Introduction

Originating in the oil and gas industries – but having branched out to other sectors, too – CFIHOS is an international standard aimed at interoperability. The standard has been developed with the express ambition to improve and streamline the handover of information between partners within the industry value chain. It encompasses the various lifecycle phases of an operation. The shared vision: to improve data and information sharing, make data more accurate and minimize delays at handovers caused by errors, inefficiencies, and manual work — advantageous for both Principal and Contractor. 

Realizing this ambition will demand the successful integration of multiple data sources and moving from traditional, document-based data exchange to a digitized and software-neutral one. The correct information may be available without successful integration but will be scattered across multiple proprietary applications that each party is working with. Bottlenecks, duplications, and errors during information handover are, in that case, more likely to occur. Such hurdles can be overcome by combining essential resources and bringing them together in an appropriate digital environment for reuse by multiple parties.

The remainder of this paper will discuss an approach in which the following resources are integrated: the CFIHOS RDL, ISDDs, and extensions. With the use of these resources and by employing knowledge graph technology for their integration, parties involved in data handovers will be in a better position to specify their expectations for the management of asset lifecycles stored in an environment that can be accessed and retrieved in a software-neutral environment, accessible for use in data handovers. Moreover, this approach allows parties to dynamically assess the information needs of an asset rather than first accumulating all data requirements that may or may not be needed in communication at some point in time, effectively separating the wheat from the chaff.

Projects and essential information

On projects, the Principal, responsible for the lifecycle management of a plant facility, exchanges information with a Contractor. The information exchange during a project tends to concern a lifecycle phase, e.g., design, construction, or maintenance. The principal and Contractor require a shared notion of the current assets in the plant and the desired state or specifications of those objects. It is exactly on this shared ground, and the data exchange on any changes proposed or made during the project, that knowledge graph technology can greatly facilitate communication on the assets involved and for further management of these assets. Before discussing where this technology can be of use, we will first describe essential resources for expressing the state and specifications of assets in a plant facility.

• CFIHOS RDL. The CFIHOS Reference Data Library (or CFIHOS RDL) is the industry standard for information exchange and interoperability during the complete asset lifecycle. The RDL defines essential terminology for the types of objects and relevant properties that are common in plant facilities, such as ‘control valve’ and ‘circuit breaker.’
• ISDD’s. Industry Standard Datasheet Definitions (or ISDDs) describe what properties ought to be registered or made available for equipment and devices. Where an RDL defines what is meant by an asset, an ISDD prescribes what data is required to be handed over (such as the ‘design pressure of any pressure valve’). These information needs, which are to be filled out per registered asset, are elementary in data handovers.
• Extensions. In addition to the two resources mentioned above, further information may be demanded to describe the data necessary for exchange. Requirements on what the desired state is, specifications of equipment that can be purchased, and additional terminology that is more specialized than the standardized set in the CFIHOS RDL are examples of the extensions that may be required for data exchange on plant facilities.

Together RDLs, ISDDs, and extensions form an instrument for owner-operators, contractors, and manufacturers to share common cross-sector information and lower costs and risks in capital and operational expenditures (Oostinga, 2019). By specifying Exchange Information Requirements (EIR), the Principal and Contractor can agree on how data is to be exchanged in pre-specified contexts, such as on projects. By incorporating knowledge graph technology as part of these requirements, both parties can benefit from the resulting accessibility and interoperability of the information from these essential resources that they need for data handover. They can work with the tools they prefer if they exchange information according to the EIR. They can further enrich the knowledge from shared libraries and other resources by linking to additional standards and specifications where appropriate. Material that used to be handed over mostly as documents, such as ISDDs, can then be handed over in a manner in which software applications can dynamically access and parse the information within. So what is knowledge graph technology, and how can it help?

A knowledge-graph approach

Knowledge graphs are the result of arranging important information as networks. They consist of knowledge related to one another, forming a network or graph, with each node identified by its unique web address. As a result, knowledge graphs link data similar to how pages are linked on the Web. Data within the knowledge graph can be easily accessed from anywhere at any time, by authorized software applications, through the use of the Internet protocol. Since the nodes in the knowledge graph are related to one another using standardized technology, referred to as Linked Data, it is possible to organize the exchange and validation of information sustainably and efficiently for an entire value chain. This presents the opportunity for essential resources – including the CFIHOS RDL, ISDDs, and extensions – to be connected and shared between the Principal and Contractor.

Using Linked Data, a knowledge graph can be extended with additional data for different purposes. Industry standards, for instance, can be extended with company standards or even department and team standards. The term ‘pressure valve’,  managed by CFIHOS, defined in their RDL, and stored on a server in Norway, can be linked to a Supplier X pressure valve’ as part of a product catalog of a supplier, stored somewhere in Germany. This way, the sector can consistently standardize, distribute, and extend the required characteristics of ‘a Supplier X pressure valve’. Linked Data offers the ability to create consistent and coherent data whilst being managed by different authors.

