Torkel Thime from the Norwegian State Archive (Statsarkivet) described a program to capture documents relating to the early days of Norwegian oil and gas exploration. Oil and gas is a crucial business to Norway and it is important to understand how it developed. The Statsarkivet has signed deals with Statoil, ConocoPhillips, Total, Exxon, OLF and others for the archival of documents from the 1970s relating to the discovery and development of Norway’s major oilfields like Ekofisk and Frigg.
The real story
Documents include scanned memoranda, handwritten notes and video archives. These include ‘behind the scenes,’ internal non public records that ‘contain the real story.’ They show for instance the unpublished negotiations on Ekofisk crude prices and tax discussions between Phillips and the Norwegian government. Management Committee minutes show Bartlesville memoranda on the Ekofisk Bravo blowout. Discussions with trades unions are also captured. According to Thime, ‘There is no truth here, only different views, eyewitness views bring us very close to events.’
A data migration strategy assures a move from original format documents, removing software dependency. Confidentiality issues are subject to agreements with companies as to when documents will be publicly available.
Lars Olav Grøvik (Hydro) believes there is a ‘mismatch’ between what we spend acquiring data and what is spent on managing it. This is a significant issue that ‘could turn the dream of the digital oilfield into a nightmare.’ Hydro’s operations control center (OCC) at Sandsil was the best in class when built. The OCC was replicated on the West Venture rig offshore. But because Hydro ‘omitted’ to change the underlying work process, the CAVE visualization system was underused. Grovik still believes in 3D visualization but this has largely failed because of the difficulty of keeping CAVE data current. Hydro has tried to get real time data into the CAVE but ‘software is not ready to consume field data.’
Wired pipe has given a huge increase in data rates and broadband from the bit represents a huge challenge for data managers. ‘Why is our software not real time ready?’ ‘Why do we have to hit a button to refresh the screen?’ A middleware integration layer can help with ‘plug and play’. Hydro was reluctantly ‘forced’ to use OpenSpirit to link its local, project and master data stores. The human factor and change management are critical—especially when changing software while operations are up and running! Grøvik concluded with a reference to ‘Putt’s Law’: Those who understand technology don’t manage it, and those who manage don’t understand technology!
XML in oil and gas
Bjorn Rugland (Statoil) traced the complex history of XML in the Norwegian petroleum industry. The Daily Production Reporting (DPR) standard has backing from OLF, Petoro and the main Norwegian operators. DPR embeds the ISO 15926 ontology, Tieto Enator’s Energy Components and an XML data base from LicenseWeb. Petoro uses DPR for partner surveillance across 50 plus fields. A related standard for Monthly Production Reporting solves reallocation issues that were addressed in DPR. MR requires correct volumes aligned with commercial systems. ProdML’s initial scope was the ‘office’ domain, upstream of historian. But in the future, ProdML scope may embrace process data. Statoil is testing ProdML with SIS on zonal allocation on Snorre.
Open Standards in E&P
Nils Hofseth works on system integration at Capgemini’s (CG) ‘EpiCentre,’ an E&P industry competency center which opened earlier this year to leverage CG projects in the upstream/midstream. Hofseth traced the history of integration efforts across vendor applications, data models and APIs. Today, shared data models including OpenSpirit and WITSML alleviate the fundamental problems with reusable programming interfaces and by minimizing ties to platform and language.
Eric Toogood (Norwegian Petroleum Directorate—NPD) offered a brief historical review from the early agreement between the NPD and oils to cooperate on data provision, focusing initially on post stack seismics. NPD is also a DISKOS user as it does its own interpretation to negotiate license terms with oils on a level playing field. The NPD also supervises data preservation and manages data release by the simple expedient of changing the entitlement to ‘public’. Seismic data release rules have been clarified and data in relinquished acreage is no longer confidential. DISKOS membership is up from 17 companies last year to 41. Users can now log into the database from anywhere in world through a new web interface. SINAS operations manager Steven Eames reported that there is now around 2TB of well data and nearly 90TB of seismic online and around 3TB is downloaded per month.
Anne-Lill Holme (Contesto) advocates tagging documents with contextual metadata to make it findable and understandable to other users and for version control. Holme deprecates the use of folders to mimic metadata. Storing documents in folders may not be consistent across the company and there are security, version, tampering and erasure issues with such naïve document management. Metadata capture starts with a database of controlled lists for owner, creator, document type, etc. These are presented to users as a form to fill-in before storing a document. Search can then use metadata and free text constrain the hits returned. It is important to think through tag structures to align them with business needs. This is where the Dublin Core set of metadata terms comes in. DC offers common semantics – with a dozen or so elements to choose from. Unfortunately, there are always some attributes that are not suitable for DC. Metadata capture should allow for stuff that can’t be structured this way—an issue that is not recognized by the DC community. To help and encourage users to fill in metadata, tricks include prepopulating fields and color coding documents.
According to Stephan Setterdahl (ExxonMobil), geologists are bad at databases and their DBAs are ‘singing the blues.’ Setterdahl suggests starting with an empty database and moving key data into the new repository. What should the be? A big corporate database with a 1,000 page manual is ‘a disaster,’ and an open database gets too cluttered after a couple years of use. Setterdahl confessed he was ‘tired of cleaning databases’ and decided that there had to be a better way—without sacrifying geologists’ creativity. He got together a geologist and script writer (SQL and Perl) and started developing a set of scripts to identify database inconsistencies and notify users. Next, a unique ‘pick name’ is required for the plethora of synonyms ‘Cret.,’ ‘Cret,’ ‘Kret,’ ‘SIS_Cret’ etc. Because there are so many, new users are forced to add their own to the mess! Scripts find and rename rogue names to a standard taxonomy. Other tools map from BP, Statoil, and other pick conventions to Exxon’s A17 Maximum Flooding Surface (MFS) convention. ‘Police Action Scripts’ control the database and prevent a mess before it happens. If a user diverges from the standard, the system gives the user one month to correct or their data is deleted! A debate on ‘database blues’ revealed similar issues for perforation data, faults, seismics and well bores. But fly in the ointment is Petrel. As Setterdahl says, ‘You can’t run scripts on Petrel data, I’m hosed!’
Otis Stelly outlined Shell’s move from semi-autonomous OPCOS to a ‘global’ EP IM discipline with central control. This has resulted in a Global IM Forum and an EP CIO. IM design is in response to business drivers like productivity and reporting. As an example, Shell’s ‘Focused Basins Information Platform’ uses global document resources to collate information on an area of interest, leveraging Metacarta. The Shell EP Catalog is now extending to embrace field development and to ‘put workflows atop of IM.’ Shell uses the ISO 15926 standard for its facilities where up to 300 contractors may work on a project.
Liv Maeland (Statoil) announced a new university study program designed for professionals in oil and gas information management. The ECIM-managed course starts later this year at the University of Stavanger. A second ‘IM value creation’ conference is planned for 2009.
This report is a short version of a Technology Watch report from The Data Room—more from firstname.lastname@example.org.
© Oil IT Journal - all rights reserved.