The 32nd International Workshop on Description Logics hosted in Oslo by Sirius* and the University of Oslo earlier this year, was billed as the major event of the DL research community. An industry panel heard from semantic practitioners in oil and gas and other verticals and served as a ‘reality check’ on the use of ontologies and reasoning in industry.
Christian Hansen (Aibel) spent the last five years working as a consultant in the oil and gas industry, applying semantic technologies and building ontology-based systems which use DL reasoning. The outcome is Aibel’s Material Master Data (MMD) engineering ontology that targets the management of complex requirements in new builds and modifications of oil and gas platforms. The Aibel MMD was notably deployed on Equinor’s Johan Sverdrup drilling platform. The Aibel MMD handles the complex requirements of cost estimates, material catalogs and data exchange and reuse. The result is possibly the ‘world’s largest’ industrial ontology with an OWL 2.0 encoded modular system that leverages public ontologies including SKOS, PAV (provenance), and others. The MMD includes some 1,840,769 axioms and 98,133 engineering data classes. The system was developed in Protégé, with an Oracle 12 master data repository. A Hermit OWL 2 reasoner controls export to SAP, AVEVA, EiS and CAD.
The system has eliminated data duplication and ensures that material catalogs containing only valid components. Incorrect component ordering has been reduced and system is said to be the foundation for future decision support/automation efforts and for a ‘digital twin’. On the downside, reasoning performance is poor across the large database. Temporal reasoning and lifecycle considerations are work in progress.
Johan Klüwer (DNV GL) explained how an enterprise ontology can automate ‘complex, but trivial work’ by capturing requirements in a formal language and validate solutions as an asset is developed and operated. A reference vocabulary allows for integration across project/plant lifecycle, between disciplines, between stakeholders and vendors, and across ‘thousands’ of specialized software applications’. Obstacles to the data nirvana include a lack of ontology specialists, methods and tools and the need for standard terminology to describe industrial artefacts and processes. Current standardization efforts are ‘small-scale and limited to upper ontologies’. In which context, Klüwer’s current focus is a methodology for digital requirements for the Norsok Z-TI joint industry project.
We asked Hansen if Aibel’s semantic effort has expanded from the R&D sector to sustainably impact the upstream. Here is his reply…
At Aibel we have used semantic technology and ontologies in production as support to our major EPC projects since 2015. Johan Sverdrup DP was the first EPC project to use MMD for management of material catalogs for piping bulk, valves, and structural steel. Since then, we have used the MMD ontology system on 5-6 other projects. The data quality in the project material catalogs has increased immensely, and we have almost eliminated duplicate entries in the material catalogs. Further, materials are grouped into piping specs, and for each spec, the material list only contains valid materials for that spec. As a result, project material procurement is completed with better accuracy, and we are better able to utilize warehouse stock in other projects. Also, the time needed for defining specs for new projects is dramatically reduced.
Much of the MMD content is based on international engineering standards, like ASME and ASTM, and we are actively looking into sharing those parts of MMD within the oil and gas community. At the moment, there are several initiatives in oil and gas (at least in Norway) to promoting ontologies and logic as ways to tackle (part of) the industry’s need for a common approach to digitalization and information exchange, including digital twin.
Those interested in following Norway’s persistent promotion of semantics and their extension into the field of AI might be interested in the upcoming workshop on formal methods and artificial intelligence in logistics. ‘Logistics 4.0’ advocates the universal digitalization of the supply chain to help to automate, verify and coordinate the execution of the procedures among the different stakeholders involved. The workshop will be held in Bergen on 2nd December 2019. More from Logistics 4.0.
* Centre for Scalable Data Access in the Oil and Gas Domain.© Oil IT Journal - all rights reserved.