, we rely on this).
MMD provides a common language for the information that needs to be exchangedbetween some crucial specialist systems.
� mappings can be validated against a common schema� information from different project phases is consolidated
20 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
The need for intelligent services
Any EPC project involves complex requirements, across 20+ major disciplines� Customer specifications� Public rules and regulations� Standards� Best practice
MMD provides a rule base that supports automated reasoning to assist experts inquickly and correctly identifying optimal engineering choices within the complex ofproject constraints.
21 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
The need for end-user access
Experts are wasting their time withinefficient information services
� searching for data� consolidating information oncethey get it
� validating, updating
Non-standard information needs arehandled ad hoc, by busy IT experts,with little regard for re-use of theaccess methods.
“A quite common situation in several or all myprojects so far, I’ve had the need to check and crosscheck data which in our current tool was quitedifficult, perhaps combine the data between tablesand hopefully not having to do it in Excel all thetime.
However, the program had an option to do SQLbased searches. As few or none of the engineersknow SQL coding, I had the IT support or systemsupport make me the SQL string that was needed inorder to make a specific search, and I then copied itinto a text field so that I could use it later or try tomodify it.”
(Aibel engineer)
22 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
The MMD solution
Material Master Data (MMD) supports project execution from end to end, based onan enterprise ontology:
� a common language for communication and exchange� a knowledge base of engineering and business rules
A shared language across project stages – and across projects
23 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Uniform project data with enterprise semantics
� Interface applications by way of the ontology� Automated validation of solutions vs. requirements
24 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
MMD as part of an enterprise architecture
� Enterprise semantics is architecture, not a function� Expert knowledge that the computer can process� A unified library of requirements and solutions
� Designs, Customer requirements, Standards, Regulations� Products, manufactured items, and facilities
� Enabled by state of the art automated reasoning software� Existing applications and databases are largely untouched� Support coherence through project phases� Documentation for operations and maintenance
25 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Where can MMD be useful?
� Business Process improvement by master data� Success at Aibel for project execution� Operations should offer many possibilities� Governance in general can benefit
� Knowledge Management� Capturing expert knowledge� Knowledge sharing
� (Big) Data� Integration of specialist applications and databases� Validation, data quality
26 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
MMD collaboration opportunity
� Aibel developed MMD with help from DNV GL� Standards based – relevant across O&G domains and roles� They will share it for a reasonable return� The whole industry could benefit
→ Re-implementation would be wasteful
27 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Value of MMD
� Not one business model, but many� An extension of standardization that can benefit all players
� Quality – due to better control of complex content� Speed/efficiency/agility – respond to changing requirements� Reliability – trace changes, find original design basis� Transparency – less room for glossing over irregularities
28 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Timeline: Aibel MMD project
2013 2014 2015 2016
Life of aWell
Enterprise registryconcept
Process andpiping
modelling
Aibel PoCdemo
SemanticDayspresentation
Webapplicationfor pipingspecs
Extendintonewdomains
Aveva PDMSintegration
SAP integration
OptiquesummitHøvik
JohanSverdrupDP
piping, structural
Extendinto new
domains
A dedicated team of specialists has developed the 2012 semantic enterpriseconcept into a material master system that is in production for the Johan SverdrupDrilling Platform project (8 BNOK).
