Session: 01-08-01 Digitalization, AI/ML, Digital Twins - I

Submission Number: 156093

Data-Centric Approach to Design Digital Twins 

Objectives/Scope:

'Digital Twin' is a broad term used for the digital replica of real-world facilities in Oil &Gas or Energy business whether it's downstream or upstream, chemical plants, refineries, and electric grids.Many companies are building Digital Twins, but the scope definition could mean different things todifferent people. On one end, Digital Twin could be considered a simple visualization and easy-to-find-information tool, but on the other end it could be an industrial work surface where operators, planners,technicians, supervisors and maintenance professionals prefer to start and end their workday on. Inaddition, a Digital Twin serves as a dashboarding and ML/AI platform for analysing current data andpredicting future trends to drive important decision-making by the industry executives.

Methods, Procedures, Process:
The purpose of this work is to prescribe a recipe for the design of DigitalTwins based on a real deployment for one of the world's largest producers of LNG. The main feature set ina Digital Twin depends on its ability to ingest multiple data sources in various formats that include:
3D models generated by various PDMS software in different formats that could be available onlineor offline
Laser scans and surveys taken through aerial drones
High resolution 360 and Panoramic Images
ESRI and ArcGIS maps available online or offline
Documents (various types, formats, and source systems)
CMMS Objects from SAP like ERP systems available through API or offline CSV files
Real-time data - high frequency/large volumes-based data in the form of Alarms or Pi data ingestedthrough

Results, Observations, Conclusions:

The paper presents various critical factors to be considered with"data" at the epicentre to ensure that the application design serves all the customer needs. Most important factors include:
Security
Storage
Use-case agnostic data model
Extensible data model
Ingestion
Contextualization
Accessibility
Performance
Enable ML and AI drive decision making

Keeping all the above factors into consideration this paper provides recommendations for the design of aneffective Digital Twin with data-centric approach as well as a prescription for must-have foundationalblocks for a data-driven design.

Novel/Additive Information:
This work provides a design approach for Digital transformation usingDigital Twins. The paper presents a pragmatic approach by considering all the aspects important to andprioritized with operational customers. The work outlines starting with basics of data storage, and securitybut would go into advanced presentation of creating agnostic models, and contextualized enough to serveas useful mines for smart real-time decision making using AL/ML enablement.

Presenting Author: Gaurav Gupta Kongsberg Digital

Presenting Author Biography: Gaurav is a Technical Director at Kongsberg Digital, passionate about solving customer problems through innovative solutions. He is a Mechanical Engineer by education, and a curious software developer by work experience. He has been managing and delivering software products for last 20+ years in the Oil and Gas Industry.
He holds a bachelor's in engineering and an MBA in Strategy. He has a patent in Cementing technology to his name.