Digital twin: An enabler for optimisation?

31 October 2022

Yanling Wu explains how digital twins are getting used to realize course of design and operational efficiency enhancements.

The time period digital twin is getting used extensively and has a wide range of interpretations. For the method business, digital twins ought to supply correct digital representations of bodily belongings, processes, and the management and automation methods that remodel information into actionable insights. They need to present evaluation and decision-making capabilities all through the venture or plant lifecycle. 

Digital twins embody first principle-based course of simulation, data-driven ML fashions, asset fashions and connections to real-world information synchronised from a number of sources together with operational information, engineering information, upkeep information and monetary info.

Digital twins can have functions throughout verticals servicing upstream and midstream oil and fuel, LNG, refining, petrochemicals, MMM (Metals, Mining and Minerals) and the rising sustainability segments. They’ll ship worth by means of the whole plant lifecycle throughout design, commissioning and start-up, and operation and sustainment by decreasing the time taken for design, enhancing operational productiveness, agility and reliability, and reducing down operational prices. 

Course of digital twins have been key instruments for course of industries for many years. So far, the use in preliminary alternative evaluation and know-how choice part and the Entrance-Finish Engineering Design (FEED) part is nicely accepted and executed for many initiatives. The design of a course of is, usually, modelled in a steady-state mannequin to overview various course of design and configuration choices. Customers can then choose a beneficial design foundation and help the event of course of move diagrams, warmth & materials stability, engineering research, and choice and sizing of key course of gear.  

Dynamic simulation-based digital twins type the inspiration of coaching programmes that embrace first-principle course of fashions and replication of management and security configurations and Human-Machine Interface (HMI). This strategy can present a protected, repeatable and interactive setting for operators to ‘study by doing’. Along with the first software of operator coaching, dynamic simulation-based digital twins are additionally helpful as instruments for transient engineering research, process growth, management and security system testing, validation, and delivering a protected start-up and commissioning. Dynamic simulation can serve a mess of makes use of in the course of the execution of the venture, and so, the time period Multi-Objective Dynamic Simulation (MPDS) has discovered its method into the method business.

Turning information into KPIs

In the course of the operation and sustainment part, plant digital twins can be utilized to remodel real-time information into key efficiency indicators (KPI) corresponding to catalyst actions, compressor effectivity and warmth switch coefficients. These can be utilized for efficiency monitoring, can detect early efficiency deterioration enabling preventive upkeep, determine constraints, and run what-if evaluation for knowledgeable determination making. 

Digital twins are enablers of end-to-end optimisation and autonomous operations by producing LP vectors for planning and scheduling instruments. They’ll additionally present necessary good points/constraints updates for APC functions so APC can precisely deal with non-linear processes. They can be used to generate artificial information for creating AI/ML fashions. 

Digital twins have been profitable in the course of the design part, however a quickly rising space with wider functions is within the lengthy operation and sustainment part. Challenges of digital twin functions on this part embrace easy mannequin conversion from the design part to operation part, steady mannequin efficiency monitoring and upkeep because the plant operation evolves and new information turn into obtainable. Different challenges embrace compatibility with different software program options, and quick mannequin execution for processing real-time information. 

These challenges could be overcome by leveraging a simulation platform that has each regular state and dynamic modelling below the identical setting. This modelling can permit straightforward transition between modes, equations-oriented modelling for robustness and pace, and extensible functionality. The extensibility allows finish customers to develop and deploy particular and proprietary modelling apps from easy to advanced fashions with out the necessity for specialised programming abilities.  Finish customers can allow the graceful integration of fashions and information and interoperability with third-party applied sciences together with normal and reusable asset fashions. Cloud-hosted deployment can improve information streaming and accessibility for distant cross-functional collaboration. AI/ML hybrid fashions and reduced-order fashions are choices to develop a fit-for-purpose mannequin for quicker mannequin execution and manageable mannequin upkeep. 

Yanling Wu is a product advertising supervisor at Honeywell Related Industrial.

Contact Particulars and Archive…