29 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Semantics for the EPC project
30 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Master data in a modular ontology
AL-PROFILES_DIM
structural-geometry
structural-coreS-Working
DIN_1025-1_DIM
DIN_1026-2_DIM
EN_10056-1_DIM
EN_10210-2_DIM
EN_10219-2_DIM
EURONORM_19_DIM
EURONORM_53_DIM
NS_1911_DIM
STR_AIBEL_DIM
STR_BAR_DIM
STR_FASTENER_DIM
STR_HP_DIM
STR_NUT_DIM
STR_PLATE_DIM
STR_THREAD_FEA
STR_TUB_FEA
STR_WASHER_DIM
core-collect
ALLOY_MAT ASTM_BOLT_MATASME_B16_5_MAT
ASME_B31_3_MAT
ASTM_MAT
API_594
piping-valve
piping-collect
API_600_VALVES
API_602_VALVES
API_609
API_6D_DESIGN
API_6D_DIM
ASME_B16_10_DIM
ASME_B16_11_DIMS-Proposed
ASME_B16_11_MOD
ASME_B16_20_DIMS-Working
ASME_B16_21_DIMS-Proposed
ASME_B16_34_VALVES
ASME_B16_36_DIM
ASME_B16_36_MOD_DIM
ASME_B16_47_DIM
ASME_B16_5_DIMS-Proposed
ASME_B16_9_DIMS-Working
ASME_PCC_1_DIM
BOLTING_DIM
BOLTING_DIM_2HH0W
BOLTING_DIM_2HH1W
BOLTING_DIM_2HH2W
ELBOW_5D_DIM
FLANGE_TAPER_B_DIM
MANIFOLD_FIT_DIM
MANUFACTURER_VALVES
MECH_J_DIM
NORSOK_L_005_DIM
OLET_DIMS-Working
STATOIL_ENSS-Proposed
STATOIL_SBR3_DIM
STATOIL_SLBNS-Proposed
domains-collect
API_MATNORSOK_M630_MAT
materials-coreS-Working
ASME_B16_20_MATS-Proposed
ASME_B16_21_MATS-Proposed
EN_10025_MAT
EN_10210_MAT
EN_10219_MAT
EN_10225_MAT
EN_485_MAT
EN_755_MAT
TR2000_COMMODITY
ASME_B16_11_CAD_ATTRS-Working
piping-bulk
ASME_B16_11_MOD_CAD_AT
ASME_B16_20_CAD_ATTR
materials-collect
ASME_B16_21_CAD_ATTR
ASME_B16_25_FEA
piping-coreS-Working
ASME_B16_47_FEA
ASME_B16_5_FEAS-Proposed
ASME_B36_10_19_FEAS-Proposed
END_PREP_FEA
FITTING_PATTERN_FEA
ISO_4200_FEA
MECH_J_FEA
NORSOK_L_005_FEA
OLET_FEA
ASME_B16_47_CAD_ATTR
ASME_B16_5_CAD_ATTRS-Proposed
ASME_B16_9_CAD_ATTRS-Working
ASME_B36_10_19_CAD_ATT
BOLTING_MAT_CAD_ATTR
STR_FASTENER_MAT
BOLTING_INTERNAL_DIMPIP_BOLT_CAD_ATTR
statoil-spec
STATOIL_MAT
MDS_TEST_CRIT_STATOIL
ELBOW_5D_CAD_ATTR
NORSOK_M120_MAT
NORSOK_M121_MAT
MANIFOLD_FIT_CAD_ATTR
VALVE_COMMODITY
commodity
MECH_J_CAD_ATTR
NORSOK_L_005_CAD_ATTR
OLET_CAD_ATTRS-Working
statoil-client
STR_STOCK_NO
STATOIL_SLBN_CAD_ATTRS-Working
structural-collect
STR_TUB_DIM
SYSTEMSYSCODE_MMD
TR2000_VDS
UNSPSC_REF
client-core
admin
manuf-core
client-spec da-core
controlS-Proposed
coreS-Working
MMD_USERS
WCO_HSS-Working
disciplinesS-Working
skos-MMDS-Official
standardsS-Working
ANNULUS
IEC_60079_EX
IEC_60529_IP
MMD_UOM
PLATE_DIM
RAL_COD
UNSPSCS-Working
da-collect
pdc
PIP_DA
STR_DA
manuf-data
manufacturers
prod-core
UNS_MAT
pav-MMDS-Official
iso15926-2S-Official
PDC_PIP_0000000100
PDC_PIP_0000000180
PDC_STR_0000000323
PDC_STR_0000000324
PDC_STR_0000000325
PDC_STR_0000000326
product
PIP_QN
STR_QN
statoil-plants
30 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Customer requirements
(Datasheet from Statoil’s TR2000 Piping Class Specification repository)
31 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
The disciplines
� Architecture and Building� Civil Engineering� Drilling� Electrical� Engineering Management andAdministration
� Geoscience� Health, Safety andEnvironment
� HVAC� Instrumentation� Marine Engineering� Material Technology� Mechanical� Multidiscipline� Operation andMaintenance
� Pipelines
� Piping� Process� Procurement� Reservoir� Structural� Subsea� Telecommunication� Weight and QuantityControl
(NORSOK)
32 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
The standards
LargeRJiameterSteelRjlangesNPSR8VRThroughRNPSRVWMetrick ånchRStandard
3NR 3M =Rå Y3N R N3Tå ON 3LR S T3NJ3RJ
3SM=R@:V,yU98W::-RevisionRofR3SM=R@:V,yU98WWV(
YopyrightR3SM=RånternationalProvidedRbyRåqSRunderRlicenseRwithR3SM= Licensee5VetcoR3ibelkzH8zUb::WU)RUser5=lsfjordstrand)RStian
NotRforRResale)RWzk8pk8W:bRW8´:y´zpRMJTNoRreproductionRorRnetworkingRpermittedRwithoutRlicenseRfromRåqS
--`,`,`,``,``,,,,,,,``,`````,`,`-`-`,,`,,`,`,,`---
