Towards a Semantic Information Platform for Subsea Petroleum Processes

by Jon Atle Gulla

The Integrated Information Platform project is defining an international semantic standard – or ontology – that will assist oil companies in making decisions and organizing collaborations. This article describes how the standard is developed and how it is used in intelligent information retrieval and reasoning.

The subsea petroleum industry is a technically challenging business with complex projects and operational structures. The projects are expensive, and they often include several large companies and span disciplines like drilling, reservoir, production, operations and maintenance. The European petroleum business is now facing a number of challenges that threaten its profitability. The costs of large older fields increase as they enter the decline phase, and new fields tend to be smaller and less scalable. We also produce more oil and gas than is added through exploration and improved oil recovery. For the Norwegian Continental Shelf for example, the addition and production numbers are at about 100 million Sm3 and 250 million Sm3 respectively. Finally, there has been an increase in the number of small and highly specialized service companies that need to collaborate closely both with each other and with the traditional bigger companies. All of this suggests that future petroleum projects need to be more cost-effective, make better use of small fields, and take advantage of the distributed and specialized skills of new service companies.

In 2004 the Norwegian Oil Industry Association launched an Integrated Operations program that proposed the use of new information and communication technology to integrate processes onshore and offshore. OLF's own estimates indicated that the implementation of this program on the Norwegian Continental Shelf would increase oil recovery by 3-4%, accelerate production by 5-10%, and lower operational costs by 20-30%. Central to this program was the semantic and uniform treatment of heterogeneous data, which originate from various disciplines and companies, at various locations, and with various degrees of precision and formality.

The Integrated Information Platform (IIP) project was initiated in June 2004 and is a collaborative project involving academic institutions and companies active on the Norwegian Continental Shelf. Led by Det Norske Veritas, the project has a budget of about 22 million NOK (about 2.8 million Euro), and is closely coordinated with various standardization efforts in the petroleum industry. The overall objective is to use semantic technology to improve decision-making processes and reduce risks and costs in petroleum projects. In particular, the project will result in an open platform that supports semantic interoperability and intelligent information management for subsea production systems.

IIP is now completing one of the largest industrial ontologies for the terminology used in the petroleum business. Support from central industrial partners on the Norwegian Continental Shelf has been secured, and the project intends to propose this ontology as part of a new ISO standard. As such, it will also be available to companies and institutions that are not currently part of the project. Parts of the new ontology have been converted from ISO 15926 Integration of life-cycle data for oil and gas production facilities, but after looking into the terminologies used in selected petroleum projects, we also included concepts from other ISO standards. More than 40 000 concepts have now been defined and modelled in hierarchical conceptual structures. Figure 1 shows some of the concepts that must be defined just for the representation of wellheads. A Christmas tree is the set of valves, spools and fittings connected to the top of the well and used to control it's the fluid flow. The final ontology, which will be available in 2007 in OWL (Web Ontology Languages) with all the properties and rules incorporated, will be used to integrate petroleum applications, interpret real-time data from subsea installations, and access the information needed in decision processes.

Figure 1: Wellhead with Christmas tree and associated ontology classes.

Finding information quickly is important in petroleum processes. With the vast number of sensors and amount of communication equipment added to new installations, the problem is more to do with relevance than lack of information. There is an overwhelming amount of information available, from project documentation to streams on real-time data from subsea installations, and it may all be relevant to decisions that need to be made quickly and accurately. The project is implementing an ontology-driven approach to searching that interprets the user's query in terms of ontological concepts and associates these concepts with weighted terms used in documents and data records. The idea is to let these weighted terms define the concepts with respect to both the information available and the preferences of the user. They may be constructed on the basis of a training set provided by the user and/or based on her past behaviour. As the user is interacting with the system, her behaviour is observed and the system refines its understanding of the user's perception of the concepts. Rather than using the ontological hierarchies to expand the search and simply increase the amount of retrieved data, the system builds personalized descriptions of ontological concepts that help us better rank the documents with respect to users' interests and needs.
Future work includes a rule-based notification component that will be used to analyse anomalies in real-time data coming from sensors measuring the production of oil and gas. Rules in OWL specify properties of the equipment used and what actions should be taken if a constraint has been violated. Since the ontology is specified in OWL Full however, it is not clear how the reasoning capabilities can be added. OWL Full does not lend itself to standard Description Logic reasoning. We are therefore looking into alternative representations of these parts or other ways of adding reasoning capabilities to our semantic information management framework.

Links:
NTNU: http://www.ntnu.no
Norwegian Oil Industry Association: http://www.olf.no

Please contact:
Jon Atle Gulla, NTNU, Norway
Tel: +47 73591847
E-mail: jon.atle.gullaidi.ntnu.no