20130910 petronas technical data quality...
TRANSCRIPT
PETRONAS E&P Technical Data Quality Metrics Initiative – Going Green with Data
Hassan B SabirinRegulatory Compliance & Technical Assurance, Technical Data
Hassan B Sabirin, 8 October 2013, Page 1
Regulatory Compliance & Technical Assurance, Technical Data
Petroleum Management Unit (PMU)
PETRONAS E&P
08 October 2013Impiana Hotel, Kuala Lumpur
Digital Energy Conference, Kuala Lumpur:Improving E&P Subsurface Data Management
© 2013 PETROLIAM NASIONAL BERHAD (PETRONAS)
All rights reserved. No part of this document may be reproduced, stored in a retrieval
system or transmitted in any form or by any means (electronic, mechanical,
photocopying, recording or otherwise) without the permission of the copyright
owner.
Contents
• Organization Background• PETRONAS EP Technical Data
• The role of data management in value creation
• What problems are we trying to solve?
• Data quality metrics defined
Hassan B Sabirin, 8 October 2013, Page 2
• Example – Some basic metrics
• Data quality metrics – high level going green roadmap
• The data quality metrics and standards connection
• Concluding remarks
Technical Data was formed to elevate functional excellence and delivery discipline of data management
PETROLEUMMANAGEMENT
PETROLEUMRESOURCE
EXPLORATION
PETROLEUMRESOURCE
DEVELOPMENT
PETROLEUMOPERATIONSMANAGEMENT
STRATEGICPLANNING
FINANCE & ACCOUNTS
PSC SCMGOVERNANCE
HSETECHNICAL
DATA
Hassan B Sabirin, 8 October 2013, Page 3
Knowledge Management
& Institutional Capability
Institutional
Capability
Knowledge
Management
Business Process
Advisory &
Solution
Planning &
Performance
Program
Excellence
Data Facility
Management
Regulatory
Compliance
Technical
Assurance
Geospatial
Geophysical
& Geological
Application
& Workflow
Project Data
Application
& Workflow
Production
Technology
Petrophysics
Reservoir
Engineering
Application
& Workflow
Facility
Engineering
Well Engineering
Process Data
Application
& Workflow
Planning&
Delivery
RegulatoryCompliance
& Technical Assurance
GeoscienceData
Project Data
PetroleumEngineering
Data
Production&
OperationData
Data Quality Metrics
TD
PETRONAS MANAGEMENT
PSCs / RSCs
3rd Parties
Government Depts
Industry Standards Peer
EP
Business
Strategy External Data
Benchmarking
& Request Data
Consultancy
Services
Technical Data Department Relationship Model
PETRONAS
Hassan B Sabirin, 8 October 2013, Page 4
TD
EP IT
Industry Standards Peer
Groups
Technical Apps &
Data Vendors
Consulting Services
ICT &
Infra
Host Authority
Data
Requirement
DM
Services
PETRONAS
EP Business
Units
> 1,000,000 media
> 700,000 technical
>766 active projects
~ 159 ventures in 25
countries
~ Average 100 wells
drilled per year
~ 1,000 2D/3D seismic
surveys
~ 12,000 wells
- All EP data types
>250 specialized
technical applications
(Multiple disciplines &
Measured Data> RM35 Billion investment on
data acquisition
Interpreted (Result) Data
Subsurface understanding which has direct
impact on business decisions
Dimensions of
Background – PETRONAS EP Data Dimensions
Hassan B Sabirin, 8 October 2013, Page 5
> 700,000 technical
docs
> 3,500 TB disk space
>120 data bases
Tools (Technical Software Application)
> RM1.2 Billion investment
(Multiple disciplines &
related application
workflows )
End Users> 400 geoscientists and engineers from all
technical domains in E&P
End Users> 400 geoscientists and engineers from all
technical domains in E&P
Media & Corporate DBsMedia, hardcopies and Geosamples
External StakeholdersPotential investors, PSCs, government
agencies, universities
External StakeholdersPotential investors, PSCs, government
agencies, universities
Dimensions of
PETRONAS EP
Data
Management
Workflow and Standardization
Project
Database
ApplicationSource / feeder
databases
Master (Reference) Data
Data Quality Metrics
Corporate
Database
Publishing
Previous studies : Geologists spend up to 60% of their time looking for & quality checking data for their projects
Re-useValueLoop
Corporate Data
2 2
3
3
unique master sourceof quality checked dataunique master sourceof quality checked data
Role of Data Management in Value Creation
Hassan B Sabirin, 8 October 2013, Page 6
Data
collation
Project set up
Database Database
Publishing
Re-use
ValueLoop
Cost management
Process improvement
Optimization
Reduce the 60% to __%
Standards & workflowsBring to acceptable qualityCapture work done for re-use
‘One version of truth’ datasetsRe-use many times
1
1
1
2
2
3
3
Approved & final Approved & final
Feeding new projectsFeeding new projects
The project space is where the big decisions are made
Reduction of time for data collation and project setup
Step 1
Step 2
Step 3
What Problems Are We Trying to Solve?