3NR 3M =Rå Y3N R N3Tå ON 3LR S T3NJ3RJ
3SM=R@:V,::98W::-RevisionRofR3SM=R@:V,::98WWH(
jorgedRjittings)Socket9WeldingandRThreaded
Lopyright T5P[5TbyTtheTRmericanTSocietyTofTMechanicalT:ngineers9No reproductionTmayTbeTmadeTofTthisTmaterialTwithoutTwrittenTconsentTofTRSM:9
cYopyrightR3SM=RånternationalProvidedRbyRåqSRunderRlicenseRwithR3SM= Licensee5VetcoR3ibelkzH8zUb::WU)RUser5=lsfjordstrand)RStian
NotRforRResale)RWzk8pk8W:bRW8´:z´yHRMJTNoRreproductionRorRnetworkingRpermittedRwithoutRlicenseRfromRåqS
--`,`,`,``,``,,,,,,,``,`````,`,`-`-`,,`,,`,`,,`---
Jesignation´ 3:p8k3:p8M − :8b =ndorsed by Manufacturers StandardizationSociety of the Valve and jittings åndustry
Used in USJO=9N= Standards
Standard Specification forjorged or Rolled 3lloy and Stainless Steel Pipe jlanges)jorged jittings) and Valves and Parts for qigh9TemperatureService:
This standard is issued under the fixed designation R[“53R[“5MD the number immediately following the designation indicates the yearof original adoption or4 in the case of revision4 the year of last revision9 R number in parentheses indicates the year of last reapproval9R superscript epsilon )´ I indicates an editorial change since the last revision or reapproval9
This standard has been approved for use by agencies of the Department of Defense.
1. Scope*
[9[ This specification5 covers forged low alloy and stainlesssteel piping components for use in pressure systems9 Includedare flanges4 fittings4 valves4 and similar parts to specifieddimensions or to dimensional standards4 such as the RSM:specifications that are referenced in Section 59
[95 _or bars and products machined directly from bar )otherthan those directly addressed by this specificationD see q9]I4refer to Specifications R]H”3R]H”M and RHk” for the similargrades available in those specifications9 Products made to thisspecification are limited to a maximum weight of [P PPP lb[]-]P kg]9 _or larger products and products for otherapplications4 refer to Specifications Rkkq3RkkqM and R”q-3R”q-M for the similar ferritic and austenitic grades4respectively4 available in those specifications9
[9k Several grades of low alloy steels and ferritic4martensitic4 austenitic4 and ferritic7austenitic stainless steelsare included in this specification9 Selection will depend upondesign and service requirements9 Several of the ferritic3austenitic )duplexI grades are also found in SpecificationR[P]”3R[P]”M9
[9] Supplementary requirements are provided for use whenadditional testing or inspection is desired9 These shall applyonly when specified individually by the purchaser in the order9
[9- This specification is expressed in both inch7pound unitsand in SI units9 ’owever4 unless the order specifies theapplicable “M” specification designation )SI unitsI4 the mate7rial shall be furnished to inch7pound units9
[9q The values stated in either SI units or inch7pound unitsare to be regarded separately as the standard9 Within the text4the SI units are shown in brackets9 The values stated in eachsystem may not be exact equivalentsD therefore4 each systemshall be used independently of the other9 Lombining valuesfrom the two systems may result in non7conformance with thestandard9
2. Referenced Documents
59[ In addition to the referenced documents listed in Speci7fication R”q[3R”q[M4 the following list of standards apply tothis specification9
595 ASTM Standards:k
R5q5 Practices for –etecting Susceptibility to IntergranularRttack in Rustenitic Stainless Steels
R5H-3R5H-M Practice for Magnetic Particle :xamination ofSteel _orgings
Rkkq3RkkqM Specification for Rlloy Steel _orgings forPressure and ’igh7Temperature Parts
Rk““3Rk““M Practice for Ultrasonic :xamination of Steel_orgings
R]H”3R]H”M Specification for Stainless Steel jars andShapes for Use in joilers and Other Pressure Vessels
