The Digital Twin in a Brownfield Environment: How to Manage Dark Data

Abstract

Abstract This paper addresses the challenges of creating a reliable and accurate as-operated digital twin for a brownfield process plant. Besides massive amounts of information spread over silos, there is so-called "dark data", which is a type of unstructured/untagged data in different formats: 1D (datasheets, lists, records), 2D (drawings, logical connectivity), 3D (physical layout, sizes). This paper describes an approach/solution that makes it possible to validate, connect and make use of this data. The solution described in this paper was developed collaboratively between Siemens and Bentley Systems. Using a combination of tools and experience from both companies, it is now possible to leverage traditionally inaccessible data to create a single digital twin. The paper will not focus on the solution itself, but rather the process and steps that operators can employ to aggregate, contextualize, validate, and visualize their data so that it can be used to make quicker/more informed decisions. The solution is an open, cloud-based platform that federates data sources, provides functionalities in the form of microservices, and facilitates the management of the digital twin of the asset, throughout its entire lifecycle. The solution was built as a set of microservices including collaborative process engineering and functional asset information management (i.e., 1D information such as specifications and data sheets and 2D schematics), and 3D design and physical asset modeling. It also contains information management from asset and project management perspectives, including maintenance history, reliability data, and failure mode analyses as well as analytics services and the capability to trace back versions. Many functionalities draw on software that has been used extensively across process industry for many years. The user is able to use pre-existing 3D plant models or projects authored in most available 3D design tools and mix and match those with models based on photogrammetry or 3D point clouds to create a "reality mesh." The portal accommodates the need for an up-to-date version of accurate asset data, which is critical to making the correct maintenance-related and operational decisions quickly as well as the requirement for historical snapshots (essential to engineering purposes and project progress reviews). This approach is key in that it enables O&G operators to solve the challenge of creating a digital twin without disruption to the existing physical or virtual environment. Combining 1D, 2D, and 3D in a single environment provides functional context to physical representations and vice versa. The more "dark data" that can be made visible, tagged, validated, and linked to other information, the more valuable and context-rich information will become.