Speaking at the SMi E&P Data Management conference in London earlier this year, Paul Duller (Tribal) and Alison North (AN Information)
provided a salutary tale of the importance of information management.
In 2010, a natural gas pipeline operated by Pacific Gas & Electric
exploded in San Bruno, a suburb of San Francisco causing eight deaths
and considerable property damage.
The US
National Transportation Safety Board determined that inadequate quality
assurance in a 1956 line relocation project was a likely cause, along
with inadequate maintenance. North was called in as an expert witness
because of her experience with old paper records. Her investigations
showed that the GIS system that PG&E used for its integrity
management program contained ‘inadequate and misleading information.’
The paper records showed that what the GIS system reported as
‘seamless’ in fact had a longitudinal seam that should have had a more
stringent maintenance program. North emphasised the risks that
companies run from such ‘dark data’ of uncertain provenance that can
show up in a legal discovery process. Similar risks are inherent to
other corporate data sources such as email.
Robert Best (Petroweb)
asked ‘why standardize a standard?’ It turns out that there are good
reasons to want to standardize a Ppdm implementation and Petroweb,
along with OpenSpirit, TGS and OilWare are working to develop tools
(data load, export, triggers etc.) and techniques to make for a robust
Ppdm deployment. While the vanilla Ppdm provides the data model and
some guidelines, it does not say how primary keys are defined, how well
coordinates should be stored or how CRSs are defined. While the Ppdm
AREA can be used to define a hierarchy of state, county (or country,
block), this is hard to implement. Ppdm is kicking off a work group on
implementation standards that is to develop a reference implementation.
ENI’s
Paul Richter likened data management to cleaning up a messy child’s
room. No sooner has it been done than you are back to square one. Enter
data ‘utopia’ where things stay nice and tidy all the time. This can be
done with a ‘robust modular framework’ that fits the business’ needs.
First, identify your master database and develop a method to reconcile
CRS/UOM across different data stores and applications. Then develop a
strategy for reconciliation and data maintenance. ENI is working with
One Virtual Source (OVS) on such a system, using data quality metrics
to trigger clean-up processes. Richter is to present more on this
project at next month’s PNEC.
KOC’s
Hussain Zaid Al Ajmi described a similar approach using ‘front end’
integration and data QC project in multi vendor environment. Dispersed
data has been cleansed and ‘stringent’ naming and UoM/CRS conventions
applied and all copied to a new OpenWorks master project database a.k.a
KOC’s single source of truth. The plan is now to promote the master to
a ‘central authenticated data bank’ and add automated workflows and
processes. BP’s Project Chile is reported as taking a similar approach.
© Oil IT Journal - all rights reserved.