Around 600 attended the Society of Petroleum Engineers’ (SPE) Digital Energy Conference in Houston last month. Shell Chief Scientist Charlie Williams’ keynote traced the history of ‘smart fields.’ These were initially more about communications than IT, with microwave links from offshore platforms to the Shell building in New Orleans. Early experiments in the mid 1970s with computer assisted operations equipment were abandoned and it was not until much later that the smart fields (SF) concept got traction. SF drivers include deep complex reservoirs, secondary recovery and the need for energy efficiency. Communications have made a lot of progress, but Katrina showed the limits of today’s infrastructure.
Williams enumerated Shell’s SF technologies including smart well control valves and monitoring with permanent downhole gauges, flowmeters and distributed temperature sensors. Getting data to the surface involves a tortuous path through packers, tubing hangers and the well head and into the surface control system. But these techniques have enabled ‘smart snake wells’ to produce from Shell’s Champion West field in Brunei, long considered un-developable because of its stacked, faulted reservoirs. Champion West now contributes 25% of Brunei Shell’s production from wells 8 km long with 4km in the reservoir. A proof of concept on another Brunei field, Iron Duke, has involved ‘retrofitting’ smarts. This resulted in a 15% production hike and a two year reprieve on water breakthrough.
Looking to the future, Williams sees computer assisted smart fields, proactively managed as a single dynamic system. ‘Discrete’ technologies are important enablers but it is the holistic management that is key. SF screening is important; Shell has a SF opportunities framing methodology. Williams warns of systems designed by engineers and management, but that operations can’t handle. SF is not about technology but more about closing the value loop with technology used by and embedded within people, process and tools.
Randy Krotowski (Chevron CIO) thinks the digital oilfield is an idea whose time has come. Most of the challenges have been solved, with 4D seismic, visualization and decision support centers. What remains are the ‘people’ challenges—convincing them that this is a good idea and especially, getting operations and engineers on board. Chevron has iField engineers who are optimizing operations, data architectures, and making it easier to introduce new technologies. Chevron is leveraging the PPDM data model and ‘PPDM XML web services’ in its SEER data warehouse (OITJ April 07). Ricardo Beltrao described Petrobras’ in-house developed ‘GeDIg’ system for integrated digital field management. Petrobras uses its own software to optimize operations on the 2,000 well Alto do Rodrigues offshore heavy oil pilot. The Barracuda-Caratinga collaborative offshore control center for digital operations (CGEDIG) is located in Rio, some 450km distant. Iraj Ershaghi, who works at the Viterbi School of Engineering at USC, commented on the ‘IT-ization’ of petroleum engineering. A lot of today’s buzz words and technology are not what engineers learn at school. Today we need ‘renaissance’ engineers with IT knowledge. The question then arises, do we train IT specialists in engineering or vice versa? Ershaghi answered with a medical analogy—where the same problem exists. Would you prefer to be operated on by an IT tech who had retrained in surgery, or by a surgeon trained to operate sophisticated equipment? Ershaghi also suggested that the SPE do more to communicate the fact that its technology is cutting edge. We need to sell better at universities where there is a huge problem with the low number of petroleum engineering students. The Chevron USC CiSoft’s MS in smart oilfield technologies is part of the answer. In the Q&A, one perspicacious observer contrasted Ershaghi’s ‘cutting edge’ description of oil country technology with the SPE’s recent self-flagellation as an industry characterized as a technology laggard!
Mike Strathman’s suggested that we should look again at time-based (as opposed to depth-based) drilling. The idea is to have a unified view of all real time data. By using a data historian, as deployed in the process control industry, all real time data is collected in one place. This allows for detailed analysis of current situations in the light of historical data, leveraging drillers’ expertise and allowing for a quick response to operational issues. The data Historian is much more performant than an RDBMS. Time based information supports queries along the lines of ‘what else was going on at that time?,’ ‘How long have we been drilling in in this formation etc.’ AspenTech’s solution in this space is the InfoPlus.21/Web.21 combo of Historian and analytics. A new ‘WellTrends’ package displays log data and allows drag and drop of data streams to log tracks.
Mike Weber (BP North America) explained how its Arkoma Basin unit was co-visualizing Pipesim and SCADA data in a Microsoft Virtual Earth-based solution from IDV Solutions (OITJ February 07). Phase I of the project resulted in map-based visualization of Pipesim model data. Phase II extends this to data from OSIsoft’s PI Historian over the 520 wells in the Red Oak. Microsoft SharePoint services and Web Parts also ran as did Schlumberger’s Avocet production data server. The plan is to offer ‘evergreen’ models, to perform ‘what if’ operating scenarios and in general, to ‘break down silo boundaries’ with visualization. A movie showed VirtualEarth with check boxes turning map layers on and off and bubble maps of production. Data can be exported to Excel for further analysis. An animated pipeline display warns operators when pigging needs to be done. GIS data is ‘piggy-backed’ on the Virtual Earth data server, limiting data and bandwidth requirements on site.
Schlumberger Fellow Bertrand du Castel gave a suitably erudite and rather obscure presentation on the application of semantic web technology to homogenize security across various digital oilfield subsystems. The Internet, SCADA systems, and ISO 27001 all expose different security models. What is missing is a common security ‘ontology,’ a problem that extends across other facets of ‘remoting’ operations, real time, automation and ‘augmentation.’ Citing Thomas Sheridan’s book, ‘Humans and Automation*,’ du Castel described the ‘prize’ as ‘goal-oriented, distributed workflows’ enabled by ontology, process model, interaction and ‘interstriction’ models. Baysian logic leveraging Norsys’ Netica also ran. All will interact through a services-oriented, semantic web, OWL and UML and BEPL. Du Castel showed pilot smart fields where composite business process modeling has been deployed.
Peter Oyewole’s presentation covered an ‘intelligent, closed loop integrated digital system’ as deployed on BP’s San Juan Basin prolific coal bed methane field. The system manages tubing flow control, plunger lift and other artificial lift systems and has resulted in increased gas production, better equipment reliability and in the development of an efficient cheap deliquification process. Remote operations connect RTUs to Invensys’ Industrial Application Server. Data is fed to Maximo for work order generation, E-Choke (for choke analysis and FDA (field data analysis). Multiple SCADA interfaces are integrated through an abstraction layer to Modbus. The system allows BP to switch to and from plunger and artificial lift as appropriate and to control tubing flow. Condition-based soap injection and other remedial actions are also enabled.
Operations Support Center
Schlumberger was showing its Operations Support Center—a crossover solution combining Petrel’s new real time capability with the wireline division’s Interact Server, Perform toolkit and real time logs. A compelling mock-up showed a twin head display with a portrait mode screen for the log/real time data and a landscape view of the Petrel model.
* John Wiley, 2002.
This article is a summary of a longer, illustrated report produced as a part of The Data Room’s Technology Watch Service. For more information please email email@example.com.
© Oil IT Journal - all rights reserved.