R]“]3R]“]M Specification for @eneral Requirements forStainless Steel jars4 jillets4 and _orgings
RHk” Specification for Steel jars4 Rlloy4 ’ot7Wrought4 for:levated Temperature or Pressure7Lontaining Parts4 orjoth
RHqk Practices for –etecting Susceptibility to IntergranularRttack in _erritic Stainless Steels
RH““3RH““M Specification for Steel _orgings4 @eneral Re7quirements
R”q[3R”q[M Specification for Lommon Requirements for[ This specification is under the jurisdiction of RSTM Lommittee RP[ on Steel4
Stainless Steel and Related Rlloys and is the direct responsibility of SubcommitteeRP[955 on Steel _orgings and Wrought _ittings for Piping Rpplications and joltingMaterials for Piping and Special Purpose Rpplications9
Lurrent edition approved Nov9 [4 5P[59 Published –ecember 5P[59 Originallyapproved in [”k-9 Last previous edition approved in 5P[5 as R[“53R[“5M–[5a9–OIW [P9[-5P3RP[“5_RP[“5M7[5b9
5 _or RSM: joiler and Pressure Vessel Lode applications see related Specifi7cation SR7[“5 in Section II of that Lode9
k _or referenced RSTM standards4 visit the RSTM website4 www9astm9org4 orcontact RSTM Lustomer Service at serviceGastm9org9 _or Annual Book of ASTMStandards volume information4 refer to the standard’s –ocument Summary page onthe RSTM website9
*A Summary of Changes section appears at the end of this standard
Yopyright © 3STM ånternational) :WW @arr qarbor Jrive) PO @ox YUWW) West Yonshohocken) P3 :Hy8p98HzH, United States
[YopyrightR3STMRånternationalProvidedRbyRåqSRunderRlicenseRwithR3STM Licensee5VetcoR3ibelkzH8zUb::WU)RUser5=lsfjordstrand)RStian
NotRforRResale)RWbk8Vk8W:bRWH´zb´zbRMJTNoRreproductionRorRnetworkingRpermittedRwithoutRlicenseRfromRåqS
--`,,`,`,,`,`,,```,`,,,`,```,``-`-`,,`,,`,`,,`---
NEK:TS:606:2009=ngelskRversjon
Utg,Ry,WR8WWH
NorskRelektrotekniskRnorm
Kabler:for:offshore:installasjonerhalogenfrie:og/eller:borevæskebestandigeTeknisk:spesifikasjon
NorwegianRelectrotechnicalRstandard
Cables:for:offshore:installationshalogen-free:and/or:mud:resistantTechnical:specification
NORSKR=L=KTROT=KNåSKRKOMåT=NorskRnasjonalkomiteRforånternationalR=lectrotechnicalRYommission)Rå=YYomitéR=uropéenRdeRNormalisationR=lectrotechnique)RY=N=L=Y© N=KRharRopphavsrettRtilRdenneRpublikasjon
Pro
vide
dRby
RSta
ndar
dRO
nlin
eR3
SRfo
rRS
tato
ilqyd
roR3
S3
R9R3
bonn
emen
tRpåR
web
R8WW
H9Wz
98z,
RRep
rodu
ctio
nRis
Rnot
Rallo
wed
,
ThisRNORSOKRstandardRisRdevelopedRwithRbroadRpetroleumRindustryRparticipationRbyRinterestedRpartiesNorwegianRpetroleumRindustryRandRisRownedRbyRtheRNorwegianRpetroleumRindustryRrepresentedRbyRTheOilRåndustryR3ssociationR-OLj(RandRTheRjederationRofRNorwegianRåndustry,RPleaseRnoteRthatRwhilstRevebeenRmadeRtoRensureRtheRaccuracyRofRthisRNORSOKRstandard)RneitherROLjRnorRTheRjederationRofRNoråndustryRorRanyRofRtheirRmembersRwillRassumeRliabilityRforRanyRuseRthereof,RStandardsRNorwayRisRresponadministrationRandRpublicationRofRthisRNORSOKRstandard,
StandardsRNorwayR Telephone´RfRyURVURpbRpVRWWStrandveienR:p)RP,O,R@oxR8y8R jax´RfRyURVURpbRpVRW:N9:b8VRLysakerR =mail´Rpetroleum+standard,noNORW3YR Website´Rwww,standard,nokpetroleum
YopyrightsRreserved
NORSOK:STANDARD:Editio
Architectural:components:and:equipment
Specification:for:Line:Pipe
3PåRSP=YåjåY3TåONRzLjORTY9jåjTqR=JåTåON)RJ=Y=M@=RR8W:8
=jj=YTåV=RJ3T=´RJULYR:)R8W:b
--`,,,`,````````,`,`,`,````,,,,-`-`,,`,,`,`,,`---
33 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Many standards!
� Shapes and Dimensions� Just for Piping, around 90 standards listed in ASME B31.3
� Materials and Fabrication� around 150
� Standards have a long life� Revisions are issued on a regular basis
� The body of standards grows with introduction of new technologies� Standards bodies include ASME, ASTM, API, ISO, and NORSOK
34 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Representing a standard as an ontology module
