UK-based Flare Solutions has released V2.0 of its E&P Catalog, in a ‘complete redesign’ of the product following analysis of usage patterns over the product’s four years of existence. The Catalog now holds over 100,000 terms and relationships. The Catalog is made up of a ‘taxonomy,’ a list of terms used in the upstream and an ‘ontology,’ hierarchical relationships between the terms. The taxonomy/ontology combo can be licensed in a variety of ways. EPCat is a ‘shrink-wrapped’ web-based search and publishing tool for electronic documents, hardcopy and data. An EPCat-Hub version offers a web service for company naming standards and pick lists. These can be installed inside the firewall or hosted as a service by Flare. The Flare taxonomy and ontology can be licensed separately and a web service API has been provided to third parties including Schlumberger, Halliburton, Enigma and the OVID subscription service. These companies use the API to push content to the Catalog from their systems and keep it up to date.
Flare’s Paul Cleverley told Oil IT Journal, ‘The Catalog assures consistent, enterprise scale tagging of documents and data with a standard set of E&P keywords. The system can then lead the user to the right information or suggest related information topics without the need for a fully qualified query. The Catalog complements Google-type full text search which is often high in recall, but low in precision. Our technology integrates disparate sources of meta-information into a single semantic layer.’
One way of visualizing Flare’s taxonomies is as a ‘hyperbolic tree.’ The public domain source code for this representation was originally developed by Toulouse University and it is now released under the GNU General Public License. A patent application that restricted commercial use of the tool in the US was lifted in February this year when Xerox dropped its claim. The graphical interface allows users to navigate the ontology, ‘pruning’ the branches for fine-tuning a search. Flare contrasts the Catalog with geographical search tools such as MetaCarta which display search results on a map. The Catalog displays search results within the E&P domain itself, in both text and graphical form.
Updates to the Catalog are supplied in the W3C’s ‘semantic web’ resource description format (RDF) making Flare an early upstream adopter of the technology. But Flare’s real value add is in its content rather than a particular delivery mechanism. The Catalog’s detailed E&P subsurface model differs from the knowledge maps used by other tools. According to Cleverley, ‘Search tools, despite all their statistical algorithms and knowledge maps, simply do not (and will never) understand, that ‘well test’ is always related to ‘reservoir pressure,’ or that a ‘limestone’ is also a ‘carbonate.’ They don’t understand that if you search on ‘petrophysical information’ for a well, the results should include ‘density logs.’ Our detailed ‘knowledge map’ is the differentiator. This also helps disambiguate non-unique terms.’
A subset of the taxonomy is sent by Flare to Energistics each year, although, according to Flare, the potential of a true global standard in this area is so far un-realized. Flare is currently working to see how it can help remedy this and is investigating how best to continue to expand and release this taxonomy and ontology into the industry while retaining ‘some element of intellectual property’ on its own work. Early versions of the E&P Catalog were a joint development with Shell (OITJ December 2002). Flare’s clients include Shell, BP, GDF Suez, Nexen and BG Group. Flare received a Queen’s Award for Enterprise Innovation this month. More from p.cleverley@flare-solutions.com.
This article originally appeared in Oil IT Journal 2009 Issue # 4.
For more information or to comment on this topic email here.