1- Mitigate business risks created by data
quality issues
• Financial
• Confidence/Satisfaction
• Productivity
Hassan B Sabirin, 8 October 2013, Page 7
• Productivity
• Compliance/Legal
• Safety
What Problems Are We Trying to Solve?
2 - Value leakage in business
• Lost time due to re-checking of data
• Not re-using data because of trust issues
• Lack of awareness & enforcement on standards –
inconsistent business practice
Hassan B Sabirin, 8 October 2013, Page 8
inconsistent business practice
• You can’t manage what you can’t measure – quantification of
leakage
What Problems Are We Trying to Solve?
3 - Managing change introduced by new technology
New data content may introduce a requirement to improve
• Data standards/guideline
• Data integration process
Hassan B Sabirin, 8 October 2013, Page 9
• Data integration process
• Data analytics method
Data Quality Metrics Defined
• Data Quality Metrics: A measure of data quality performance
• Data Quality Dimensions: Perspectives from which data quality is being looked at
Hassan B Sabirin, 8 October 2013, Page 10
• Business Rules: Rules that must be adhered to in order to keep the business running in good/optimum condition
Project A
Data Quality Metrics Defined
Completeness
Validity
Deviation Survey
Data Type Quality Dimension Metrics Target
Data Quality Metric
Hassan B Sabirin, 8 October 2013, Page 11
Validity
Uniqueness
Conformity
Consistency
Integrity
Database B
Business Rules:
Well/wellbores with a
completion date 3
months back from
today must have a
definitive/final
deviation survey
Data Quality Metric
Total Error Total Records
)%1 – (
Metrics
Well measured depth must be more or equal than true vertical depth
Offshore well must have a water depth information
A well deviation header must have end/bottom depth of the survey defined
A well deviation header must have start/top depth of the survey defined
A well deviation header must have the survey contractor/company performing the job defined
A well deviation header must have the survey tool name defined
Borehole survey
depth range
Borehole
Examples of Data Quality Metrics
Well Header
Hassan B Sabirin, 8 October 2013, Page 12
A well deviation header must have the survey tool name defined
A well deviation survey point must be identifiable to a deviation survey header
A well deviation survey point must have the calculated dog leg defined
A well deviation survey point must have the calculated vertical section defined
A well deviation survey point must record the measured azimuth
A well deviation survey point must record the measured inclination
Borehole
survey tool
identity
Borehole
survey
measurements
Data Quality Metrics Aggregation
Metric/QuerySurvey must have at
least azimuth,
inclination, MD and
TVD Metric/Query Well Deviation Survey
Metric
Target: Drilling
Database
Borehole Survey
Business RulesAggregation
Hassan B Sabirin, 8 October 2013, Page 13
TVD Metric/Query
Metric/QuerySurvey dogleg must
not be more than 10
degrees Metric/Query
Metric
Queries, Query Sets, Targets & Aggregation
Single Query
Single Query
Single Query
Single Query
Query Set
OPENWORKSPROJECTOPENWORKS
PROJECT
OPENWORKSPROJECT
OPENWORKS
PROJECT
OTHER
DataBasesOTHER
DataBasesOTHER
DataBases
Single Query
Single Query
Single Query
Single Query
Query Set
OPENWORKSPROJECTOPENWORKS
PROJECT
OPENWORKSPROJECT
OPENWORKS
PROJECT
OTHER
DataBasesOTHER
DataBasesOTHER
DataBases
Single Query
Single Query
Single Query
Single Query
Query Set
OPENWORKSPROJECTOPENWORKS
PROJECT
OPENWORKSPROJECT
OPENWORKS
PROJECT
OTHER
DataBasesOTHER
DataBasesOTHER
DataBases
Multiple Query Multiple Query Multiple Query Multiple Query
SetsSetsSetsSets
Multiple
Target SetsSingle
Metric
A G
G R
E G
A T
I O
N MULTIPLE QUERY SETS MULTIPLE TARGET SETS
Hassan B Sabirin, 8 October 2013, Page 14 14
Single Query
eg. select well_name from well_master;
PROJECT
DATABASE
