Registration for the 2010 edition of Norway’s Expert Center for Information Management (ECIM) E&P information management conference in Haugesund this month topped out at over 330 making ECIM the largest upstream data conference in Europe. International (non-Norwegian) participants made up 45% of the attendees.
Focus was on Statoil’s migration from its legacy OpenWorks/GeoFrame data environment to Halliburton/Landmark’s R5000 data infrastructure. Dynamic Consulting’s Egil Helland outlined the move to ‘one platform to manage it all.’ Since the 2007 Statoil Hydro merger the number of applications has been halved. But until last year, Statoil was still running Halliburton and Schlumberger’s data infrastructures—a state of affairs that was deemed ‘costly, complex and inefficient.’ Moving to R5000 will save Statoil around 30 NOK ($5) million/year through lower maintenance, support and data loading costs. Various reorganization scenarios were envisaged such as a (costly) storage upgrade, demobilizing data to ‘non G&G’ storage at half the cost, and data deletion. To date, eight projects have been upgraded and some 600 more are in various stages of decision taking and preparation. Cross framework migration is tricky as data models really are different. Work done for one asset may not be applicable to another where things are done differently.
In a joint presentation, Catherine Gunnesdal and Wim Ahuis offered more details on, respectively, Statoil’s and Shell’s R5000 implementation. Statoil reports that while the R5000 database works ‘satisfactorily,’ currently, PowerView is not stable and performant. SeisWorks is poorly integrated to R5000 and it is ‘time to phase it out.’ The move to R5000 requires major changes to exploration workflows. Shell reported limited use of Landmark applications, but the move to R5000 was an opportunity for a major data cleanup exercise.
Speaking through an interpreter, Xun Ma outlined PetroChina’s ‘multi year, multi million’ upstream IT and data management revamp. This involved a move from scattered, siloed systems to centralized systems by domain, sharing a common earth model. Following a successful pilot at the Daqing oilfield business unit, Landmark’s Information Management & Infrastructure consultants rolled out an enterprise-wide information management system at 14 sites throughout China. The solution included Landmark’s Engineering Data model, extended with input from PPDM to include geoscience and production data. To date, 210 TB of seismics and 230,000 wells have been loaded to the system. A PowerExplorer front end gives access to the whole of China’s basin-wide geological and reservoir data set for modeling and future exploration. PetroChina’s vision is for more automated data capture to a ‘digital oilfield, digital basin and a digital PetroChina.’
Chris Bradley (IPL and author of ‘Data Modeling for the Business’) described an attempt to deploy a consistent, enterprise information architecture (EIA) at BP, where a ‘vertical silo-based approach dominates’ and ‘the accountants rule.’ The application environment consisted of many different SAP deployments, some 5,000 ‘other’ applications and a ‘huge’ Microsoft Office SharePoint environment. The move to a service architecture mandated XML model management (with ER/Studio) and a set of consistent definitions in the EIA. These were based on PPDM and used a ‘data virtualization’ approach. Bradley bemoaned the fact that local ‘empowerment’ meant that many don’t follow corporate guidelines. Some see data modeling in a bad light—so subterfuge was required, with ‘data models’ called a ‘business glossary,’ ‘data dictionary,’ etc.. A data Wiki proved popular with folks keen to provide feedback on definitions.
Bradley railed against the evils of data ‘mine-ing’ and data ‘ours-ing’ advocating a move from the hoarding mentality. Other key tools in the EIA effort include DataFlux (quality), Composite Software (integration), Business Objects and Kalido. Prior to 2006, anarchy ruled—anyone could do anything, projects created their own models and definitions. Today BP has a global ER/Studio license with 300 users and a models repository available through SharePoint. BP is engaged in a quality ‘push,’ there is no more ‘make a new model for every project.’ But there is a ‘constant battle with the bean counters.’
Geographical information systems (GIS) continues to interest the ECIM community—with a full track devoted to GIS ending with a new GIS User Forum and Workshop. Andrea Le Pard (Nexen) advocated the use of GIS as a tool for corporate data management with ESRI’s ArcGIS as ‘common industry-standard platform.’ Data suppliers, vendors and government all supply ‘GIS ready’ data. But this is incompatible with Petrel, Landmark and Petrosys, which ‘has its own non ESRI GIS.’ Nexen uses OpenSpirit to bridge the GIS gap—particularly for non spatially aware applications like Petrel. Le Pard made a plea for Petrosys to ‘get together with OpenSpirit.’ On the downside, Le Pard noted that ‘GIS is 80% about data management and only 20% about playing with the data.’
Grahame Blakey and Bernie South (ExxonMobil) gave an insightful presentation on the evolution of GIS over the past couple of decades (in 1989, Exxon was ESRI’s 26th client!). By way of a caveat to GIS enthusiasts, Blakey put things into context with the fact that ‘true enterprise integration is SAP—50% of ExxonMobil’s IT support is around SAP.’ Geologist South described geology as a ‘4D problem, not just maps but time.’ The challenge is to synthesize all regional geological disciplines and here, GIS is a big hitter. Exxon has managed to standardize on 13 INFO tables used for all ARC-based geoscience data sets. Windows migration means a move ‘from Perl and vi to Python, Silverlight and .NET.’ But the new programming tools have forced a wedge between G&G and the programmer. South found life easier back in the day of AML. Operating systems were more stable than on todays’ Windows systems, ‘Unix was a rock, a chuck of granite!’ Performance in Windows is another headache. Database joins and relates are less robust today. There are benefits in the modern GIS in terms of usability and a reduced requirement for hands-on data management. Visualization is hugely improved but Exxon warns of the danger of form over substance. Today it is too easy to ‘swallow’ bad data. More from www.ecim.no.
This article is abstract of a longer Technology Report produced by Oil IT Journal’s publisher, The Data Room. More from www.oilit.com/tech and tw@oilit.com.
© Oil IT Journal - all rights reserved.