The 2005 Houston meet of the Public Petroleum Data Model (PPDM) Association witnessed a modest return of the corporate data model with deployment reports from Nexen, Woodside and Anadarko. But it is in the service sector that the data modeling landscape is being re-defined, with PPDM being the solution of choice for pretty well all new development. PPDM CEO Trudy Curtis is building on this success by hooking up with the Pipeline Open Data Standard (PODS) organization, with plans to share data infrastructure and modeling best practices.
Petris CTO Jeff Pferd showed how the loose coupling and service-oriented architecture favored by his company can be used to avoid the dependencies and versioning issues encountered when upgrading, say from PPDM 3.6 to 3.7. Pferd called for the standards bodies to create web services interfaces to their data models, documenting recommended granularity and providing XML data question and response.
Kenneth Greer (CenterPoint Energy) described PODS as a ‘software and vendor-independent’ database for pipeline and location data. Work is in progress on a GPS data dictionary, rights of way, inline inspection, compliance and documentation. Greer believes that in the future, PODS and PPDM may share data infrastructure where there is overlap—e.g. for geodetics, units of measure and partnerships. Tracy Thorliefsen (Eagle Information Mapping) elaborated further on the PODS data model showing how pipelines are similar to wellbores to the data modeler. PODS has no spatial capability ‘out of the box’ and is therefore ‘GIS-neutral’. Spatialization options are a trade-off between loose-coupled systems and the tight coupling as described in the ESRI manual. This has unwanted side-effects and ‘creates havoc for interoperability at the enterprise level.’ Loose coupling links GIS features and classes to PODS via foreign keys. The database becomes GIS-independent albeit with a performance and maintenance penalty. The PODS model lends itself to network modeling and engineering studies using tools like PipePhase.
In a video presentation, Gwen Kelly showed how PPDM has been coupled to Woodside’s SAP ERP system. Woodside’s AFE* Project tracks expenditure from business proposal to financial settlement for items like G&G studies, seismic surveys and wells. Each project contains both financial and technical data, requiring a joint finance and production management approach. The project links permit information in PPDM with joint venture finances in SAP. The development was needed because a) ‘SAP is not very good at detailed project management,’ b) ‘It is hard to make a neat customized form in SAP’ and c) ‘There’s no word for ‘fluffy’ in German!’ PPDM is good at managing the complexity of joint ventures and farmouts. SAP’s Business Warehouse was used to produce permit reports and an ArcMap interface shows AFE spend as color coded map.
Trudy Curtis reported good take up for PPDM’s latest 3.7 version which now has 1,200 tables and 24,000 columns. The model now runs on Oracle, SQL Server, MySQL and PostGres. Work is in progress on a PPDM metamodel and on reference values. These should include data provenance and units of measure. EPSG geodetics are now embedded in PPDM. The next version will leverage PPDM’s collaboration with POSC on XML components and schema, although some of this work is ‘on hold’ pending WITSML developments. An EnCana-funded project will align the PPDM schema with GML (see below). Work is also ongoing in the field of taxonomy with arbitration between UNSPSC and NASA code sets and a ‘confusing and incomplete’ POSC Discovery EPICAT.
Dave Buggraf (Galdos Systems) presented a backgrounder on Geography Markup Language (GML), XML for web-based geospatial information. GML has a constellation of related standards—ISO 19199 GML Web Feature Server (WFS), XMML (mining), PPDM GML, O&M (observation and measurement) etc. The basic idea is that a browser can access data in any geo-database via a web feature service. A distributed geo infrastructure will allow land, pipeline, hydrography etc. to be consolidated from multiple databases. Raster, contour, digital terrain models and are handled by attaching an external binary file in JPEG 2000-based JPX ‘package’. Bugraff described JPX as ‘GeoTiff on steroids.’
John Jacobs described how Anadarko used to manage data flows throughout the enterprise with ‘bubble gum and bailing wire solutions’ written in Perl and Unix shell scripts. These have been replaced with Informatica’s extract transform and load (ETL) tools. Metadata mapping has allowed integration of data from Anadarko’s ERP system with its in-house PPDM corporate data base. Integrating production data from external sources has been hard to achieve, particularly with 3.5 million wells in its domestic database. PPDM proved a ‘clear and easy winner’ for well data. Documentum and ESRI SDE also ran—leveraging a standard taxonomy. Anadarko’s system has been developed in-house ‘to save cost’. Although Anadarko is in general a ‘buy not build’ company, business rules and enhancements ‘need to be done by people in the organization.’
Kim Thomas (ExxonMobil and PPDM Board) described the industry as ‘at a watershed,’ with an average age of 48 years—even older in the data management sector. Turning to data management, Thomas stated that in ExxonMobil, database use is being driven by reporting requirements. So the more standards bodies work on common issues, the better for everyone. In some areas though, there are plenty of standards already. Units of measure can be managed in IEEE, ANSI, POSC, API, Mathematica and NIST. ExxonMobil is ‘in transition’ regarding units of measure and is planning to select one of the above and share best practices with the industry. Thomas noted that such openness and sharing reflects a change in ExxonMobil’s outlook and ‘would not have happened a few years ago.’
* Authorization for expenditure.
© Oil IT Journal - all rights reserved.