TARGETTARGETTARGETTARGET
Single Query
Single Query
Single Query
Single Query
Query
Set
PROJECT
DATABASEPROJECT
DATABASE
PROJECT
DATABASE
PROJECT
DATABASE
OTHER
DataBasesOTHER
DataBasesOTHER
DataBases
MULTIPLE QUERIES MULTIPLE TARGETS
Single
Metric
Single
Metric
A G
G R
E G
A T
I O
N
Data Quality Metrics Aggregation to Dashboard
A B C D E F etc.Regions
of Interest
The (Enterprise) Dashboard
Bu
sin
ess F
ocu
s
The (Enterprise) Data Quality Trend
Hassan B Sabirin, 8 October 2013, Page 15 15
Databases
of Interest
Data
of Interest
Data Element
of Interest
Assets
of InterestBu
sin
ess F
ocu
sD
ata
Fo
cu
s
Data Quality Metrics Dashboard
Metric
A B
Metric1
2
3
Metric
Metric
Metric
Metric
Metric
Metric
Metric
C
Hassan B Sabirin, 8 October 2013, Page 16
3 Metric Metric Metric
GoodMediumPoor
95-10090-950-90 Total Error Total Records
)%1 – (
Data Quality Metrics Dashboard
Hassan B Sabirin, 8 October 2013, Page 17
Data Quality Metrics Tool
Hassan B Sabirin, 8 October 2013, Page 18
Data Quality Metrics Dashboard – By Drilling Rig
Hassan B Sabirin, 8 October 2013, Page 19
The Data Quality Metrics – Data Standards Connection
Business Rules Data Standards
Data Quality Metrics
measure adherence to
Hassan B Sabirin, 8 October 2013, Page 20
measure adherence to
data standards
Metric
A B
Metric1
2
3
Metric
Metric
Metric
Metric
Metric
Metric
Metric
C
Data Quality Metrics- High Level Roadmap
2013 2014 2015
Infrastructure
ArchitectureArchitecture
Data Quality Focal Person
Da
ta Q
ua
lity
Me
tric
sD
ata
Qu
ali
ty M
etr
ics
Top 4 data types
Corrective Action
Measure Validity,
Conformity
Measure Validity,
Conformity
Measure Completeness Measure Spatial Layers
Top 8 data types
Hassan B Sabirin, 8 October 2013, Page 21
Data Ownership – Roles & Responsibilities
Da
ta Q
ua
lity
Me
tric
sD
ata
Qu
ali
ty M
etr
ics
Oth
er
Are
as
Oth
er
Are
as
Data Standards
Business Rules
Management Support
Project Quality Metrics
Conclusion
•Data Quality Metrics is an important part of the data
management portfolio.
•A data quality metrics program needs to closely
engage business lines• Business must play a role to define the business rules that adds
value to the data.
• Data Ownership – roles and responsibilities must be clearly
Hassan B Sabirin, 8 October 2013, Page 22
• Data Ownership – roles and responsibilities must be clearly
defined
• Proactive participation at the working level, and drive from the top
•Data Quality Metrics results have to tie to business owners
as a measure of performance
Hassan B Sabirin, 8 October 2013, Page 23
Thank You
Paper Abstract
People, process & technology are often touted as the 3 important components an organization must have in order to ensure that the data is managed properly and effectively. While that is true to a large degree, the existence of this tripartite in
the organization does not necessarily support the sustainability of good data management practices over the longer term. Data quality today still remains as a
significant challenge we face in the E&P industry.
For the data we use within upstream E&P, data quality metrics, when implemented
Hassan B Sabirin, 8 October 2013, Page 24
For the data we use within upstream E&P, data quality metrics, when implemented in the right databases, creates a transparency that cannot be easily achieved by any other means, and provides us with a powerful tool to highlight issues before the problem becomes too big. A proper data quality framework also address data issues in holistic way and combines with data ownerships and accountabilities to
ensure the right interventions and corrections get done.
This paper describes PETRONAS E&P’s data quality metrics initiative that will progressively improve the quality of data used by the business and result in a
traffic-light format data quality dashboard that gets more and more green.
Philip Lesslar, 20thMay 2013, Page 2