� What to include?� Tabular material listing standard-compliant variants� Reading the text� Rules to calculate commonly used attributes
� Example: Inner diameter from OD and WT� Separate out generic classes as “core”
36 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Reference data in standards
� Geometry
� Attributes� NPS� Outside diameter� Center-to-End
� A “standard class”� Elbow 90 Long RadiusASME B16.9 NPS 4
Table 1 Dimensions of Long Radius ElbowsA
A
B
B
Center6to6EndNominalPipe Outside 906deg 456degSize Diameter Elbows8 Elbows8vNPS2 at Bevel A B
1⁄2 2173 38 163⁄4 2677 38 191 3374 38 2211⁄4 4272 48 2511⁄2 4873 57 29
2 6073 76 3521⁄2 7370 95 443 8879 114 5131⁄2 10176 133 574 11473 152 64
(ASME B16.9)
37 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Standard classes in ontology
(Protégé 4.3)
38 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Compositional definition
(OWLViz, Protégé 4.3)
39 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Levels of specificity: The design article
generic class Pipe Flange
core class Pipe Flange and’has Flanged End’ some ’Annulus 21.3 x 2.77’ and’has Pipe End’ some ’Annulus 21.3 x 2.77’
standard class ASME B16.5 Pipe Flange NPS 1/2 CL 1500 Weld Neck Ring type joint
design article class FLG WN NPS 1/2 CL 1500 WT 2.77 MDS DF101
product class (compliant product – vendor numbering scheme)
40 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Compact flange with connectivity
41 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Query across standards
� DL Query fits the engineering task of identifying components� Combine with SPARQL for data retrieval
42 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Query across standards
� Find duplex steel components in nominal size 4
43 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Validating project data against requirements
44 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Integrating PDMS and SAP with MMD
� Aveva PDMS: Leading 3D CAD tool� Get rich non-geometric object information from MMD� Maintain and update with changing project requirements
� SAP: The inventory� Get product type identifiers from MMD� Refer to MMD for full content and link back to requirements
45 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Tabular sources for bulk classes
� Engineers know how to work with spreadsheets� “Lifting” with rules in SPARQL
46 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Architecture overview
� Oracle semantic server� SQL, XML, RDF, …� Access control� Code on the server
� ETL for populating ontology� R2RML� Linked Data� Mediawiki� Spring Java wizards� VCS
Standards SQLOntology
core
standard
design
product
Maintainer apps End user/project apps
Look up
Visualize
Exchange
Verify
Extend
47 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Team composition
DNVDevelopers:
� Ontology� Linked Data� Web application� Database application
Aibel� Discipline specialist engineers
� Deep knowledge of governing standards� Experience from project work
� IT developers� Deep knowledge of specialist applications� Deep knowledge of enterprise IT architecture
� Training and communcation staff
48 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Development process
� Traditional development: Domain experts hand over specifications, programmersprogram
� Semantic development: Domain experts work with ontology designers.� Identify core notions� Find the governing standards� Identify what to model� Build a skeleton core� Build and test some examples� Set up spreadsheets for data capture� Rules for lifting tables to ontology� Consolidation with other modules� Test, rinse, repeat
49 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Ontology metadata
� SKOS for various code books� Provenance: POV, based on PROV-O� VoID� Revision histories� Permissions� Status proposed/approved/etc.