The desired environment for a knowledge graph is one in which data is stored once in a centralized way but is used repeatedly in a distributed way. The data should be available to an ecosystem of contributors and consumers wishing to update specifications and offerings, as well as extendable to allow for agreements between parties in smaller environments within the ecosystem. Such needs call for interoperability between software, which is the ability of systems to work with other systems and is especially relevant in situations where multiple parties exchange data and which wish to select their software to operate on that data. A shared understanding of the information to be exchanged and processed is key in such situations. Knowledge graphs are an apt fit for capturing and sharing that common ground, facilitating interoperability. The information is not locked into a specific technology or service provider. Still, it is available and accessible to all parties involved, meaning that true data interoperability can be achieved in this manner.

Image 1.1. A graphical representation of the automation of EPC handovers in a software-neutral environment, using Linked Data to facilitate the linking and enrichment of datasets.

What does this look like in practice?

The Automation of EPC Handovers
Imagine, if you will, that an organization is the project principal responsible for the lifecycle management of a plant facility. The principal would like a contractor to work on the design of a related asset. The principal can indicate their CFIHOS-based specifications in the form of connected knowledge graphs – of the CFIHOS RDL, ISDDs, and extensions specific to the principal. Here, one can interlink these sets of information and include everything from a description of each type of asset to their properties and units of measurement (e.g., Imperial vs. Metric, etc.). The principal can then issue this knowledge to the contractor as their requirements, or master data, for the work required. 

Image 1.2. The CFIHOS RDL is configured. Physical objects, as described and specified in the ontology, are available for selection as part of a design specification for comparison against equivalent information in a separate dataset.

The contractor with information on the current or planned status of the plant can supplement the information from the knowledge graph. This information allows easy access to the specifications demanded. The information is returned to the principal and can be automatically read into configurable software packages. When the contractor is able to return the work they have completed, they can extend the knowledge graph with on-site information and perform a data handover accordingly. The return of the data set as a zipped, non-graphical data file to the contract principal, meeting the required standards, allows the principal to receive this data into their environment and consume the information in a format and language that fits their organization.

Image 1.3 Datasets stored in varying databases can be mapped to one another in Linked Data. Data transformation is carried out to facilitate the transfer and sharing of information in a language and format that the various parties can work within their native environment.

The benefits of this way of work and realizing this data transformation are:

For both parties:

• There is no substantial investment in IT infrastructure to change storage environments or data management principles by either party
• Both parties continue to work in their preferred environments, using the tools they have already invested in and configured with information from the knowledge graph shared for this purpose.
• Both parties are satisfied that they speak a common language and that their intentions and information are clearly understood. It is guaranteed that the product that is designed and supplied will meet the design specifications and requirements of the asset under management.
• There is no time delay in validating two data sets against one another. The parties work to the standards outlined, and such data transformations as measurements in cm to inches, in case these are necessary between principal and contractor, can be performed effortlessly through the information contained within the knowledge graph.

For Principals:

• Information needs can be indicated through links to the CFIHOS RDL. According to standardized ISDDs, several needs are already being connected, allowing Principals to profit from prior work by others.

For Contractors:

• Classifying an asset as a particular object (e.g., as a pump) will automatically provide insights into the information that needs to be registered for that asset. These information fields can be retrieved dynamically and employed in the software of preference. Thus, only the essential information needed for the context will be brought forward, providing an easy way to ensure conformance in data handovers.

Evidence

In their paper on the application of knowledge graphs and Linked Data technology at TenneT, a transmission system operator in the Netherlands and Germany that owns a large number of power stations, Stolk et al. describe their current use for asset and facility management (Stolk, 2022):

• Over the last decade, linked data and Semantic Web technology in information exchange have significantly increased in several European countries. The Netherlands, for instance, has widely adopted these technologies in sectors such as infrastructure and construction. In these sectors, information on assets owned by provinces, municipalities, and national organizations is not uncommonly exchanged using a linked data format that incorporates terminology captured in an ontology. 
• Linked data, including ontologies and JSON-LD, forms an essential part of this digitalization strategy in facilitating efficiency and accuracy in the communication between TenneT and contractors in a software-neutral manner.

Conclusion

Knowledge graphs and Linked Data technology make Reference Data Libraries and ISDDs easily accessible, directly applicable, and extendable for parties sharing data to manage the phases of an asset’s life cycle. By applying these principles, parties adopting the CFIHOS standard can rapidly and accurately exchange specifications for the management of plant asset lifecycle without the intervention of IT specialists. They can actively enrich the specifications for rapid digitization of asset lifecycle management.