50 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Bad tooling
� OWL Full is still favored by some� Reasoner support
Example, TBC faq:How can I run OWL 2 inferences in TopBraid Composer?To configure TopBraid Composer for RDFS/OWL 2 reasoning using TopSPIN, go tothe Profile sub-tab of the Ontology Home (see figure below). Choose the reasoningprofile you wish to use.It’s crucial to be able to verify consistency during ontology development. Thedegree of integration of reasoners does make a difference.
51 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Ontology quality control
QA is key.
� Top-down across imports� Reproducibility!� Syntax checking – SPARQL test methods� DL Reasoning a hard requirement for maintaining consistency� Integrity constraint checking� Logic of ontology modules� Metaphysically motivated sanity checks� Heuristics of “balance” – topical modules� Domain specific rules� Continous Integration (Jenkins – Chris Mungall)
Q. What are the rules for building large, robust ontologies?
52 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
The long road to adoption
� Timing is crucial to success� Show business value along the way
Enterprise process
observe implementappraiselearn evaluate
join testchooseusecase
buildteam
milestones
activities
53 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Standards as ontologies
We need the standards organizations to start issuing their standards as ontologies:as machine-readable, and compatible across different standards, even those issuedby other organizations. We need conformance criteria to be openly available, withreproducible tests and support for verification by independent bodies. Proprietarysoftware has not been up to the task.
54 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Competing upper ontologies – semantic silos
� ISO 15926: No success for standardization of industry classes through ISO� BFO: Lack of organisation, coordination across projects� Both have OWL troubles
� BFO: Time troubles� ISO 15926: Higher-order troubles
55 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Panoptique: Application, 2016
“Ontologies and declarative mappings will be used to capture domainconceptualisations, facilitate the specification of application level data needs, andtransform them into highly optimised queries over a wide variety of Big Datasources. PANOPTIQUE will contribute to the development of ontology standards forthe process industry that can function as the organizing principle of an ecosystemof tools and work processes that is open to a range of specialized service providersand tool developers, and that covers the whole value chain. To promoteinteroperability, ontologies used in the PANOPTIQUE pilots will share a commoncore; the generic parts of Aibel’s ISO 15926 based MMD ontology can provide sucha core, and one that has already proven its usefulness in a production setting.PANOPTIQUE will use state of the art ontology engineering tools, many of whichhave been developed by members of the consortium, to ensure that ontologiesmeet rigorous QA standards.” (p. 19).
56 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Transformative improvement
What we need:1. Automated requirements management2. Automated verification of solutions against requirementsPart 1. requires cross-disciplinary collaboration on machine-interpretablestandards, to be publisehd by the respective entities responsible – government,standardisation organisations, and industry groups. This needs to happen in bothnational and international contexts.Part 2. requires technology to make information conformant, and verificationmethods available during development and use in projects/work processes.
57 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry
Thanks!
Johan W. Klüwer
58 / 58 DNV GL © 2016 May 23, 2016 Ontology-based projects in industry